Desmond User Manual

Desmond 3.1
User Manual

Schrödinger Press
Desmond User Manual Copyright © 2012 Schrödinger, LLC. All rights reserved.

While care has been taken in the preparation of this publication, Schrödinger
assumes no responsibility for errors or omissions, or for damages resulting from
the use of the information contained herein.

BioLuminate, Canvas, CombiGlide, ConfGen, Epik, Glide, Impact, Jaguar, Liaison,

LigPrep, Maestro, Phase, Prime, PrimeX, QikProp, QikFit, QikSim, QSite, SiteMap,
Strike, and WaterMap are trademarks of Schrödinger, LLC. Schrödinger and
MacroModel are registered trademarks of Schrödinger, LLC. MCPRO is a trademark
of William L. Jorgensen. DESMOND is a trademark of D. E. Shaw Research, LLC.
Desmond is used with the permission of D. E. Shaw Research. All rights reserved.
This publication may contain the trademarks of other companies.

Schrödinger software includes software and libraries provided by third parties. For
details of the copyrights, and terms and conditions associated with such included
third party software, see the Legal Notices, or use your browser to open
$SCHRODINGER/docs/html/third_party_legal.html (Linux OS) or
%SCHRODINGER%\docs\html\third_party_legal.html (Windows OS).

This publication may refer to other third party software not included in or with
Schrödinger software ("such other third party software"), and provide links to third
party Web sites ("linked sites"). References to such other third party software or
linked sites do not constitute an endorsement by Schrödinger, LLC or its affiliates.
Use of such other third party software and linked sites may be subject to third
party license agreements and fees. Schrödinger, LLC and its affiliates have no
responsibility or liability, directly or indirectly, for such other third party software
and linked sites, or for damage resulting from the use thereof. Any warranties that
we make regarding Schrödinger products and services do not apply to such other
third party software or linked sites, or to the interaction between, or
interoperability of, Schrödinger products and services and such other third party
software.

Revision A, September 2012

Contents
Document Conventions ..................................................................................................... ix

Chapter 1: Introduction ....................................................................................................... 1

1.1 Installation and Configuration ................................................................................ 2

1.2 The Maestro Interface to Desmond ........................................................................ 3

1.3 Desmond Calculations Overview ........................................................................... 4

1.4 Running Schrödinger Software .............................................................................. 5

1.5 Citing Desmond in Publications ............................................................................. 7

Chapter 2: Building a Model System ........................................................................ 9

2.1 Adding Solvent .......................................................................................................... 9

2.2 Setting Up the Boundary Box ............................................................................... 11

2.3 Adding a Membrane ................................................................................................ 11

2.4 Using Custom Charges .......................................................................................... 13

2.5 Specifying the Force Field ..................................................................................... 14

2.6 Adding Ions .............................................................................................................. 14

2.6.1 Defining an Excluded Region............................................................................ 15
2.6.2 Ion Placement ................................................................................................... 15
2.6.3 Adding a Salt..................................................................................................... 16

2.7 Running the Job ...................................................................................................... 16

2.8 Rebuilding a Model System ................................................................................... 17

2.9 Quick Setup Instructions ....................................................................................... 17

Chapter 3: Running Simulations from Maestro................................................ 19

3.1 Overview of the General Desmond Panels ......................................................... 19

3.2 Selecting a Model System ..................................................................................... 20

3.3 Minimizations ........................................................................................................... 21

3.4 Molecular Dynamics Simulations ......................................................................... 22

Desmond 3.1 User Manual iii

 Contents

3.5 Simulated Annealing Simulations ........................................................................ 24

3.6 Replica Exchange Simulations ............................................................................. 26

3.7 Metadynamics Simulations ................................................................................... 28

3.8 Simulations on Systems with Membranes ......................................................... 30

3.9 Setting Options for Desmond Simulations ......................................................... 31

3.9.1 The Integration Tab ........................................................................................... 32
3.9.2 The Ensemble Tab ............................................................................................ 33
3.9.3 The Minimization Tab ........................................................................................ 34
3.9.4 The Interaction Tab ........................................................................................... 35
3.9.5 The Restraints Tab............................................................................................ 36
3.9.6 The Output Tab ................................................................................................. 37
3.9.7 The Misc Tab..................................................................................................... 38

3.10 Running a Simulation Job ................................................................................... 39

Chapter 4: Running FEP Simulations .................................................................... 41

4.1 FEP Panels................................................................................................................ 41

4.2 Ligand Functional Group Mutation ...................................................................... 42

4.3 Ring Atom Mutation ................................................................................................ 44

4.4 Protein Residue Mutation ...................................................................................... 46

4.5 Total Solvation Free Energy Calculation ............................................................. 48

4.6 Selecting the Environment and FEP Protocol .................................................... 49

4.7 FEP Results .............................................................................................................. 51

4.7.1 Relative Free-Energy Differences ..................................................................... 51
4.7.2 Total Solvation Free Energies ........................................................................... 52
4.7.3 FEP Analysis..................................................................................................... 53

4.8 Customizing and Restarting FEP Simulations................................................... 53

Chapter 5: Analyzing Simulations ............................................................................. 57

5.1 Viewing Trajectories................................................................................................ 57

5.2 Simulation Quality Analysis .................................................................................. 61

iv Schrödinger Suite 2012 Update 2

 Contents

5.3 Simulation Event Analysis ..................................................................................... 62

5.4 Replica Exchange Review...................................................................................... 63

5.5 Metadynamics Analysis ......................................................................................... 64

Chapter 6: Running Simulations from the Command Line ...................... 67

6.1 The desmond Command ........................................................................................ 67

6.2 Running Multiple Simulations ............................................................................... 70

6.2.1 Examples of Running MultiSim ......................................................................... 72
6.2.2 Sample MultiSim Job (.msj) File ....................................................................... 72
6.2.3 Treatment of Intermediate Files ........................................................................ 75

6.3 Building a Model System ....................................................................................... 75

6.3.1 Reading the Structures .................................................................................... 76
6.3.2 Adding a Membrane......................................................................................... 78
6.3.3 Setting the Box Shape and Dimensions............................................................ 78
6.3.4 Setting Force Field Information ......................................................................... 79
6.3.5 Setting the Number and Location of Ions.......................................................... 79
6.3.6 Solvating the System ....................................................................................... 80
6.3.7 Writing the Output File ...................................................................................... 80

Chapter 7: Using VMD for Desmond Trajectories .......................................... 81

7.1 Reading a CMS File and a Desmond Trajectory ............................................... 81

7.2 Writing a Maestro File ............................................................................................. 82

Chapter 8: Alternate Force Field Parameters and Constraints............. 85

8.1 The viparr Utility ...................................................................................................... 85

8.2 The build_constraints Utility ................................................................................. 87

8.3 Input and Output Files ............................................................................................ 88

8.4 Specifying Multiple Force Fields .......................................................................... 89

8.5 User-Defined Force Fields ..................................................................................... 90

8.6 Known Issues ........................................................................................................... 91

Desmond 3.1 User Manual v

 Contents

Chapter 9: Utilities................................................................................................................ 93
9.1 solvate_pocket ......................................................................................................... 93
9.1.1 Methodology ..................................................................................................... 93
9.1.2 Command Syntax ............................................................................................. 94
9.1.3 Command File Syntax....................................................................................... 95

9.2 manipulate_trj.py ..................................................................................................... 98

9.3 amber_prm2cms.py ................................................................................................ 99

9.4 mold_gpcr_membrane.py .................................................................................... 100

9.5 trajectory_extract_frame.py ................................................................................ 101

9.6 desmond_restraints.py ........................................................................................ 102

9.7 rebuild_cms.py ...................................................................................................... 103

9.8 Other Scripts .......................................................................................................... 104

Appendix A: The multisim Utility............................................................................... 105

A.1 Running multisim ................................................................................................ 105
A.1.1 Template multisim Commands ....................................................................... 105
A.1.2 Node Locking.................................................................................................. 106
A.1.3 Restarting multisim Jobs ................................................................................ 106
A.1.4 Obtaining Information from multisim Checkpoint Files ................................... 108

A.2 The multisim File Syntax .................................................................................... 109

A.2.1 General Keywords .......................................................................................... 111
A.2.2 Desmond-Specific Common Keywords .......................................................... 111
A.2.3 The restrain Keyword...................................................................................... 112
A.2.4 The atom_group Keyword .............................................................................. 113
A.2.5 The task Stage ............................................................................................... 114
A.2.6 The system_builder Stage .............................................................................. 116
A.2.7 The build_geometry Stage ............................................................................. 117
A.2.8 The assign_forcefield Stage ........................................................................... 118
A.2.9 The simulate, replica_exchange, and lambda_hopping Stages ..................... 118
A.2.10 The minimize Stage ...................................................................................... 119

vi Schrödinger Suite 2012 Update 2

 Contents

A.2.11 The solvate_pocket Stage ............................................................................ 119

A.2.12 The extern Stage .......................................................................................... 121
A.2.13 The analysis Stage ....................................................................................... 123
A.2.14 The fep_analysis Stage ................................................................................ 124
A.2.15 The vrun and fep_vrun Stages ..................................................................... 125
A.2.16 The trim Stage .............................................................................................. 125

Appendix B: The Configuration File ....................................................................... 127

B.1 Ark Format Syntax Summary ............................................................................ 127

B.2 Units ....................................................................................................................... 128

B.3 Macros in String Values ...................................................................................... 128

B.4 Common Front-End Parameters ....................................................................... 129

B.4.1 fep................................................................................................................... 129
B.4.1.1 fep.lambda .......................................................................................... 130
B.4.1.2 fep.i_window ....................................................................................... 130
B.4.1.3 Default Lambda Schedules .................................................................. 130
B.4.2 cutoff_radius ................................................................................................... 131
B.4.3 taper ............................................................................................................... 131
B.4.4 coulomb_method ............................................................................................ 132
B.4.5 temperature .................................................................................................... 132
B.4.6 annealing ....................................................................................................... 133
B.4.7 pressure.......................................................................................................... 133
B.4.8 surface_tension .............................................................................................. 133
B.4.9 ensemble ........................................................................................................ 133
B.4.10 time ............................................................................................................... 134
B.4.11 elapsed_time ................................................................................................ 134
B.4.12 timestep ........................................................................................................ 134
B.4.13 cpu ................................................................................................................ 135
B.4.14 glue ............................................................................................................... 135
B.4.15 trajectory....................................................................................................... 135
B.4.16 eneseq .......................................................................................................... 136
B.4.17 checkpt ........................................................................................................ 136
B.4.18 maeff_output ................................................................................................ 137

Desmond 3.1 User Manual vii

 Contents

B.4.19 randomize_velocity ...................................................................................... 137

B.4.20 simbox_output .............................................................................................. 138
B.4.21 energy_group .............................................................................................. 138
B.4.22 backend ........................................................................................................ 138

B.5 Parameters for Minimization .............................................................................. 139

B.5.1 max_steps ..................................................................................................... 139
B.5.2 convergence ................................................................................................... 139
B.5.3 steepest_descent_steps ................................................................................ 139
B.5.4 num_vector ..................................................................................................... 139

B.6 Parameters for Replica Exchange Simulations .............................................. 140

B.6.1 replica ............................................................................................................. 140

B.7 Parameters for Metadynamics Simulations .................................................... 141

B.7.1 meta................................................................................................................ 141
B.7.2 cv .................................................................................................................... 141

B.8 Parameters for vrun ............................................................................................ 142

B.8.1 vrun_frameset................................................................................................ 142

Appendix C: Analyzing a Simulation from the Command Line ........... 143

C.1 simulation_block_data.py ................................................................................. 143

C.2 simulation_block_test.py .................................................................................. 144

C.3 Simulation Block Analysis (.sba) File Syntax ................................................. 144

C.4 Simulation Block Test (.sbt) File Syntax .......................................................... 145

References .............................................................................................................................. 147

Getting Help ........................................................................................................................... 151

Index ............................................................................................................................................ 155

viii Schrödinger Suite 2012 Update 2

Document Conventions

In addition to the use of italics for names of documents, the font conventions that are used in
this document are summarized in the table below.

Font Example Use

Sans serif Project Table Names of GUI features, such as panels, menus,
menu items, buttons, and labels
Monospace $SCHRODINGER/maestro File names, directory names, commands, envi-
ronment variables, command input and output
Italic filename Text that the user must replace with a value
Sans serif CTRL+H Keyboard keys
uppercase

Links to other locations in the current document or to other PDF documents are colored like
this: Document Conventions.

In descriptions of command syntax, the following UNIX conventions are used: braces { }
enclose a choice of required items, square brackets [ ] enclose optional items, and the bar
symbol | separates items in a list from which one item must be chosen. Lines of command
syntax that wrap should be interpreted as a single command.

File name, path, and environment variable syntax is generally given with the UNIX conven-
tions. To obtain the Windows conventions, replace the forward slash / with the backslash \ in
path or directory names, and replace the $ at the beginning of an environment variable with a %
at each end. For example, $SCHRODINGER/maestro becomes %SCHRODINGER%\maestro.

Keyboard references are given in the Windows convention by default, with Mac equivalents in
parentheses, for example CTRL+H (H). Where Mac equivalents are not given, COMMAND
should be read in place of CTRL. The convention CTRL-H is not used.

In this document, to type text means to type the required text in the specified location, and to
enter text means to type the required text, then press the ENTER key.

References to literature sources are given in square brackets, like this: [10].

Desmond 3.1 User Manual ix

x Schrödinger Suite 2012 Update 2
 Desmond User Manual

Chapter 1

Chapter 1: Introduction

Desmond is an explicit-solvent molecular dynamics program developed by D. E. Shaw

Research. Desmond was created from scratch with an emphasis on accuracy, speed and scal-
ability. It supports many of the most sought-after features in a modern molecular dynamics
program, including:

• Highly scalable parallel execution

• Explicit solvent simulations with periodic boundary conditions using cubic, orthorhom-
bic, and triclinic simulation boxes. Truncated octahedron and rhombic dodecahedron are
supported via their triclinic analogues.
• Support for isotropic, semi-isotropic and anisotropic pressure coupling
• Smooth particle mesh Ewald method for accurate and efficient evaluation of long-range
electrostatics
• NVE, NVT, NPT, NPAT, NPγT ensembles with Berendsen, Langevin, or Nosé-Hoover
thermostats, and Berendsen, Langevin, or Martyna-Tobias-Klein barostats
• Symplectic integration of the equations of motion using a multiple time step approach,
RESPA
• Elimination of many sources of numerical error, permitting accurate and fast calculations
using single-precision arithmetic
• Accurate implementation of constraints to eliminate high-frequency motions and thus
permit larger time steps
• Exploitation of modern computer chip features to enhance speed (SIMD)
• Efficient calculation of the pressure
• Accurate checkpointing mechanism for continuing or restoring simulations
• Template-based support for widely-used force fields (using viparr).
• Viewing of trajectories with VMD using a Desmond plug-in.

A description of Desmond was published, along with performance data, as part of the confer-
ence proceedings of the ACM/IEEE Conference on SuperComputing 2006 (SC06) [1]. While
developing Desmond, D. E. Shaw Research has introduced and extended a number of scientific
algorithms, including new parallelization strategies and numerical techniques, some of which

Desmond 3.1 User Manual 1

 Chapter 1: Introduction

have been published [2–5].

Problem-solving often involves using a wide range of modelling techniques, so integrating

Desmond into Schrödinger’s premier molecular modelling suite for drug development
enhances the utility of both. Examples of such synergies include:

• The Protein Preparation Wizard, LigPrep (ligand structure) and Epik (ligand protonation
state) preparation tools can be used to ensure that the structures provided to Desmond are
chemically correct. Such careful system preparation often represents a crucial step prior
to initiating a molecular dynamics simulation.
• Prime can be used to create homology models for use in simulations and to repair protein
structures.
• Glide can be used to generate relevant poses within protein binding sites for use in simu-
lations. Desmond in turn can be used to thermally relax, refine, and sample conforma-
tions related to the docked poses.
• Strike can be used to generate statistical models from the results of simulations.
• Desmond can be used to sample protein structures prior to performing docking calcula-
tions with Glide.
• SiteMap can be used to identify potential binding sites from simulation results.
• WaterMap analyzes specially designed Desmond simulations to characterize the thermo-
dynamics of water in protein binding sites.

1.1 Installation and Configuration

Desmond is supported on x86 hardware under Linux and Mac OS X, and is available in both
32-bit and 64-bit versions. Detailed requirements and installation and configuration instruc-
tions are given in the Installation Guide. Desmond itself is not supported on Windows, but the
Maestro graphical interface is supported, and you can run jobs on remote Linux or Mac hosts
from Windows.

Although Desmond can run serially, for most purposes, you will want to make use of the
parallel execution capabilities. Desmond uses Open MPI for parallel execution. Before you can
run jobs, however, you must add entries to the hosts file for parallel execution with Open MPI,
in addition to any configuration that is needed for the hosts and the queueing system. See the
Installation Guide for instructions, especially Chapter 7 and Section 7.3.3.

2 Schrödinger Suite 2012 Update 2

 Chapter 1: Introduction

1.2 The Maestro Interface to Desmond

A number of Maestro panels have been provided to streamline the process of setting up,
running and understanding the results of Desmond jobs so that you can focus on what you are
studying. These panels can be used on all supported platforms to prepare and analyze jobs
(Linux, Mac, and Windows).

In addition, much of the framework for running Desmond jobs has been written in Python to
facilitate adaption to user-specific requirements, including the automation of larger and more
specific workflows.

System Builder panel

• Constructs systems suitable for simulation using periodic boundary conditions

• Bulk solvent and membrane environments supported
• Solvation and neutralization largely automated yet customizable
• Seamless force-field parameter assignment
• Can be run locally on Windows
Minimization panel
Molecular Dynamics panel
Simulated Annealing panel
Replica Exchange panel
Metadynamics panel

• Desmond job launching

• Ability to intuitively see and adjust the key parameters used by the Desmond program for
both minimization and dynamics calculations
• Default parameters are suitable for many simulations
• Intelligent coupling of related settings
• Access to both minimization and simulation settings
• Easy simulation continuation and restoration
• Optional automated relaxation and equilibration procedures
Ligand Functional Group Mutation by FEP panel
Ring Atom Mutation by FEP panel
Protein Residue Mutation by FEP panel
Total Free Energy by FEP panel

• Easy to use with a focus on the real problem of interest rather than the details of the cal-
culation
• Support for absolute and relative solvation free energy calculations
• Support for relative binding free energy calculations
• Restarting and customization of FEP jobs via FEP panel

Desmond 3.1 User Manual 3

 Chapter 1: Introduction

Trajectory Viewer panel

• Integrated into Maestro

• Comprehensive speed controls
• Replication of primary simulation box for viewing
• Easy creation of images and movies.
Simulation Quality Analysis panel

• Intuitive tool for examining certain markers of simulation quality

Simulation Event Analysis panel

• Intuitive tool for investigating what happened during a simulation

The use of these panels is described in subsequent chapters of this manual.

Most Desmond-related tools are available from the Desmond submenu of the Applications
menu in Maestro. The exceptions are the trajectory viewer, which is opened from the Project
Table using the output entry for a job. Desmond panels can also be opened from Tasks →
Molecular Dynamics.

1.3 Desmond Calculations Overview

Desmond jobs should be started from well-prepared structures. For proteins it is recommended
that the protein be prepared with the Protein Preparation Wizard (see the Protein Preparation
Guide for details). For other types of molecules, such as ligands, the molecule should have a
fairly good Lewis structure (although there are some built-in capabilities for adjusting incorrect
or non-optimal Lewis structures).

If you have MacroModel you can perform a quick check on the structure by performing a
Current Energy calculation (available from the MacroModel submenu of the Applications menu)
using the OPLS_2005 force field with the Solvent set to None. If that calculation succeeds it is
almost certain that Desmond and its associated tools will be able to work with this structure as
well. If the structure is problematic Maestro and MacroModel often provide useful diagnostics
for what might be wrong.

Performing a study based on a Desmond molecular dynamics simulation usually involves a

number of stages, including simulation setup, relaxing the system (this could just be a minimi-
zation), running the simulation, viewing trajectories, and analyzing the results. Simulation
setup is described in Chapter 2, the basic Desmond minimization, molecular dynamics, simu-
lated annealing, replica exchange, and metadynamics tasks are described in Chapter 3. Simpli-
fied FEP setup for relative binding and solvation free energies and absolute solvation free
energies is described in Chapter 4, along with restarting and customizing FEP simulations.

4 Schrödinger Suite 2012 Update 2

 Chapter 1: Introduction

Examining the results, including viewing a trajectory, and analysis of results, is described in
Chapter 5.

FEP jobs are handled differently due to the complexity of the calculations and the fact that the
overall goal for an FEP job is to produce one number: the free energy change. FEP jobs are
supported for specific types of calculations, using automated procedures that differ from those
used for individual, general purpose Desmond simulations.

The basic outline of a Desmond simulation as run from Maestro is as follows:

1. Import the structure file for the system of interest into Maestro.
2. Prepare the structure for simulation with the Protein Preparation Wizard. This step
involves removing ions and molecules (which are artifacts of crystallization), setting cor-
rect bond orders, adding hydrogens, filling in missing side chains or whole residues as
necessary, reorienting various groups and varying residue protonation states to optimize
the hydrogen bonding network, and then checking the structure carefully.
3. If your system is a membrane protein, embed the protein in the membrane. This step and
the next two steps are performed in the System Builder panel.
4. Generate a solvated system for simulation.
5. Distribute positive or negative counter ions to neutralize the system, and introduce addi-
tional ions to set the desired ionic strength (when necessary).
6. Relax the system either by minimization or by selecting the panel option to relax the
model system before simulation.
7. Set the simulation parameters in one of the general Desmond panels, for molecular
dynamics, simulated annealing, or replica exchange.
8. Run the simulation.
9. Analyze your results using the Trajectory Viewer and other analysis tools.

1.4 Running Schrödinger Software

Schrödinger applications can be started from a graphical interface or from the command line.
The software writes input and output files to a directory (folder) which is termed the working
directory. If you run applications from the command line, the directory from which you run the
application is the working directory for the job.

Desmond 3.1 User Manual 5

 Chapter 1: Introduction

Linux:

To run any Schrödinger program on a Linux platform, or start a Schrödinger job on a remote
host from a Linux platform, you must first set the SCHRODINGER environment variable to the
installation directory for your Schrödinger software. To set this variable, enter the following
command at a shell prompt:

csh/tcsh: setenv SCHRODINGER installation-directory

bash/ksh: export SCHRODINGER=installation-directory

Once you have set the SCHRODINGER environment variable, you can run programs and utilities
with the following commands:

$SCHRODINGER/program &
$SCHRODINGER/utilities/utility &
You can start the Maestro interface with the following command:

$SCHRODINGER/maestro &
It is usually a good idea to change to the desired working directory before starting Maestro.
This directory then becomes Maestro’s working directory.

Windows:

The primary way of running Schrödinger applications on a Windows platform is from a graph-
ical interface. To start the Maestro interface, double-click on the Maestro icon, on a Maestro
project, or on a structure file; or choose Start → All Programs → Schrodinger-2012 > Maestro.
You do not need to make any settings before starting Maestro or running programs. The default
working directory is the Schrodinger folder in your documents folder (Documents on Windows
7/Vista, My Documents on XP).

If you want to run applications from the command line, you can do so in one of the shells that
are provided with the installation and that have the Schrödinger environment set up:

• Schrödinger Command Prompt—DOS shell.

• Schrödinger Power Shell—Windows Power Shell (if available).
You can open these shells from Start → All Programs → Schrodinger-2012. You do not need to
include the path to a program or utility when you type the command to run it. If you want
access to Unix-style utilities (such as awk, grep, and sed), preface the commands with sh, or
type sh in either of these shells to start a Unix-style shell.

6 Schrödinger Suite 2012 Update 2

 Chapter 1: Introduction

Mac:

The primary way of running Schrödinger software on a Mac is from a graphical interface. To
start the Maestro interface, click its icon on the dock. If there is no Maestro icon on the dock,
you can put one there by dragging it from the SchrodingerSuite2012 folder in your Applications
folder. This folder contains icons for all the available interfaces. The default working directory
is the Schrodinger folder in your Documents folder ($HOME/Documents/Schrodinger).

Running software from the command line is similar to Linux—open a terminal window and
run the program. You can also start Maestro from the command line in the same way as on
Linux. The default working directory is then the directory from which you start Maestro. You
do not need to set the SCHRODINGER environment variable, as this is set in your default envi-
ronment on installation. If you need to set any other variables, use the command

defaults write ~/.MacOSX/environment variable "value"

1.5 Citing Desmond in Publications

The use of this product should be acknowledged in publications as:

Desmond Molecular Dynamics System, version 3.1, D. E. Shaw Research, New York, NY,
2012. Maestro-Desmond Interoperability Tools, version 3.1, Schrödinger, New York, NY,
2012.

Please also include a reference to the following paper:

Kevin J. Bowers, Edmond Chow, Huafeng Xu, Ron O. Dror, Michael P. Eastwood, Brent A.
Gregersen, John L. Klepeis, Istvan Kolossvary, Mark A. Moraes, Federico D. Sacerdoti, John
K. Salmon, Yibing Shan, and David E. Shaw, “Scalable Algorithms for Molecular Dynamics
Simulations on Commodity Clusters,” Proceedings of the ACM/IEEE Conference on Super-
computing (SC06), Tampa, Florida, November 11-17, 2006.

Desmond 3.1 User Manual 7

8 Schrödinger Suite 2012 Update 2
 Desmond User Manual

Chapter 2

Chapter 2: Building a Model System

Performing simulations on aqueous biological systems requires the preparation of biological

molecules such as proteins and ligands, addition of counter ions to neutralize the system, selec-
tion of simulation box size, solvation of the solutes using explicit solvent molecules, and align-
ment of proteins to a membrane bilayer (if used). This procedure is often tedious if it has to be
performed manually. Tools for all these tasks are provided with Maestro.

Protein and ligand structures used in a Desmond simulation must be complete all-atom 3D
structures with a reasonable geometry. The preparation of protein and ligand structures for use
in a simulation can be done with the Protein Preparation Wizard and LigPrep. The Protein
Preparation Wizard corrects structural defects, adds hydrogen atoms, assigns bond orders, and
can selectively assign tautomerization and ionization states, and optimize the hydrogen
bonding network. For more information, see the Protein Preparation Guide. LigPrep performs
2D-to-3D conversion if necessary, adds hydrogen atoms, generates tautomers, ionization
states, ring conformations, and stereoisomers, as requested, and produces minimized 3D struc-
tures. For more information, see the LigPrep User Manual.

Once you have prepared the protein and ligand structures, you can proceed to the remaining
tasks in building a model system that can include proteins, ligands, explicit solvent, a
membrane, and counter ions. The System Builder automates this process and significantly
reduces the effort required. You can set up and run a System Builder job from the System
Builder panel, or from the command line. See Section 6.3 on page 75 for information on
running the System Builder from the command line.

To open the System Builder panel, do one of the following:

• Choose Applications → Desmond → System Builder

• Choose Tasks → Molecular Dynamics → System Setup
Before you start working in this panel, display the solutes in the Workspace.

2.1 Adding Solvent

The solvation model is selected in the Solvation tab. You can choose from a set of predefined
solvent models, or specify a custom solvent model:

• None—Do not use a solvent. This option allows you to run a simulation on a pure liquid,
for example, or in vacuum (with a sufficiently large box).

Desmond 3.1 User Manual 9

 Chapter 2: Building a Model System

Figure 2.1. The Solvation tab of the System Builder panel.

• Predefined—Use one of the predefined solvent models, which you can select from the
option menu. The models include four water models, SPC, TIP3P, TIP4P, and TIP4PEW,
and three organic solvents, methanol, octanol, and dimethyl sulfoxide (DMSO).
• Custom—Import a custom solvent system from file. Enter the solvent system file name in
the text box, or click Browse and navigate to the solvent system file in the file selector
that is displayed.

The solvent is placed by replicating “boxes” of solvent molecules and deleting molecules
whose center of mass lies outside the periodic box boundary, and molecules that are inside or
have significant overlap with the solute or the membrane (if one is used).

10 Schrödinger Suite 2012 Update 2

 Chapter 2: Building a Model System

2.2 Setting Up the Boundary Box

The periodic boundary conditions are set up by specifying the shape and size of the repeating
unit, or box, which you can do in the Solvation tab.

To set up the box, first choose the shape from the Box shape option menu. Three basic shapes
are provided: Cubic, Orthorhombic, and Triclinic. As special cases of the triclinic box shape,
three other shapes are supported: Truncated octahedron, Rhombic dodecahedron xy-square,
and Rhombic dodecahedron xy-hexagon.

When you have chosen the box shape, you can choose whether to specify the size of the box in
terms of a buffer distance or as an absolute size, by selecting one of the Box size calculation
method options:

• Buffer—The simulation box size is calculated by using the given buffer distance between
the solute structures and the simulation box boundary.
• Absolute size—Specify the lengths of the sides of the simulation box size (and angles if
necessary).

Having chosen a method, you can specify the distances and angles in the Distances and Angles
text boxes. The text boxes that are available depend on the box shape. For all choices except a
truncated octahedron, the box can be displayed in the Workspace by selecting Show boundary
box.

If you want to calculate the volume of the box that encompasses the solutes, click Calculate.
The volume is displayed in the Box volume text box. To minimize the volume of the box, click
Minimize Volume. The solutes are reoriented so that the box volume is minimized.

2.3 Adding a Membrane

A membrane can be added to the system using the Set Up Membrane dialog box, which you
open by clicking Set Up Membrane in the Solvation tab. This dialog box allows you to select
and position the membrane; the actual membrane is added when the system builder job is run.

There are four predefined membranes, DPPC, DMPC, POPC, and POPE, which you can
choose by selecting Predefined, and choosing the membrane from the option menu. The
temperature at which the membrane patch was preequilibrated is given in parentheses after the
membrane name. Because DPPC has a gel transition temperature around 313 K, the recom-
mended minimum simulation temperature is also 325 K.

If you want to position a custom membrane, select Custom, and enter the name of the Maestro
file containing the membrane model in the text box, or click Browse and navigate to the file.

Desmond 3.1 User Manual 11

 Chapter 2: Building a Model System

Figure 2.2. The Set Up Membrane panel.

If you have an existing membrane in a project entry that you want to use for the current model
system, you can load it by selecting the entry and clicking Load Membrane Position from
Selected Entry. The membrane from this entry is then used for the model system you are
building.

When you click Place Automatically, the membrane position is determined according to the
information available, as follows:

• If you have a protein from the OPM database (http://opm.phar.umich.edu), the membrane
is placed using the information provided with the protein.
• Otherwise the surface of the membrane is placed perpendicular to the longest axis of the
protein.
• If transmembrane atoms are defined, they are placed inside the membrane. Placement of
transmembrane atoms inside the membrane takes precedence over placement perpendicu-
lar to the longest axis.
To define the transmembrane atoms, you can do one of the following:
• Click Select, and use the Atom Selection dialog box to select the desired atoms. For
more information on this dialog box, see Section 6.5 of the Maestro User Manual.
• Click Set to Helices. This button sets the ASL expression to res.sec helix,
which selects the residues whose secondary structure assignment is “helix”. The
protein must have a secondary structure assignment before you use this button. To
do the assignment, choose Tools → Assign Secondary Structure in the main win-
dow.

12 Schrödinger Suite 2012 Update 2

 Chapter 2: Building a Model System

If you have a protein that is prealigned, you can click Place on Prealigned Structure to place
the membrane. The membrane is positioned symmetrically about the coordinate origin so that
its surfaces are parallel to the xy plane (perpendicular to the z axis). This means that the protein
must be aligned accordingly.

When you have placed the membrane, a representation of the membrane is displayed in the
Workspace. The representation consists of two red slabs for the surfaces, with a yellow line
perpendicular to the slab planes. After the membrane has been placed, you can adjust its orien-
tation by selecting Adjust membrane position, and rotating the membrane. The actual
membrane molecules are placed when the system builder job is run. The molecules are placed
by replication of a membrane segment and deletion of molecules whose center of mass lies
outside the periodic box boundaries. Molecules that are inside the solute or have significant
overlap with it are removed to accommodate the solute.

If you click Place Automatically after adjusting the membrane, the membrane is returned to its
default position and orientation.

The membrane position and orientation can be stored in Project Table entries, by selecting the
entries in the Project Table, and clicking Save Membrane Position to Selected Entries. This
enables the membrane position and orientation to be loaded at a later time by selecting the
entry and clicking Load Membrane Position from Selected Entry.

It can be difficult to set up GPCR systems properly. The mold_gpcr_membrane.py script can
be used to swap your GPCR protein into a system that has already been constructed for a
related protein. Templates for a number of GPCR systems are available. See Section 9.4 on
page 100 for details.

2.4 Using Custom Charges

If you want to use partial charges from a source other than the force field, you can do so by
selecting Use custom charges in the Solvation tab. You can then choose to use the partial
charges from the structure, or enter the name of the property that defines the custom charges in
the Custom text box. The property name is the internal name, which should start with r_ (i.e. a
real-valued property). For example, the property r_j_ESP_Charges selects Jaguar-generated
ESP charges.

When you have selected the property, click Select to select the atoms for which these charges
are to be used. There is no default. The selection is made in the Atom Selection dialog box,
which is described in detail in Section 6.5 of the Maestro User Manual. If the property you
chose has values only for some of the atoms (e.g. the ligand), you can select these atoms by
specifying the entire range of values. Atoms that do not have a value for the property will not
be selected.

Desmond 3.1 User Manual 13

 Chapter 2: Building a Model System

Figure 2.3. The Ions tab of the System Builder panel.

2.5 Specifying the Force Field

The force field is specified from the choices on the Force field option menu. The default choice
is OPLS_2005. To specify other force fields, you can run the utility viparr—see Chapter 8.

2.6 Adding Ions

It is usually desirable to have an electrically neutral system for simulation (though not strictly
necessary, as Desmond applies a uniform background charge distribution to neutralize the
system in the Ewald summation). You can choose to add ions to neutralize the system in the Ion
placement section of the Ions tab. The system can also be set up in a salt solution rather than a
pure solvent in the Add salt section of the Ions tab. To limit the locations in which ions can be
placed, you can define regions from which ions are excluded, in the Exclude region section of
the Ions tab.

14 Schrödinger Suite 2012 Update 2

 Chapter 2: Building a Model System

2.6.1 Defining an Excluded Region

To define an excluded region, click Select in the Excluded region section of the Ions tab, and
use the Atom Selection dialog box to select the desired set of residues. You should select resi-
dues that are within or near the binding site. When ions are placed, they will not be placed near
these residues. The residues that you select are colored blue and rendered in CPK. See
Section 6.5 of the Maestro User Manual for more information on the Atom Selection dialog
box. The region is defined by the distance in the Exclude ion and salt placement within text box.
Ions will not be placed within the specified distance of the selected atoms.

2.6.2 Ion Placement

Ions are placed in the solvent according to your selection in the Ion placement section of the
Ions tab. Each ion replaces a solvent molecule. You can, of course, choose not to add ions, by
selecting None.

If you select Neutralize, the minimum number of sodium or chloride ions required to balance
the system charge is placed randomly in the solvent.

If you select Add, you can choose the ion type from the option menu and enter the number of
ions to add (which need not neutralize the system). The option menu only displays ions that are
opposite in charge to that of the system. Ions are not placed in the excluded regions.

Instead of placing ions automatically, at random, you can locate and select suitable regions for
ions to be placed. Usually these regions are near residues that have the same charge as the
system charge and are not near the active site. You can define these regions in the Advanced Ion
Placement dialog box, which you open by clicking Advanced Ion Placement in the Ions tab.

To place the ions, you must identify suitable candidate residues. When you click Candidates,
the Candidates table is populated with a list of residues in regions that have not been excluded
and have the same charge as the overall charge of the system. These residues are colored red
and rendered in CPK. Ions are placed near the residues that you select in the table, replacing
the closest solvent molecule to the average position of the atoms in the residue. The number of
ions placed (initially 0), along with the number of ions remaining to be placed and the total
number of candidate residues are displayed above the table.

You can add candidates to the table by clicking Select, and selecting the residues in the Atom
Selection dialog box. When you click OK in the dialog box, the residues are added to the table,
and can be selected along with the automatically located residues. To clear the table, click
Reset.

Desmond 3.1 User Manual 15

 Chapter 2: Building a Model System

Figure 2.4. The Advanced Ion Placement dialog box.

When the system builder job is run, ions that are placed using the Advanced Ion Placement
dialog box are placed first. Once these ions are placed, random placement is performed to
place any remaining ions that are needed to neutralized the system or complete the number of
ions selected for placement in the Add text box.

2.6.3 Adding a Salt

Adding a salt is relatively simple. To do so, first select the Add salt option. The controls in the
Add salt section are then activated, and you can enter the salt concentration, in mol dm–3, and
select the desired ions. If you select multiply charged ions, the concentration is taken from the
empirical formula for the salt. For example, for MgCl2 the concentration of Mg2+ would be the
specified concentration and the concentration of Cl– would be twice the specified concentra-
tion. A value of 0.15M is approximately the physiological concentration of monovalent ions.

When the salt ions are placed, they are randomly distributed in the solvent, and replace solvent
molecules. They are not placed in the excluded region defined in the Exclude region section.

2.7 Running the Job

When you have finished making settings, you can set up and start the job immediately, or write
out the input file and run the job from the command line.

To set up and run the job, click Start. The Start dialog box opens, allowing you to name the job,
choose the host and set the user name (if necessary). System Builder jobs do not usually take
more than a few minutes, so you can run the job locally on Linux systems (but not on
Windows). You can also choose whether to incorporate the output CMS file back into the
Maestro project, by choosing Append new entries from the Incorporate option menu. This is
useful if you want to continue on to set up a simulation in Maestro. If you choose Do not incor-
porate, the CMS file is placed in the current working directory, but is not added to the project.

16 Schrödinger Suite 2012 Update 2

 Chapter 2: Building a Model System

If you want to run the job from the command line, click Write. The Write dialog box opens, in
which you can specify a name and then write the file. The name is used to construct the file
names, by adding the appropriate extension.

2.8 Rebuilding a Model System

If you want to modify an existing model system (full system CT), and run a simulation on it,
you do not need to completely rebuild the system in the System Builder panel. Instead, you can
simply regenerate the model system. To do so, include the full model system in the Workspace,
then choose Applications → Desmond → Model System Regeneration. A Start dialog box
opens, so you can specify the usual job parameters.

If you want to rebuild a model system from the command line, you should run one job from
Maestro, and then edit the multisim input file (.msj) for other model systems.

2.9 Quick Setup Instructions

The sets of instructions below take you through the simplest setup procedures. It is assumed
that you have imported the prepared protein and ligand structures into Maestro, and displayed
them in the Workspace.

To add solvent:
1. Select Predefined for the Solvent model option, and choose a model from the option
menu.
2. Choose a box shape.
3. Choose a box size calculation method—Buffer for adding a buffer region to the solutes, or
Absolute size for specifying the actual box size.

4. Enter buffer distances or side lengths in the available text boxes.

5. Enter angles if you selected Triclinic for the box shape.

To add ions:
1. In the Ions tab, choose an option for the addition of ions.
2. If you selected Add, enter the number of ions to add in the text box.
3. Choose an ion type from the option menu.
4. If the solvent is intended to be a salt solution, select Add salt.
5. Enter the desired salt concentration in the Salt concentration text box.

Desmond 3.1 User Manual 17

 Chapter 2: Building a Model System

6. Choose positive and negative ion types from the Salt positive ion and Salt negative ion
option menus.

To add a membrane:
1. Click Add Membrane in the Solvation tab.
2. In the Membrane tab, select Predefined for the membrane model, and choose a membrane
type from the option menu.
3. Click Place Automatically.
4. Select Adjust membrane position and adjust the orientation of the membrane in the Work-
space.
5. Click OK.

Click Start to run the job or click Write to write the input file.

18 Schrödinger Suite 2012 Update 2

 Desmond User Manual

Chapter 3

Chapter 3: Running Simulations from Maestro

The general Desmond panels enable you to set up and run the main tasks available with
Desmond: molecular dynamics, minimization, simulated annealing, and replica exchange jobs.
The panels are designed to make setting up these types of jobs as easy as possible, and provide
the most common simulation controls. The default values provided in the panels represent a
good balance between accuracy and performance, and are adequate for most jobs without
change. For more control over the simulation parameters, you can make settings in the
Advanced Options dialog box, which is described in Section 3.9 on page 31.

A much more automated approach is provided for FEP simulations of binding and solvation
free energies in four specialized panels, Ligand Functional Group Mutation by FEP, Ring Atom
Mutation by FEP, Protein Residue Mutation by FEP, and Total Free Energy by FEP, for which a
model system and the additional parameters are set up automatically. These panels, and the
FEP panel for restarting and customizing these jobs, are described in Chapter 4.

In addition to setting up simulations, you can use the general panels to restart a simulation
from a checkpoint file as generated by a previously interrupted simulation.

All jobs run from these panels require a model system to be built first, in the System Builder
panel—see Chapter 2 for details.

Desmond simulations can also be run from the command line—see Chapter 6.

3.1 Overview of the General Desmond Panels

The general Desmond panels have two main sections: Model system, in which the model
system is chosen, and Simulation, in which the parameters for the task are set up (or Minimiza-
tion for minimization tasks). The controls in the second section depend on the panel. Speci-
fying a model system is described in Section 3.2 on page 20, and the various tasks are
described in the subsequent sections.

At the bottom of the panel are the action buttons for the job:

• Start—Start the job. Opens the Start dialog box to set job parameters and submit the job
for execution. See Section 3.10 on page 39 for details. A general description of this dia-
log box and its features is given in Section 2.2 of the Job Control Guide.
• Read—Read a configuration (.cfg) file, to set up the simulation. Opens a file selector in
which you can navigate to the desired file.

Desmond 3.1 User Manual 19

 Chapter 3: Running Simulations from Maestro

• Write—Write the input files for the job but do not start it. Opens a dialog box in which
you can provide the job name, which is used to name the files. The job can be run from
the command line, as described in Chapter 6.
• Reset—Reset the values in the panel to their defaults.

To run a job:
1. Choose the task from the Desmond submenu of the Applications menu.
2. Specify the model system, either by loading it from the Maestro Workspace or importing
it from a file.
3. Adjust the simulation parameters as necessary.
For parameters that are not available in the main panel, open the Advanced Options dialog
box.
4. Click Start.
5. Set the job parameters in the Start dialog box, and click Start to run the job.

To restart a molecular dynamics job:

1. Import the checkpoint file generated by the interrupted simulation.
The default name of this file is jobname.cpt.
When the checkpoint file is imported, the Molecular Dynamics panel enters a read-only
state, in which most of the controls are set by the information read and cannot be
changed.
2. Adjust the total simulation time if necessary.
3. Click Start.
4. Set the job parameters in the Start dialog box, and click Start to run the job.

3.2 Selecting a Model System

In the Model system section, you select the model system that you will use for the simulation.
A valid model system for simulations must contain both the coordinates of the particles and the
force field parameters. In the case of FEP simulations, the model system should also contain
additional FEP-specific parameters. A model system file normally has the .cms extension.

There are two options on the option menu, and the tools in this section depend on which option
you choose.

20 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

• Load from Workspace—Load the model system from the Workspace. The Workspace
must contain a model system that has already been prepared with the System Builder
panel. When you choose this item, the Load button is displayed, which you click to load
the model system from the Workspace.
• Import from file—Import the model system from a file. You can choose to import a model
system file (.cms) or a checkpoint file (.cpt).
If you import a model system file, it must contain a model system that has already been
prepared with the System Builder panel. When the file is imported, a message about the
system is displayed below the option menu.
If you import a checkpoint file, most of the panel controls are unavailable. The purpose of
the checkpoint file is to restart an interrupted simulation, so the parameters of the simula-
tion cannot be altered. You can change the total simulation time, and then start the job.

3.3 Minimizations
Minimization jobs relax the system into a local energy minimum. The model system is mini-
mized using a hybrid method of the steepest decent and the limited-memory Broyden-Fletcher-
Goldfarb-Shanno (LBFGS) algorithms. This task is set up in the Minimization panel, which you
open by choosing Applications → Desmond → Minimization in the main window.

There are only two generally useful parameters for this task:

• Maximum iterations—Enter the maximum number of iterations in this text box, or use the
arrow buttons to change the maximum number of iterations in steps of 10.

Figure 3.1. The Minimization panel.

Desmond 3.1 User Manual 21

 Chapter 3: Running Simulations from Maestro

• Convergence Threshold—Enter the convergence threshold for the gradient in units of

kcal mol–1 Å–1.

3.4 Molecular Dynamics Simulations

Molecular dynamics jobs simulate the Newtonian dynamics of the model system, producing a
trajectory of the particle coordinates, velocities, and energies, on which statistical analyses can
be performed to derive properties of interest about the model system. The molecular dynamics
task performs a single MD simulation under the chosen ensemble condition for a given model
system, generating simulation data for post-simulation analyses.

This task is set up in the Molecular Dynamics panel, which you open by choosing Applications
→ Desmond → Molecular Dynamics in the main window.

The controls at the top of the Simulation section allow you to specify the simulation time in ns
and the recording interval in ps for the energy and for the trajectory.

For the recording intervals you can enter a value in ps in the text boxes. The values are rounded
to an integer multiple of the far time step size. This time step size is set in the Integration tab of
the Advanced Options dialog box, in the RESPA integrator section.

Figure 3.2. The Molecular Dynamics panel.

22 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

The controls in the lower part of the Simulation section allow you to choose the ensemble class,
from NVE, NVT, NPT, NPAT, and NPγT, to set the temperature (except for NVE) and the pres-
sure (except for NVE and NVT), and set the surface tension (NPγT only). It also allows you to
relax the model system before performing the simulation, and choose the protocol for the
relaxation.

When Relax model system before simulation is selected, a series of minimizations and short
molecular dynamics simulations are performed to relax the model system before performing
the simulation you set up. This option is selected by default, and a default protocol is used.
Usually, if the model system was just created from the System Builder panel, it needs to be
relaxed; if the model system has been relaxed before, it does not need to be relaxed again. As
an alternative you can run a minimization before the molecular dynamics calculation.

The stages in the default relaxation process for the NPT ensemble are:

1. Minimize with the solute restrained

2. Minimize without restraints
3. Simulate in the NVT ensemble using a Berendsen thermostat with:
• a simulation time of 12ps
• a temperature of 10K
• a fast temperature relaxation constant
• velocity resampling every 1ps
• non-hydrogen solute atoms restrained
4. Simulate in the NPT ensemble using a Berendsen thermostat and a Berendsen barostat
with:
• a simulation time of 12ps
• a temperature of 10K and a pressure of 1 atm
• a fast temperature relaxation constant
• a slow pressure relaxation constant
• velocity resampling every 1ps
• non-hydrogen solute atoms restrained
5. Simulate in the NPT ensemble using a Berendsen thermostat and a Berendsen barostat
with:
• a simulation time of 24ps
• a temperature of 300K and a pressure of 1 atm
• a fast temperature relaxation constant
• a slow pressure relaxation constant
• velocity resampling every 1ps
• non-hydrogen solute atoms restrained

Desmond 3.1 User Manual 23

 Chapter 3: Running Simulations from Maestro

6. Simulate in the NPT ensemble using a Berendsen thermostat and a Berendsen barostat
with:
• a simulation time of 24ps
• a temperature of 300K and a pressure of 1 atm
• a fast temperature relaxation constant
• a normal pressure relaxation constant
This protocol is used for the NPAT and NPγT ensembles as well. A similar protocol is used for
the NVT ensemble.

The protocol files can be found in $SCHRODINGER/mmshare-vversion/data/desmond. The

procedure follows a similar pattern as for NPT. If you are relaxing a protein in a membrane, we
strongly recommend using the desmond_membrane_relax.msj protocol from that directory.
This protocol can take a while to run, as it involves about 1.2 ns of simulation. For more infor-
mation, see Section 3.8 on page 30.

If you want to modify the protocol, you can copy these files and edit them. To make use of the
modified protocol, click Browse and navigate to the new protocol file, which has a .msj exten-
sion. The file name is then listed in the Relaxation protocol text box.

When the simulation finishes, the output structure file (.cms) is written to disk and incorpo-
rated into the Maestro project. In addition, a new trajectory directory is created, called
jobname_trj by default. Checkpoint files are written during the simulation, but are not written
during the relaxation process.

3.5 Simulated Annealing Simulations

Simulated annealing methods use a temperature program rather than a single temperature for
the simulation. A temperature program is a series of times and target temperatures. The
temperature is linearly interpolated as a function of time between adjacent target temperatures
and is controlled by a thermostat.

One of the predominant strategies used is to raise the temperature to a high value one or more
times before relaxing the system to the desired temperature. The goal is to permit the system to
relax out of an initial state that corresponds to a high energy potential minimum into a lower
state by crossing barriers in the free-energy landscape, which is achieved more effectively
during the periods of elevated temperatures. The default temperature program in the Simulated
Annealing panel falls into this category.

Another common use for simulated annealing is to perform an effective minimization with
some relaxation of the system by slowly decreasing the temperature down to very low temper-
atures. This slow cooling should permit at least some shifts from higher energy minima to
lower minima in the energy landscape.

24 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

Figure 3.3. The Simulated Annealing panel.

Simulated annealing calculations can be set up and run from the Simulated Annealing panel,
which you open by choosing Applications → Desmond → Simulated Annealing.

In the Simulation section, you can make settings for the simulated annealing job. The settings
for the simulation time, recording interval, ensemble class and model system relaxation are the
same as for a molecular dynamics simulation, and are described in Section 3.4 on page 22. The
main specific task for simulated annealing is to provide information on the stages by providing
a schedule of reference temperature changes.

The number of stages is set in the Number of stages text box. When a value has been entered,
the table below is adjusted to display text boxes for each stage. The stages are indexed from 0.
For each stage you can specify a starting time in the Time text box, and a starting temperature
in the Temperature text box. The temperature is linearly interpolated between adjacent time
points. The last stage runs until the specified total simulation time.

Desmond 3.1 User Manual 25

 Chapter 3: Running Simulations from Maestro

Figure 3.4. The Replica Exchange panel.

3.6 Replica Exchange Simulations

Many molecular systems have conformations that are separated by significant free energy
barriers. It can be difficult to sample such conformations if they differ by concerted or collec-
tive shifts of many atoms. This commonly occurs in protein-ligand complexes. Random
methods such as Monte Carlo conformational searches have trouble generating such collective
changes, while thermal methods such as molecular dynamics have trouble surmounting the
free energy barriers. Replica exchange simulations [41, 42] attempt to tackle this problem by
allowing the system to spend some time at elevated temperatures in addition to the temperature
of interest. Time spent at elevated temperatures permits the system to evolve faster, in part by
more readily crossing free energy barriers.

26 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

Desmond supports replica exchange simulations in which multiple copies of the system are
simulated at different temperatures, which usually range from the temperature of interest up to
700 K or more. Periodically during the simulation, attempts are made to exchange the coordi-
nates of copies that are at different temperatures. The exchange is processed in a Monte Carlo-
like process: select the systems to attempt to exchange and then use a Metropolis-like criterion
to decide whether to accept the change [41]. The exchange acceptance ratio satisfies the
detailed balance or balance condition so that each replica remains in equilibrium after the
exchange. When many such exchanges are accepted over the course of an extended simulation,
multiple systems with very different histories can visit the temperature of interest. While
systems spend time at higher temperatures they explore conformational space significantly
more rapidly than if they remained at the target temperature. Thus the composite trajectory at
the temperature of interest may contain a more diverse collection of conformations than if
multiple simulations were performed at the target temperature.

As with a regular molecular dynamics simulation each replica may be run on multiple proces-
sors. Since the simulations of each replica proceeds independently between exchange attempts
the additional level of parallelization achieved by running multiple replicas is highly efficient.

Replica exchange simulations can be set up and run from the Replica Exchange panel, which
you open by choosing Applications → Desmond → Replica Exchange in the main window.

In the Simulation section, you can make settings for the replica exchange job. The settings for
the simulation time, recording interval, ensemble class and model system relaxation are the
same as for a molecular dynamics simulation, and are described in Section 3.4 on page 22. The
default ensemble for replica exchange is NPT. If your highest temperature is above 373 K you
might want to change the ensemble to NVT. Exchanges are done between nearest neighbors.
The main specific task is to set the temperature range and temperature profile.

The temperature range is set in the Temperature range text boxes. The defaults are 300 K for
the low temperature and 310 K for the high temperature. You should adjust the low and high
temperature values to suitable values. There are four options for the temperature profile:

• quadratic—Set the temperatures by quadratic interpolation between the minimum and

maximum, with the high temperature at the maximum of the quadratic curve.
• linear—Set the temperatures by linear interpolation between the maximum and the mini-
mum.
• auto—Automatically set the number of replicas and the temperature schedule so that the
acceptance ratio is approximately 30%.
• manual—Set the temperatures manually, by editing the temperatures in the replica table.
When you select this option the table becomes editable.

Desmond 3.1 User Manual 27

 Chapter 3: Running Simulations from Maestro

Information on the temperatures is displayed in the replica table. You can edit the temperatures
by selecting manual for the temperature profile. Some guidance on selecting temperatures is
available in Ref. 43. Setting up the temperatures and the number of replicas for a meaningful
simulation can be difficult. For assistance with this task, contact help@schrodinger.com.

The replica exchange simulation produces one trajectory for each replica, labeled
jobname_replicanum_trj, where num is the index of the replica, starting from 0, and corre-
sponds to the replica index in the replica table. You can display a temperature versus time plot
of the replicas and the exchanges that were made—see Section 5.4 on page 63.

3.7 Metadynamics Simulations

Metadynamics is a technique in which the potential for one or more chosen variables (“collec-
tive variables”) is modified by periodically adding a repulsive potential of Gaussian shape at
the location given by particular values of the variables. These repulsive Gaussians eventually
fill up the well that is being sampled, and force the calculation to sample elsewhere. For prop-
erly converged simulations, the sum of the Gaussians and the free-energy surface (FES)
becomes flat, and therefore the sum of Gaussians is the negative image of the FES.

The parameters that control the accuracy of the simulation are the height and width of the
Gaussian potential and the interval at which the Gaussians are added. The accuracy of the
results is inversely proportional to the height for a given interval for addition of the Gaussian.
However, if smaller heights are used, the simulation takes longer. The accuracy of the results
increases as the time interval increases, but the simulation time also increases. The accuracy is
not very sensitive to the ratio of the height to the interval, but smaller values of this ratio
increase the accuracy. The width of the Gaussian should be roughly 1/4 to 1/3 of the average
fluctuations of the collective variable during a free MD run.

In the Simulation section, you can make settings for the metadynamics job. The settings for the
simulation time, recording interval, ensemble class and model system relaxation are the same
as for a molecular dynamics simulation, and are described in Section 3.4 on page 22. The
default ensemble for metadynamics is NPT. The main specific task for metadynamics is
defining the collective variables and the Gaussian.

The height of the Gaussian and the interval at which it is added are set in the Height and
Interval text boxes, under Metadynamics parameters.

To set up a collective variable, first select an option for the type of variable: Distance, Angle, or
Dihedral. (More flexible definitions can only be made from the command line.) Select Pick
atoms to pick the atoms in the Workspace. The atom numbers are listed in the AtomN text
boxes. You can set the width of the Gaussian in the Width text box. The default values are 0.05
Å for distances, 0.03 rad (1.8°) for angles, and 0.05 rad (3°) for dihedrals.

28 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

Figure 3.5. The Metadynamics panel.

For distance variables, you can place a barrier (“wall”) at a given distance in the Wall at text
box. Introducing a wall prevents the system from moving too far in the direction defined by the
collective variable.

Once you have set up the variable, you can add it by clicking Add. The variable is added to the
table, and is used in the simulation. To remove a variable from the simulation, select it in the
table and click Remove. Setting up more than two collective variables can create sampling
problems.

Desmond 3.1 User Manual 29

 Chapter 3: Running Simulations from Maestro

More flexibility in the collective variables is available from the command line, via multisim.
You can use ASL expressions to define a group of atoms whose center of mass is used as the
variable, for example.

3.8 Simulations on Systems with Membranes

Simulations of systems that contain membranes require some special consideration. This is
because nearly all current all-atom membrane potential models in existence do not, on their
own, maintain the appropriate surface areas on the time scale of tens of ns in simulations of
pure membranes. If non-lipid components make up a significant fraction of the membrane
region (such as a protein in a relatively small amount of lipid), this issue is much less
pronounced and many not require special treatment. In this case the semi-isotropic NPT
ensemble may work well. However, if the simulated membrane is pure or only contains a small
solute (e.g. ligand-sized) the following practical approach may be useful.

When a solute is placed in a pure membrane, some lipids are usually removed to make room
for the new molecule during the system building process. As a result, adjustment of the surface
area of the solute + membrane system is often needed. This can usually be done using a fairly
short semi-isotropic simulation of up to about 0.5 ns. When simulating beyond that time range
it is recommended to switch to either a constant surface area, constant normal pressure simula-
tion (NPAT) or a constant surface tension simulation (NPγT). If the latter is selected we
suggest using a surface tension of 2000 bar/Å for DPPC and 4000 bar/Å for POPE and POPC.
We recommend examining the results of all membrane simulations carefully.

It can be difficult to relax freshly built protein-membrane systems. In particular, penetration of

the water between the protein than the lipids can be problematic and require very lengthy simu-
lations to correct. A relaxation protocol that should reduce or eliminate such problems is avail-
able by running the script relax_membrane.py from the command line.

To use the membrane relaxation protocol:

1. Save your newly built protein-membrane system in a CMS file (referred to here as pro-
tein-membrane.cms.
2. Run the script to prepare the necessary input files:
$SCHRODINGER/run relax_membrane.py -i protein-membrane.cms
-t temperature -j protein-membrane
3. Run the membrane relaxation protocol using the command
$SCHRODINGER/utilities/multisim -JOBNAME protein-membrane
-HOST myhost -mode umbrella -cpu cpus -i protein-membrane-in.cms
-m protein-membrane.msj -o protein-membrane-out.cms

30 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

This process may take hours to days since it equilibrates the system in stages for about 1.2 ns.
The file protein-membrane-out.cms should be reasonably well equilibrated and can be used
as input for the production simulation for your study.

3.9 Setting Options for Desmond Simulations

The default settings used in the Desmond panels were selected to produce good results in the
majority of cases. At times, you may want greater control over the parameters of the calcula-
tion. The Advanced Options dialog box provides access to advanced options for control of the
simulation or the minimization, such as the frequency of data output, integration time step
sizes, thermostat and barostat parameters, restraints, cutoff radii, and so on.

To open the Advanced Options dialog box, click Advanced Options in the Desmond panel that
you have open. This panel has several tabs, which are described in the following subsections.

• Integration tab
• Ensemble tab
• Minimization tab
• Interaction tab
• Restraints tab
• Output tab
• Misc tab

The selection of tabs that is displayed depends on the task. For minimizations, only the Minimi-
zation, Interaction, Restraints, and Misc tabs are displayed. For MD simulations all but the Mini-
mization tab are displayed.

The settings in this dialog box and the settings in the Desmond panels are not entirely indepen-
dent, and can affect each other. For example, changing the far time step can affect the values of
the recording periods in the panel, because the latter are automatically rounded by the far time
step. As another example, changing the temperature or pressure in the Desmond panel updates
the temperature or pressure parameters in the dialog box. Changes in the Desmond panel take
effect immediately and update parameters in the dialog box, whereas changes made in the
dialog box only take effect when you click OK or Apply.

If you want to clear changes that have not been committed with the Apply button, click Reset.
Any changes made since the last set of changes were committed are discarded, and the values
in the dialog box are reset to the last set committed.

Desmond 3.1 User Manual 31

 Chapter 3: Running Simulations from Maestro

Figure 3.6. The Integration tab of the Advanced Options dialog box.

3.9.1 The Integration Tab

In this tab, you can set parameters for the integration algorithm. The tab has a single section,
RESPA integrator.

RESPA integrator section

Specify the time steps in fs for bonded, near, and far, by entering values in the text boxes, or
using the arrow buttons to change the value in increments of the bonded time step. Because the
bonded, near, and far time steps must maintain a certain ratio, when a new value is set for the
bonded time step, the other two time steps are automatically updated according the current
ratio. Changing the near or far time steps adjusts this ratio.

Selecting Set time step automatically based on constraint setting couples the RESPA time step
settings with those in the Constraint section. The time steps will be automatically set based on
the settings in the Constraint section, and are not available for editing.

32 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

Figure 3.7. The Ensemble tab of the Advanced Options dialog box.

3.9.2 The Ensemble Tab

In this tab you can set the thermostat method and the barostat method and adjust the settings
for these methods. The thermostat method and the barostat method are coupled: the choice you
make from the Thermostat method option menu changes the selection from the Barostat
method option menu, and vice versa. The exception is that you can choose None for the baro-
stat method for any of the thermostat methods.

The Thermostat method option menu offers four choices: Nose-Hoover chain, Berendsen,
Langevin, and None.

Although in most circumstances, only one thermostat group is needed, you can specify
multiple thermostat groups by entering the number of groups in the Number of groups text box
and supplying information on these groups in the Thermostat group settings table. The
maximum number of groups is 8. The selection of atoms that is in each group can be set up in
the Misc tab, by defining multiple groups named thermostat, with the group numbers corre-
sponding to the entries in the Values column. Any atoms not explicitly added to a group are
automatically assigned to group 0, the default group. This means that you do not need to define
a group if you only want to use one thermostat, and that you only need to define groups for the
extra thermostats, starting from thermostat 1.

The Thermostat group settings table provides text boxes for making settings for each thermo-
stat group. The settings that can be made are:

Desmond 3.1 User Manual 33

 Chapter 3: Running Simulations from Maestro

• Temperature (K)
• Relaxation time (ps)
The Barostat method option menu also offers four choices: Martyna-Tobias-Klein, Berendsen,
Langevin, and None. For each of these methods you can set the relaxation time (ps) in the
Relaxation time text box and choose a coupling style from the Coupling style option menu. The
coupling style choices are Isotropic, Semi-isotropic, Anisotropic, and Constant area. The pres-
sure used is atmospheric pressure.

3.9.3 The Minimization Tab

In this tab you can set parameters for the minimization, and also specify the output file. Mini-
mization is performed with the LBFGS method, with an optional steepest descent initial phase.

The Minimizer section provides the following controls:

• LBFGS vectors—Specify the number of history vectors used by the LBFGS minimizer
for the update of the Hessian. The maximum is 6.
• Minimum SD steps—Specify the minimum number of steepest descent steps to be used
before switching to the LBFGS minimizer.

In the Output section you can specify the name of the structure output file. You can use
$JOBNAME as a variable representing the job name that you will set when you start the job or
write out the input files.

Figure 3.8. The Minimization tab of the Advanced Options dialog box.

34 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

Figure 3.9. The Interaction tab of the Advanced Options dialog box.

3.9.4 The Interaction Tab

In this tab you can specify how the short-range and long-range Coulombic interactions are
handled.

To define the short-range region, choose a method from the Short range method option menu.
The controls below this menu depend on the method chosen, which can be one of the
following:

• Cutoff—Enter a value in the Cutoff radius text box. The default is 9.0 Å.
• Potential tapering—Specify the range in angstroms over which the potential is tapered off
in the two Tapering range text boxes.

There are two choices for handling the long-range Coulombic interactions:

• Smooth particle mesh Ewald—use the smooth particle mesh Ewald method. This method
requires a tolerance to be set in the Ewald tolerance text box. This tolerance affects the
accuracy of the long-range Coulombic interactions. The smaller the tolerance is, the more
accurate the computation of the long-range Coulombic interactions is, but the simulation
will be correspondingly slower.
• None—use the unmodified Coulomb interaction.

Desmond 3.1 User Manual 35

 Chapter 3: Running Simulations from Maestro

Figure 3.10. The Restraints tab of the Advanced Options dialog box.

3.9.5 The Restraints Tab

In this tab you can specify restraints on atom positions. A restraint is defined by a set of atoms
and a force constant. The restraints are listed in the restraints table. The Atoms column is filled
in automatically when you click Select and use the Atom Selection dialog box to specify the
atoms. You must enter the force constant in the table manually, by editing the table cell.

To manage the restraints, you can use the buttons beside the table:

• Select—Opens the Atom Selection dialog box to specify the atoms for the selected
restraint. Only available if a single row is selected in the table.
• Add—Adds a row to the restraints table so that a new restraint can be defined.
• Delete—Deletes the selected restraints.
• Reset—Resets the table to its default state.

36 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

Figure 3.11. The Output tab of the Advanced Options dialog box.

3.9.6 The Output Tab

In this tab you can set names for various output files, and set the times for which recording in
these files begins, and the update frequency. For each file there is a starting time text box, in
which you can enter the starting time for recording of information to this file or adjust the
starting time in increments of 50 times the far time step (except for the checkpoint file, where
the increment is 5000 times the far time step). Some files also have an interval text box, in
which you can enter or adjust the recording interval (in the same increments as the starting
time). The files are listed below, with any extra controls. The intervals for the energy sequence
file and for the trajectory can be set in the Desmond panel.

• Energy sequence file—This file contains a sequence of various energies of the system.
• Trajectory directory—This directory is used by Desmond to periodically write out files
that record coordinates and velocities (optional) of all particles in the system at a particu-
lar point in the simulation. You can provide a title for the trajectory, and you can select
Record velocities if you want the velocities to be recorded along with the coordinates.

To ensure that associated solutes appear together in the trajectory rather than on opposite
sides of the simulation box, you can select Glue close solute molecules together. This
option only affects the way the trajectory is displayed.
• Checkpoint file—This file contains information that can be used to restart an interrupted
simulation. Checkpoint files permit bitwise continuation of simulations, so they can be
large, and should be saved infrequently if at all.

Desmond 3.1 User Manual 37

 Chapter 3: Running Simulations from Maestro

Figure 3.12. The Misc tab of the Advanced Options dialog box.

3.9.7 The Misc Tab

This tab provides access to various options that do not fit into the other categories. In this tab
you can specify when the velocities are randomized, and specify atom groups for special treat-
ment in the simulation.

You can set the starting time and the interval at which velocities are randomized in the
Randomize velocities section. By default velocities are randomized at the beginning of a calcu-
lation, and not randomized again. For some kinds of simulations (e.g. in the gas phase), peri-
odic randomization of velocities can improve the sampling.

In the Atom group section you can specify atom groups. Atom groups are used for various
special treatments of atoms in the simulation.

You can define multiple groups with the same name but a different index by setting the index in
the Value column. If you use multiple thermostats, for example, you can define the atoms in
each thermostat group by naming the group thermostat and setting the index to the thermo-
stat group number in the Ensemble tab. Group 0 is the default group, to which all unassigned
atoms automatically belong.

The atom groups are listed in the table. The Atoms column is filled in automatically when you
click Select and use the Atom Selection dialog box to specify the atoms. Otherwise you can
edit this column to specify the ASL expression for the atoms in the group. The number of
atoms defined by the ASL expression is shown in the No. of Atoms column.

38 Schrödinger Suite 2012 Update 2

 Chapter 3: Running Simulations from Maestro

Beside the table are the following buttons:

• Select—Opens the Atom Selection dialog box to specify the atoms for the selected group.
Only available if a single row is selected in the table.
• Add—Adds an atom group. Clicking this button adds a row to the table.
• Delete—Deletes the selected groups. This operation can only be done on user-defined
groups.
• Reset—Resets the table to its default state.

3.10 Running a Simulation Job

Once you have finished making settings, you can click Start to set up the job parameters and
run the job, or you can click Write to write out the input files and run the job from the
command line. For information on running from the command line, see Chapter 6.

When you click Start, the Start dialog box opens. The common features of this dialog box,
such as the Output section, the Job name, and CPUs text boxes, and the Host option menu, are
described in Section 2.2 of the Job Control Guide. These features allow you to direct the job to
the appropriate host with the desired number of processors, and decide how to incorporate the
output into the Maestro project.

For Desmond jobs, the choice of the number of processors has some special requirements.
Desmond uses a Cartesian domain decomposition of the simulation for efficient processing
This decomposition is described in the Desmond User’s Guide, provided by D. E. Shaw
Research. The details of the decomposition can be displayed in the Start dialog box by clicking
the Details button. You can control the decomposition by entering values in the x, y, and z text
boxes. The values you enter must be a power of 2, 3, or 5, or products of these powers. The
decomposition is done automatically if you enter a number in the CPUs text box. However, if
you enter a number that is not a suitable product of powers, the actual number of processors
used is the largest product of powers that is smaller than the number you entered. The number
of processors actually used is displayed in the details section. To hide the details, click the
Details button again.

Before you can run Desmond jobs in parallel, you must configure the hosts file and any queues
that you want to use. Details are given in Chapter 7 of the Installation Guide.

When the job is submitted to a queueing system, the processors requested are all allocated to
the master job, and the subjobs are run directly on these processors, without being submitted to
the queueing system. (This is called “umbrella” mode—see Section A.1.2 on page 106.)

Desmond 3.1 User Manual 39

 Chapter 3: Running Simulations from Maestro

Figure 3.13. The Start dialog box.

40 Schrödinger Suite 2012 Update 2

 Desmond User Manual

Chapter 4

Chapter 4: Running FEP Simulations

Maestro provides panels for quickly setting up FEP simulations for certain common scenarios:

• Mutation of ligand functional groups: a group attached to a ligand by a single bond is

mutated to some other group.
• Aromatic ring atom substitution: an atom (with its attached hydrogens) is mutated into
another atom, usually a heteroatom.
• Protein residue mutation: the side chain of a protein residue is mutated into that of
another residue.

In these three scenarios, the relative binding free energy of the ligand can be calculated. For the
first two, the solvation free energy of the ligand relative to the original ligand can be calculated.

Total (absolute) solvation free energies can also be calculated for a ligand or a molecule, which
is done by annihilating the chosen molecule in the FEP simulation.

Free energy perturbation calculations are resource-intensive. While it is possible with the FEP
panels to set up multiple mutations in a single calculation, we recommend that only one muta-
tion be performed per calculation due to the resource requirements.

4.1 FEP Panels

The FEP panels are designed to make setting up FEP jobs as easy as possible: most of the
computational details, such as setting the force field parameters, solvation, relaxation of the
system, and simulation time, are taken care of automatically. These predetermined parameters
for the FEP calculations should work for most systems, but you can also write out the input
files and then modify the simulation parameters in the FEP panel—see Section 4.8 on page 53.
(Note that checkpoint files are not written out for the relaxation stages.)

The panels for these calculation types have a common design. Each panel has a Define Pertur-
bation tab, in which you set up the systems to be simulated, and a Plan Calculation tab, in
which you choose the environment (complex, pure solvent, vacuum) for the ligand and choose
the FEP protocol, which includes the ensemble. Information on setting up the system is
contained in the next few sections, followed by a section describing the Plan Calculation tab.

Desmond 3.1 User Manual 41

 Chapter 4: Running FEP Simulations

The panels have action buttons at the bottom:

• Start—Opens the Start dialog box, in which you can set the job parameters such as job
name, host, number of CPUs, and so on, and then start the job. The general layout of this
dialog box is described in Section 2.2 of the Job Control Guide. The Desmond-specific
features of the Start dialog box and the issues in choosing the number of CPUs are
described in Section 3.10 on page 39.
• Write—Writes the input files for the job but does not start the job. Opens a dialog box in
which you can specify a job name, which is used to name the files. This capability is use-
ful if you want to change the FEP protocol.
• Reset—Resets the entire panel to its initial state. This includes clearing the Workspace. If
you want to use the same structures, you will have to redisplay them before proceeding.

4.2 Ligand Functional Group Mutation

In ligand mutation, a group on the ligand molecule, called the substitution group, is replaced
with selected fragments. These fragments are chosen from a list of small, predefined frag-
ments. The fragments can be adjusted to reduce steric clashes, and can be modified by using
the tools on the Build toolbar or in the Build panel.

The FEP calculations for ligand mutation are set up in the Ligand Functional Group Mutation by
FEP panel, which you open by choosing one of the following:

• Applications → Desmond → Ligand Functional Group Mutation by FEP

• Tasks → Binding Energy Estimation → Ligand Functional Group Mutation → Molecular
Dynamics FEP

To run a ligand mutation job:

1. Include in the Maestro Workspace the system that contains the ligand to be mutated, the
receptor protein (optional), and any other key molecules of the system.
This step can be performed prior to opening the panel. You can use an existing model sys-
tem that may, for example, include a membrane. If you do, note that the SPC water model
is used in the FEP simulations, regardless of the choice made for your model system.
2. Select Pick the attachment bond, then pick the attachment bond in the Workspace, at the
end toward the substitution group.
This bond is the bond at which the substitution will take place. When the bond is picked,
a green arrow is drawn over the bond, pointing to the substitution group, and the ligand
core is colored with green carbons. If you clicked on the wrong end of the bond, click
again to correct the error. Once focus is returned to the panel, this option is deselected.

42 Schrödinger Suite 2012 Update 2

 Chapter 4: Running FEP Simulations

Figure 4.1. The Ligand Functional Group Mutation by FEP panel, Define Perturbation tab.

3. In the Fragment library table, select the fragments that you want to use to replace the orig-
inal substitution group, by selecting their table rows.
Use shift-click and control-click to select multiple table rows. You can view the mutated
ligand by clicking in the View column.
Since each FEP mutation requires a large amount of CPU time and space, we recommend
performing only one mutation in a given calculation.
4. (Optional) Display the mutated ligand and make adjustments to the fragment.
The adjusted fragment is used as the starting point in the FEP simulations. It is possible
that the fragment may have clashes with the protein in its default orientation or be in an in
appropriate conformation. The barriers between local minima in the potential energy sur-
face can be large, and thus take a long time to cross during a simulation. Adjusting the
conformation can reduce clashes and help the simulation to sample the appropriate
conformations.

Desmond 3.1 User Manual 43

 Chapter 4: Running FEP Simulations

Once the fragment is displayed, you can reorient the fragment with the adjustment tools,
which are available from the Adjust toolbar button:

or with the local transformation tools, which are available from the Transform toolbar
button:

For more information on these tools, see Section 5.9 of the Maestro User Manual.
You can also add to the fragment using the Build toolbar or the Build panel—for example,
replacing a hydrogen with a methyl or a hydroxyl group. Since the quality of the results
goes down as you increase the size of the fragment, it is not advisable to add too much to
the fragment.
You can revert to the original fragment by clicking the button in the Reset column. You
should not make modifications to the ligand core or other molecules in the system.
5. In the Plan Calculation tab, select the calculation types:
• For binding free energy calculations, select In complex and In pure solvent.
• For solvation free energy calculations, select In pure solvent and In vacuum.
6. Choose the FEP protocol you want to use for each calculation type from the FEP protocol
option menu.
You should choose the same ensemble type for both the complex and pure solvent
calculations.
7. Click Start, set the job parameters in the Start dialog box, and click Start in the dialog
box to run the job.

4.3 Ring Atom Mutation

Ring atom mutation changes a selected ring atom to a different atom type while preserving the
hybridization of the atom. To open the Ring Atom Mutation by FEP panel, choose one of the
following:

• Applications → Desmond → Ring Atom Mutation by FEP

• Tasks → Binding Energy Estimation → Ligand Ring Atom Mutation → Molecular Dynam-
ics FEP

44 Schrödinger Suite 2012 Update 2

 Chapter 4: Running FEP Simulations

Figure 4.2. The Ring Atom Mutation by FEP panel, Define Perturbation tab.

To run a ring atom mutation job:

1. Include in the Maestro Workspace the system that contains the ligand to be mutated, the
receptor protein (optional), and any other key molecules of the system.
This step can be performed prior to opening the panel. You can use an existing model sys-
tem that may, for example, include a membrane. If you do, note that the SPC water model
is used in the FEP simulations, regardless of the choice made for your model system.
2. Select Pick the ring atom, then pick the ligand atom to be mutated in the Workspace.
When the atom is picked, it is rendered in ball-and-stick, and the ligand is colored with
green carbons. The Mutation targets table is displayed. Once focus is returned to the
panel, this option is deselected.
3. In the Mutation targets table, select the targets that you want to use to replace the original
ring atom by selecting their table rows.

Desmond 3.1 User Manual 45

 Chapter 4: Running FEP Simulations

Use shift-click and control-click to select multiple table rows. You can view the mutated
ligand by clicking in the View column. If you want to add a functional group as well as
mutate the ring atom, you can use the build tools to add to the mutated structure.
4. In the Plan Calculation tab, select the calculation types:
• For binding free energy calculations, select In complex and In pure solvent.
• For solvation free energy calculations, select In pure solvent and In vacuum.
5. Choose the FEP protocol you want to use for each calculation type from the FEP protocol
option menu.
6. Click Start, set the job parameters in the Start dialog box, and click Start in the dialog
box to run the job.

4.4 Protein Residue Mutation

In protein residue mutation, a protein residue is mutated into another residue in the presence of
a ligand, and the change in the ligand binding free energy due to the mutation is calculated.

The FEP calculations for protein residue mutation are set up in the Protein Residue Mutation by
FEP panel, which you open by choosing one of the following:

• Applications → Desmond → Protein Residue Mutation by FEP

• Tasks → Binding Energy Estimation → Protein Residue Mutation

To run a protein residue mutation job:

1. Include in the Maestro Workspace the system that contains the ligand to be mutated, the
receptor protein, and any other key molecules of the system.
This step can be performed prior to opening the panel.
2. Select Pick atom to identify ligand, then pick a ligand atom in the Workspace.
3. In the Define Mutations section, select Pick residue.
4. Pick a protein residue in the Workspace.
The residue is added to the table as a mutation site.
5. In the Substitutes column, double-click in the table cell and choose the residues from the
list that you want to use to replace the original.
6. Repeat Step 4 and Step 5 for as many mutations as you want to perform.
You can pick the same residue multiple times to mutate it to different residues. If you pick
different residues, each residue that you pick is mutated.

46 Schrödinger Suite 2012 Update 2

 Chapter 4: Running FEP Simulations

Figure 4.3. The Protein Residue Mutation by FEP panel, Define Perturbation tab.

7. Click Generate Mutants.

8. The table is filled with the mutations that you have defined. For each of the original resi-
dues that you picked, all possible combinations of mutations are added to the table. For
example, if you picked two different residues, and for each you specified three substi-
tutes, the table would list nine mutations, corresponding to each possible combination.
9. In the mutants table, select the mutations that you want to use, by selecting their table
rows.
Use shift-click and control-click to select multiple table rows. You can view the mutated
protein by clicking in the View column.

Desmond 3.1 User Manual 47

 Chapter 4: Running FEP Simulations

Once the residue is displayed, you can reorient the side chain with the adjustment tools,
which are available from the Adjust toolbar button:

or with the local transformation tools, which are available from the Transform toolbar
button:

For more information on these tools, see Section 5.9 of the Maestro User Manual.
You can also modify the residue using the Build toolbar or the Build panel—for example
to replace sulfur with selenium, or to phosphorylate the residue. Since the quality of the
results goes down as you increase the size of the perturbation, it is not advisable to add
too much to the residue.
You can revert to the original residue orientation and structure by clicking the button in
the Reset column. You should not make modifications to the ligand core or other mole-
cules in the system.
10. In the Plan Calculation tab, choose the FEP protocol you want to use for each calculation
type from the FEP protocol option menu.
You should choose the same ensemble type for both the complex and pure solvent
calculations.
11. Click Start, set the job parameters in the Start dialog box, and click Start in the dialog
box to run the job.

For calculations that are not defined by a simple protein, ligand, and water model (such as
membrane proteins or calculations in salt), you can use existing model systems for input. To do
this you should prepare two model systems, one with and one without the ligand. You should
run the In complex calculation only, for both systems, to ensure that the model system is used
as is for both perturbations. The relative binding free energy can be calculated from the differ-
ence between the two free-energy differences.

4.5 Total Solvation Free Energy Calculation

Calculating the absolute solvation free energy for a molecule involves running an “annihila-
tion” FEP simulation, in which the selected molecule is removed from the system. See Refer-
ence 45 for a study of absolute solvation free energies.

48 Schrödinger Suite 2012 Update 2

 Chapter 4: Running FEP Simulations

Figure 4.4. The Total Free Energy by FEP panel, Define Perturbation tab.

To run an annihilation job:

1. Include in the Maestro Workspace the system that contains the molecule to be annihi-
lated.
This step can be performed prior to opening the panel.
2. Select Pick the molecule, then pick the molecule in the Workspace.
When the molecule is picked, it is colored with green carbons. Once focus is returned to
the panel, this option is deselected.
3. In the Plan Calculation tab, ensure that In pure solvent is selected.
4. Choose the FEP protocol you want to use for each calculation type from the FEP protocol
option menu and make any related settings.
The default simulation time is 2 ns, which is recommended for accurate results.
5. Click Start, set the job parameters in the Start dialog box, and click Start in the dialog
box to run the job.

4.6 Selecting the Environment and FEP Protocol

The selection of the ligand environment and FEP protocol is made in the Plan Calculation tab.
This section explains the choices, which were referred to in the preceding sections.

The FEP simulations can be performed in three environments: in the protein-ligand complex
(with solvent), in pure solvent, and in vacuum. This permits the calculation of binding free
energies, by selecting In complex and In pure solvent, and of solvation free energies, by

Desmond 3.1 User Manual 49

 Chapter 4: Running FEP Simulations

selecting In pure solvent and In vacuum. Selecting just one of the environments does not on its
own produce a meaningful result.

The FEP protocol defines in detail how system is solvated, relaxed, and simulated. Four main
protocols are provided in the FEP protocol option menu for the complex and pure solvent: two
ensembles, Desmond NPT and Desmond NVT, with two options for the relaxation part (stan-
dard, or quick relaxation). The protocol for the simulation is the same for both ensembles (NVT
and NPT). For calculations in vacuum, there is the choice of the standard or quick relaxation.
You can set the temperature for all the simulations in the Temperature text box.

In addition you can read in a protocol by choosing User defined from the FEP protocol option
menu. A file name text box and Browse button is displayed. You can enter the name of the
protocol file (which has a .msj extension) in the text box, or navigate to it in the file selector
that opens when you click Browse.

Figure 4.5. The Plan Calculation tab.

50 Schrödinger Suite 2012 Update 2

 Chapter 4: Running FEP Simulations

The default protocol is as follows. The original system is solvated by adding SPC water mole-
cules with a buffer distance of 5 Å for complexes and 10 Å for pure solvent. The vacuum simu-
lation uses a buffer distance of 100 Å. The system goes through a relaxation process that
includes two minimizations followed by 4 short molecular dynamics simulations. The produc-
tion simulation is run for 0.6 ns for each lambda window, and 12 windows are used for each
perturbation. You can change the buffer size and the production simulation time for each envi-
ronment in the Buffer size and Production simulation time text boxes. If you are using an
existing model system, the buffer size is ignored for the complex, but is used for the pure
solvent.

Lambda hopping is a form of Hamiltonian replica exchange in which exchanges of coordinates

for systems simulated in adjacent FEP windows are attempted. For mutations involving confor-
mational flexibility, lambda hopping can improve performance. To use this method, select Use
lambda hopping to enhance sampling.

After the production simulations, the results are collected and analyzed using the Bennett
method. The final result (free energy) for each perturbation is recorded as an entry-level prop-
erty in the final Maestro output file.

4.7 FEP Results

The results of FEP calculations usually need some further processing to produce the quantities
of interest. The most common of these are discussed in the sections below.

4.7.1 Relative Free-Energy Differences

The estimated changes in free energy for FEP calculations are recorded as structure-level prop-
erties in the mutated structures (rather than in the original structures) in the output structure
file. For ligand functional group mutation, protein residue mutation and ring atom mutation,
these values are displayed in Maestro as dG jobname, and recorded in the file as
s_des_dG_jobname, where jobname is the name for the multisim job and usually ends in
_vacuum, _solvent or _complex for the corresponding environments (vacuum, pure
solvent, and complex). These values are strings rather than real numbers because they include
the estimated uncertainty. To obtain a relative solvation free energy, subtract the vacuum result
from the solvent result. Similarly, relative binding free energies are calculated by subtracting
the solvent result from the complex result.

You can use the script calculate_ddg.py to take the difference between the values in two
different environments to provide a ΔΔG value. To run this script, use the following command:

$SCHRODINGER/run -FROM desmond calculate_ddg.py jobname1-out.mae

jobname2-out.mae [output-file.mae]

Desmond 3.1 User Manual 51

 Chapter 4: Running FEP Simulations

where jobname1-out.mae and jobname2-out.mae are the Maestro output structure files
from the two FEP calculations. The output file output-file.mae is optional; the default name is
ddG.mae. This file is in Maestro format and contains the ΔG value from both input files. The
final ΔΔG value is recorded for each mutant as an entry property s_des_ddG_jobname1-
jobname2, which is displayed in the Project Table as ddG jobname1-jobname2. The atom coor-
dinates in the output file are copied from the first input structure file.

4.7.2 Total Solvation Free Energies

The total absolute free energies calculated by Desmond are recorded in the output structure file
as structure-level properties for the solute structures. The property names are
s_des_dg_jobname_transfer and s_des_dg_jobname_solvation. The former, the
transfer free energy, uses the same concentration in both the vacuum and solution phases while
the latter, the solvation free energy, uses standard state concentrations in the two phases (1 bar
in vacuum and 1 Molal or 1 mole of solute per kg of solvent in solution).

When charged molecules are deleted or created in absolute free energy calculations, finite size
effects can be significant, particularly for the pure solvent FEP calculations. A script,
calculate_correction.py, has been provided to provide a correction to the free energy
once the FEP calculation is complete. To run this script, use the following command:

$SCHRODINGER/run -FROM desmond calculate_correction.py jobname [data-dir]

where jobname is the name of the λ=0 simulation job for the original FEP calculation (e.g.,
myfepjob_complex_4_lambda0). This program assumes the input and output file names are
derived from jobname. Specifically, this program looks for the following files and directories:
jobname-in.cms, jobname-in.cfg, and jobname_trj. jobname-in.cms and jobname-
in.cfg must be the input files of the simulation, and jobname_trj is the trajectory directory
produced during the simulation. The optional argument data_dir can be used to designate the
name for the directory in which to look for the for the data files. If this argument is missing, the
default directory name is jobname.

This program reviews the trajectory and calculates the correction. The final correction is
printed to standard output. During the process, the program writes out several files:
jobname_correct1.cfg, jobname_correct1.log, and jobname_correct1.dat.
Depending on your system, it may also write out the analogous files:
jobname_correct2.cfg, jobname_correct2.log, and jobname_correct2.dat. You
need not be concerned with the content of these files.

52 Schrödinger Suite 2012 Update 2

 Chapter 4: Running FEP Simulations

4.7.3 FEP Analysis

Judging the accuracy and convergence of free energy calculations is difficult, particularly if no
additional data is available. Desmond FEP jobs generate additional information for relative
free energy and total solvation free energy calculations, which can help to identify accuracy
and convergence problems. This information includes plots of the distributions of the perturba-
tion energies for mutating between adjacent windows, the time evolution of the overall free
energy and the corresponding estimates for the individual perturbations. The information is
available in the tar file of the fep_analysis stage for the calculation for absolute solvation
free energies and each of the two FEP calculations needed to generate a ΔΔG value for relative
free energy calculations. These tar files are usually the highest numbered ones.

For instance, for ligand functional group mutations in the presence of a protein these files
would normally be called jobname_solvent_12.tgz and jobname_complex_12.tgz.
These files can be extracted using the command:

tar -xzvf tarfilename

Inside the extracted directory there is an HTML file, which you can open in a web browser to
view the various plots.

4.8 Customizing and Restarting FEP Simulations

FEP jobs perform a series of MD simulations for the calculation of the free energy difference
between two systems. These simulations are set up in the specialized FEP panels described
earlier in this chapter. The FEP panel allows you to adjust FEP protocols set up in these panels,
and to rerun lambda windows for FEP jobs that failed. To open the FEP panel, choose Applica-
tions → Desmond → FEP in the main window.

The settings for the simulation time, recording interval, ensemble class, temperature, pressure,
surface tension, and model system relaxation are the same as for a molecular dynamics simula-
tion, and are described in Section 3.4 on page 22.

There are two options for the type of FEP that can be performed, that correspond to the types
of FEP that are set up with the specialized FEP panels:

• Mutation—Mutate one set of atoms into another. The atom set may be a molecule or a
fragment. Select this option for ligand functional group mutation or ring atom mutation.
• Total free energy—Select this option for total free energy simulations, in which a set of
atoms is annihilated by the perturbation.

Desmond 3.1 User Manual 53

 Chapter 4: Running FEP Simulations

Figure 4.6. The FEP panel.

The number of lambda windows you want to use can be entered in the Number of windows text
box. Each window is listed in the table below, with values for each of the six lambda parame-
ters, Van der Waals, Charge and Bonding, for the two systems, labeled A and B.

You can set the lambda values automatically, which is the default, or you can deselect Set
lambda values automatically and edit the values for each window and each type of lambda. The
lambda values must be within the range [0, 1]. The background color of the cell is changed to
pink if the value is out of range, to warn that it must be set within the range.

54 Schrödinger Suite 2012 Update 2

 Chapter 4: Running FEP Simulations

The windows can be selected or deselected for the simulation using the buttons in the Window
index row. For a new FEP job, all windows should normally be selected for simulation. If a
simulation for a particular lambda window needs to be re-run, select only that window.

For mutation FEP, you can also use Hamiltonian replica exchange to enhance the sampling, by
selecting Use lambda hopping to enhance sampling.

Desmond 3.1 User Manual 55

56 Schrödinger Suite 2012 Update 2
 Desmond User Manual

Chapter 5

Chapter 5: Analyzing Simulations

5.1 Viewing Trajectories

You can play through trajectories, examine individual frames, and export trajectory data in a
variety of forms, in the Trajectory panel. To open the Trajectory panel, click the T button in the
Title column of the Project Table for the entry whose trajectory you want to view.

If you have entries in the Workspace, a panel opens asking if you want to keep them in the
Workspace while the trajectory is played, or to remove them. Keeping the entres in the Work-
space allows you to view the trajectory against a fixed background. You can superimpose the
trajectory atoms on the background, and you can set up measurements, such as H-bonds,
between the background atoms and the trajectory atoms, which are updated during play.

The toolbar in the Trajectory panel contains a standard set of controls for playing through the
trajectory frames, which are listed below. The menu bar has one menu, Play, which contains
items that correspond to the toolbar buttons.

Go to start
Display the first frame.

Previous
Display the previous frame.

Play backward
Display the frames in sequence, moving toward the first.

Stop
Stop playing through the frames.

Play forward
Display the frames in sequence, moving toward the last.

Next
Display the next frame.

Go to end
Display the last frame.

Desmond 3.1 User Manual 57

 Chapter 5: Analyzing Simulations

Loop
Choose an option for repeating the display of the frames. Single direction displays frames in
a single direction, then repeats. Oscillate reverses direction each time the beginning or end of
the frame set is reached.

Figure 5.1. The Trajectory panel.

58 Schrödinger Suite 2012 Update 2

 Chapter 5: Analyzing Simulations

You can control the selection of frames and the speed of play in the Frame control section of
the panel.

• The Start and End text boxes define the frames at which play starts and ends. Frames are
numbered from 0.
• The Frame slider and frame text box can be used to select the frame to view. The current
frame number is displayed in the text box below the slider. The total number of frames is
also displayed in a noneditable text box.
• The Step text box sets the number of frames to step when playing through frames. This
value does not affect the Frame slider. The frames that are selected for play can be
exported as a selection of frames, using the output buttons.
• The Time text boxes display the time for the current frame and the total time for the tra-
jectory. You can enter a time in the text box to select a frame.
• The Speed slider sets the speed at which the frames are played.

By default, the clipping planes window is automatically hidden when you play through a
trajectory, as this speeds up play by about 50%. It is displayed again (if it was originally
displayed) when play stops. If you want to see the clipping planes window during play, dese-
lect Hide clipping planes during continuous play.

In the Display section you can control how frames are displayed in the Workspace and what
features are displayed:

• Use lower quality drawing to speed up play—Use a lower quality representation of objects
(tubes, spheres) in the Workspace to speed up play. This option has no effect on wire
frame representation.
• Update secondary structure—Update the secondary structure assignment for each frame.
• Show simulation box—Show the edges of the simulation box (in purple).
• Show axes—Show the coordinate axes in green.
• Replicate system—Enter the number of replicas of the system to display in each of the
three directions. This enables you to visualize the movement across the simulation box
boundaries. These text boxes are unavailable if there are no periodic boundary conditions.
• Trajectory smoothing—Smooth the trajectory by averaging the coordinates over the spec-
ified number of frames.
• Positioning—Select on of these options to control the positioning of each frame relative to
the Workspace during play.
• Do not adjust—Do not adjust the positioning of each frame.

Desmond 3.1 User Manual 59

 Chapter 5: Analyzing Simulations

• Superimpose on frame—Align the structure in each frame by superimposing a

selection of atoms on the corresponding atoms in the frame number given in the text
box. Use the Pick atoms for positioning picking controls to make the atom selection.
• Superimpose on Workspace structure—Align the structure in each frame by super-
imposing a selection of atoms on the corresponding atoms in the Workspace (that
are not part of the trajectory). Use the Pick atoms for positioning picking controls to
make the atom selection, which must match both the Workspace structures and the
trajectory structure.
• Center molecules—Center the selected molecules in the Workspace. Use the Pick
atoms for positioning picking controls to make the atom selection.

• Pick atoms for positioning—Use these picking controls to select the atoms to superimpose
or to center. You should consider picking atoms that do not change their position much
during the simulation.

If you want to superimpose the trajectory on the Workspace structure, you must take care when
you select the atoms to superimpose. The ASL expression for the selection in the Workspace
structure is applied to each trajectory frame. If the ASL expression depends on the numbering
(atom number, molecule number, etc.) and the order of the objects in the trajectory frames is
not the same as in the Workspace structure, you may get unexpected results. You should use
ASL expressions that are not order-dependent. You can use the atom selection button to choose
from a variety of structural features whose ASL expressions are not order-dependent, like
Backbone or Ligands. If the ordering is a problem, you could instead align the Workspace
structure to a particular frame, and then play the trajectory by superimposing on that frame.

The atoms that are visible in each frame can be set either with the Workspace toolbar buttons,
or with the tools in the Atoms to display in each frame section.

To select the atoms that are visible when frames are displayed, use the toolbar buttons that
control the atom display:

The choice that you make with the toolbar buttons is recorded as an atom set, and is applied to
each frame. The same atoms are always displayed in each frame, no matter where they move in
the trajectory.

To display the atoms that come within a given distance of a particular set of atoms (such as a
ligand or a binding site), use the Atoms to display in each frame section. You can select the
atoms that are always visible with the Always display these atoms selection tools. You can then
choose to display entire residues that have any atoms within a specified distance of these
atoms. This expression is evaluated for each frame, which allows residues (such as water) to

60 Schrödinger Suite 2012 Update 2

 Chapter 5: Analyzing Simulations

move in and out of this distance. The particular set of atoms that is visible can therefore change
from frame to frame.

The output buttons allow you to export the trajectory data in various forms. You can export
individual frames, all frames, or the selection of frames defined by using the Start, End, and
Step text boxes. The buttons have the following actions:

• Structure—Save structures from the trajectory to a file or create project entries from the
structures. Opens the Export Structure dialog box, in which you can specify where the
structures will go, which structures to export, and which atoms to export.
• Image—Create an image of the Workspace with the current frame displayed. Opens the
Save Image panel.

• Movie—Save a movie of the trajectory in MPEG format. Opens the Export Movie panel,
in which you can select the frames to be exported, the speed and the resolution.

If you want to view a trajectory while a simulation is running, you can do so by importing the
jobname-out.cms file for the simulation from the host on which the simulation is running.
This file contains information on the location of the trajectory, and should be in the temporary
directory for the job. There may also be a copy of the output CMS file in the working directory
on the local host, but the trajectory will not be present until the job finishes. Once you have
imported the file, click the T button for the imported entry. To update the trajectory later, you
must reimport the file.

5.2 Simulation Quality Analysis

The Simulation Quality Analysis panel displays a simulation summary and an analysis of the
total energy, potential energy, temperature, pressure, and volume over the length of the simula-
tion. The analysis includes the average value, the standard deviation, and the slope of a linear
fit to the values of the property as a function of time. You can run the analysis on a completed
simulation or while the simulation is running.

To open the Simulation Quality Analysis panel, choose Applications → Desmond → Simulation
in the main window.

To load the desired simulation, click Browse, and navigate to the desired Energy Sequence file
for the simulation, which has a .ene suffix. If you want to perform the analysis on a running
simulation, you must use the file from the host on which the simulation is running.

The analysis performs averaging of the results over short time periods (“blocks”) to reduce the
noise and eliminate correlation between consecutive reports. To change the size of the block,
enter a value in the Block length for averaging text box.

Desmond 3.1 User Manual 61

 Chapter 5: Analyzing Simulations

Figure 5.3. The Simulation Quality Analysis panel.

When you have the desired block length, click Analyze to perform the analysis. The analysis
can take a few minutes. When it finishes, the Simulation summary and Properties sections are
filled in with the results of the analysis. If you want to view a plot of the thermodynamic prop-
erties as a function of simulation time, click Plot. A panel is displayed with the results plotted.

5.3 Simulation Event Analysis

You can use the Simulation Event Analysis panel to analyze the output from a Desmond simula-
tion and display the results of the analysis in various forms. The analysis involves the selection
of properties that are extracted from the trajectory; these properties can then be exported and
plotted, and statistics calculated for the properties.

To open this panel, choose Applications → Desmond → Simulation Event Analysis. Detailed
information about using this panel is available in the online help, which you can open by
clicking the Help button in the panel. A summary is given below.

Five types of properties are available: energy, hydrogen bonds, RMS deviation of atoms rela-
tive to a trajectory frame or reference structure, RMS fluctuation of atoms over the entire
trajectory relative to a trajectory frame or reference structure, and measurements of distances,
angles, dihedrals, and radius of gyration.

The first task in the analysis is to select properties and the atoms or molecules for which they
are calculated. These are added to the list in the center of the panel, where they are represented
as “keywords”. Next, these properties must be evaluated, by selecting individual properties in
the list and clicking Analyze, or clicking Analyze All to analyze all the properties. You can read
or write list of properties (keywords), and you can export the results of the analysis.

62 Schrödinger Suite 2012 Update 2

 Chapter 5: Analyzing Simulations

Figure 5.4. The Simulation Event Analysis panel.

Once the properties have been evaluated, you can display the results in various kinds of plots:
time series, histogram, multi-variable plot, heat map, polar plot, RMSF plot, and also display
statistics on the properties.

5.4 Replica Exchange Review

The Replica Exchange Dynamics Review panel can be used to display a plot of the replicas as a
function of time, on a temperature versus time graph. Each replica is colored uniquely so you
can trace its temperature as a function of time.

Figure 5.5. The Replica Exchange Dynamics Review panel.

Desmond 3.1 User Manual 63

 Chapter 5: Analyzing Simulations

Figure 5.6. The Replica Exchange Graph panel.

To open this panel, choose Applications → Desmond → Replica Exchange Review or Tasks →
Molecular Dynamics → Replica Exchange Review. First, you must load a log file for the simu-
lation (not the multisim log file). Then you can choose whether to show all replicas or a
single chosen replica. When you have made a choice, click Display Plot. The plot is displayed
in a new window, with a standard set of plotting controls for configuring the plot and saving an
image. If you are displaying single replicas, you can add replicas to the plot by selecting a new
one, or you can plot a different replica by closing the plot window, selecting a new replica, and
clicking Display Plot again.

5.5 Metadynamics Analysis

The Metadynamics Analysis panel analyzes a metadynamics calculation and displays a plot of
the free energy as a function of the collective variables (CVs) used for the calculation. You can
analyze calculations with one or two collective variables. The free energy value can be plotted
relative to the minimum value or as an absolute value. The plots can be saved and loaded. To
open this panel, choose Applications → Desmond → Metadynamics Analysis.

64 Schrödinger Suite 2012 Update 2

 Chapter 5: Analyzing Simulations

Figure 5.7. The Metadynamics Analysis panel.

Desmond 3.1 User Manual 65

66 Schrödinger Suite 2012 Update 2
 Desmond User Manual

Chapter 6

Chapter 6: Running Simulations from the

Command Line

Desmond jobs may be started from Maestro or from the command-line. The mechanisms for
running jobs from the command line are described in this chapter. You might wish to run
Desmond from the command line for any of the following reasons:

• To exercise greater control over Desmond behavior

• To debug an aberrant run
• For convenience
• To transfer the job to a remote location
The Desmond panels can be used to write out (using the Write button) .cfg files for a simula-
tion. Those .cfg files can then be run elsewhere, routinely from a batch script or customized
prior to use. More information on the contents of .cfg files is available in Appendix B.

Desmond jobs run under Schrödinger’s Job Control facility. This facility manages the execu-
tion and monitoring of jobs, and handles the input and output files and the incorporation of
results into a Maestro project. The Job Control Guide describes how to set up the information
needed for Job Control to run on the computers to which you have access. It includes informa-
tion on remote hosts, clusters, and batch queues.

As is the case for all Schrödinger software, the environment variable SCHRODINGER must be
set to the directory where the Schrödinger software, including Desmond, was installed. In addi-
tion, there are other environment variables that can be set to override default resource values.
See Appendix B of the Job Control Guide for more information.

By default, the Schrödinger job control facility uses ssh to communicate between remote
nodes. For more information, see Section 7.2 of the Installation Guide.

This chapter describes the command syntax for running single-stage and multistage Desmond
jobs, and for building the input composite model system. For details of the file formats and
technical background, see the Desmond User’s Guide, from D. E. Shaw Research.

6.1 The desmond Command

Desmond is run from the command-line by executing the script desmond, which is located in
the $SCHRODINGER directory. The desmond script determines the proper executables to run
for the host being used. Additional utility programs are provided in the $SCHRODINGER/
utilities directory.

Desmond 3.1 User Manual 67

 Chapter 6: Running Simulations from the Command Line

The syntax for the desmond command is as follows:

$SCHRODINGER/desmond [options] -in cms-file -c config-file

$SCHRODINGER/desmond [options] -restore checkpoint-file
Use the -in option for a new job, and the -restore option for restarting a job. The input file
for a new job must be a CMS file, with extension .cms, and a configuration file (.cfg) must
also be specified. For restarting a job, a checkpoint file, with extension .cpt, must be
provided.

The options for the desmond command are listed in Table 6.1. The standard Job Control
options are also supported—see Section 2.3 of the Job Control Guide. Additional diagnostic
options, which do not run the job but provide information, are also given in the section
mentioned. In particular, you should note the syntax of the -HOST option, which is used to
specify the list of hosts used for the job.

For information on storage of temporary files, interacting with running Desmond jobs, and so
on, see the Job Control Guide.

Table 6.1. Options for the desmond command.

Option Description

Job Options
-DEBUG Print diagnostic output.
-LOGINTERVAL Interval for copying the log file back to the submission host. Default: 5s.
-INTERVAL Interval for copying monitor files back to the submission host. Default:
5s.
-p|-P|-PROCS|-NPROC Number of processors to use. Default: 1.
-JOBNAME name Specify the name for the job. The default is the input .cfg or .cpt file
name, minus the extension.
-jin filename Files to be transferred to the execution host by Job Control.
-jout filename Files to be copied back to the submission host by Job Control.
-h[elp]|-HELP Print usage information
-v print version information and exit

68 Schrödinger Suite 2012 Update 2

 Chapter 6: Running Simulations from the Command Line

Table 6.1. Options for the desmond command. (Continued)

Option Description

Program Options
-exec program Run the specified Desmond program. Available programs are:
mdsim—Run a molecular dynamics simulation (the default)
minimize—Run a minimization
remd—Run a replica exchange molecular dynamics simulation
vrun—Analyze a trajectory
If a front-end config file is provided with -c (such as those created by
Maestro), the program to use is determined automatically and this option
is not needed.
-c config-file Parameter file for simulation. Required with -in for a new simulation.
-cfg key=val Specify extra simulation parameters. Multiple instances of this option
can be supplied to specify multiple extra parameters.
-comm plugin Use the specified communication plugin. Available plugins are serial
and mpi. Default: serial.
-dp Run double-precision version. Default: run single-precision version.
-noopt Do not optimize parameters automatically.
-t temperature Specify temperature in a replica exchange MD simulation. Include an
instance of this option for each temperature in the exchange.
-tpp n Specify the number of threads per processor.

Some examples of running desmond from command line are shown below:

1. Running a 4-processor job on a queuing system or a particular machine (myhost):

$SCHRODINGER/desmond -HOST myhost -P 4 -c x.cfg -in x.cms
2. Running a 4-processor job on two different machines (host1 and host2):
$SCHRODINGER/desmond -HOST "host1:2 host2:2" -P 4 -c x.cfg
-in x.cms
3. Continuing a 4-processor simulation from a checkpoint file (x.cpt):
$SCHRODINGER/desmond -HOST myhost -P 4 -restore x.cpt
4. Running a replica exchange simulation, in which each replica uses 2 processors with dif-
ferent input CMS files:
desmond -P 4 -c x.cfg -t 300 -in x300.cms -t 310 -in x310.cms

Desmond 3.1 User Manual 69

 Chapter 6: Running Simulations from the Command Line

6.2 Running Multiple Simulations

Molecular dynamics studies usually involve performing a series of calculations. For example, a
sequence of minimization, thermalization, and relaxation calculations followed by the produc-
tion simulation itself is usually undertaken for molecular dynamics study. The MultiSim utility
(multisim) facilitates the task of running multiple Desmond-related or MCPRO+-related
calculations in a sequential manner. The Desmond general and FEP panels can be used to write
out .msj files for those tasks. These .msj files can then be run elsewhere, for example from a
batch script, or customized prior to use. More information on the contents of .msj files is
available in the Maestro online help—see the list of Desmond topics.

The syntax for the multisim utility is as follows:

$SCHRODINGER/utilities/multisim [options] [-i] structure-file -m msj-file

$SCHRODINGER/utilities/multisim [options] -RESTART checkpoint-file
The first syntax should be used for new jobs, and the second for restarting jobs. The -i for the
structure file is deprecated, but can still be used. The structure file can be either a Maestro
(.mae) file or a CMS (.cms) file. For restarting jobs, -r can be used instead of -RESTART, but
-r is also deprecated. The options are described in Table 6.2, including job options. The stan-
dard Job Control options are also supported, as are the options -LOCAL, -WAIT, and
–NOJOBID—see Section 2.3 of the Job Control Guide.

Table 6.2. Options for the multisim utility.

Option Description

-ADD_FILE filename Additional input file to copy to the working directory of the
multisim job. By default, multisim identifies and copies most files
needed for the run.
-c cfg-file File that contains the default Desmond config parameters. This option
is only useful for Desmond subjobs.
-cpu num-cpus Number of processors for each Desmond subjob. This can be a string
indicating the processor topology e.g., '2 2 2'. If this option is not
provided, a default value set by either the backend or the protocol will
be used. Processor counts must be a power of 2, 3, or 5, or products of
these powers.
-d stage-file File (.tgz) containing information on the stages from the previous
job, when restarting a simulation. Use one instance for each file.
Requires a multisim checkpoint file.
-debug Turn on multisim debug mode.
-DEBUG Turn on both multisim and Job Control debug modes.

70 Schrödinger Suite 2012 Update 2

 Chapter 6: Running Simulations from the Command Line

Table 6.2. Options for the multisim utility. (Continued)

Option Description

-description string Job description to print to the log file.

-h[elp], -HELP [topic] Print usage message and exit. You can specify an optional topic with
the following topic references:
@all—Show concise help of all valid options.
@doc—Show detailed help of all valid options.
@option—Show detailed help of the specified option, e.g., @-cpu.
@examples—Show a few command examples.
@deprecated—Show a list of deprecated options.
-SUBHOST subjob-host Host on which to run the subjobs. -host can also be used instead of
-SUBHOST, but is deprecated. This option does not support the
host:cpus syntax.
-HOST multisim-host Host to run the multisim master job.
-JOBNAME jobname Job name of this multisim job
-RETRIES maxretries Maximum number of times to restart failed subjobs. -max_retries
can also be used instead of -RETRIES, but is deprecated.
-maxjob maxjob Maximum number of simultaneous subjobs
-mode mode Run in non-default job management mode. The only allowed value is
umbrella, for which subjobs run on the same node as the master job,
-SUBHOST is ignored and -maxjob is set to 1. All processors are
assigned to the master job, so that the subjobs do not need to be
queued. You must use an MPI-parallel queue with this option.
-nopref
-notify email Send email notification to the specified address when the job finishes.
-o output-file Output .cms or .mae file. This file can be incorporated into Maestro.
-probe checkpoint-file Probe a checkpoint file and quit. The state of the job recorded in the
checkpoint file is displayed, and whether the checkpoint file is com-
patible with the current version of multisim.
-quiet Light log information (default).
-set string Specify parameters that take precedence over those in the .msj file, as
a quoted string. Can be specified multiple times to make more than
one change to the parameters.
-v[ersion] Print the multisim version and exit.
-verbose Heavy log information

Desmond 3.1 User Manual 71

 Chapter 6: Running Simulations from the Command Line

6.2.1 Examples of Running MultiSim

A simple multisim command can take the form:

$SCHRODINGER/utilities/multisim -i input.cms -m input.msj

-HOST mycluster -JOBNAME multistage
which runs the multisim utility with the starting structure from input.cms. The instruc-
tions for different stages of multisim are provided in input.msj. The master job, which
coordinates the subjobs, is run on the host labeled mycluster in the schrodinger.hosts
file. The job name is multistage.

The following command explicitly uses a different host for the master job and the subjobs:

$SCHRODINGER/utilities/multisim -i input.mae -m input_fep.msj

-HOST tabitha -host kyla_para -JOBNAME example_fep_job
-maxjob 6 -cpu “2 2 2”
The -HOST option specifies the host of the master job, whereas -host specifies the compute
host for the subjobs. The keyword -maxjob 6 means that at most 6 subjobs are simultane-
ously in the queue of the compute host. When one of the subjobs finishes the next subjob is
launched on the compute host. At any time, the number of the subjobs launched from the
master job does not exceed 6 on the compute host.

6.2.2 Sample MultiSim Job (.msj) File

An example .msj file is given below. The file syntax is described in full in Appendix A. The
file is modular, consisting of a series of stages, each of which contains a list of keywords and
value settings enclosed within braces. Lines beginning with # are comments and hence are
ignored.
# Desmond standard NPT relaxation protocol
# All times are in the unit of ps.
# Energy is in the unit of kcal/mol.
task {
task = "desmond:auto"
}

minimize {
title = "Minimization with restraints on solute"
max_steps = 2000
steepest_descent_steps = 10
convergence = 50.0
restrain = { atom = solute force_constant = 50.0 }
}

minimize {

72 Schrödinger Suite 2012 Update 2

 Chapter 6: Running Simulations from the Command Line

title = "Minimization without any restraints"

max_steps = 2000
steepest_descent_steps = 10
convergence = 5.0
}

simulate {
title = "Berendsen NVT, T = 10 K, small timesteps, and restraints on solute
heavy atoms"
annealing = off
time = 12
timestep = [0.001 0.001 0.003]
temperature = 10.0
restrain = { atom = solute_heavy_atom force_constant = 50.0 }
ensemble = {
class = NVT
method = Berendsen
thermostat.tau = 0.1
}

randomize_velocity.interval = 1.0
eneseq.interval = 0.3
}

simulate {
title = "Berendsen NPT, T = 10 K, and restraints on solute heavy atoms"
annealing = off
time = 12
temperature = 10.0
restrain = retain
ensemble = {
class = NPT
method = Berendsen
thermostat.tau = 0.1
barostat .tau = 50.0
}

randomize_velocity.interval = 1.0
eneseq.interval = 0.3
}

simulate {
title = "Berendsen NPT and restraints on solute heavy atoms"
effect_if = [["@*.*.annealing"] ’annealing = off temperature =
"@*.*.temperature[0][0]"’]
time = 12
restrain = retain
ensemble = {
class = NPT

Desmond 3.1 User Manual 73

 Chapter 6: Running Simulations from the Command Line

method = Berendsen
thermostat.tau = 0.1
barostat .tau = 50.0
}

randomize_velocity.interval = 1.0
eneseq.interval = 0.3
}

simulate {
title = "Berendsen NPT and no restraints"
effect_if = [["@*.*.annealing"] ’annealing = off temperature =
"@*.*.temperature[0][0]"’]
time = 24
ensemble = {
class = NPT
method = Berendsen
thermostat.tau = 0.1
barostat .tau = 2.0
}

eneseq.interval = 0.3
}

simulate {
cfg_file = "example.cfg"
jobname = "$JOBNAME"
dir = "."
compress = ""
}

In the above example, the first stage, task, specifies the type of job that is run so that appro-
priate defaults can be set. In this case desmond:auto indicates that it is a Desmond job and
that the type of job should be detected automatically.

The second stage is a minimization of the system over a maximum of 2000 steps. Of the 2000
steps, the first 10 steps are be steepest descent. The convergence criterion is set rather loosely
to 50.0 kcal mol-1Å-1. The solute is restrained with a force constant of 50.0 kcal mol-1Å-1. The
monitor file is not updated. The third stage is similar except that nothing is restrained.

The fourth stage sets up a 12 ps Berendsen NVT simulation. For this simulation, the tempera-
ture is set to 10.0 K, and the thermostat relaxation is set to 0.1 ps. Resampling is done every
1 ps. The solute atoms are restrained. Checkpointing and structure monitoring are turned off.
Center-of-mass motion is removed and the .ene file is updated more frequently, at intervals of
0.3 ps. Subsequent simulation stages follow with progressively more freedom until the second
last stage when conditions resemble the production run.

74 Schrödinger Suite 2012 Update 2

 Chapter 6: Running Simulations from the Command Line

The last stage is the production dynamics simulation. The simulation parameters could have
been explicitly listed here as in the preceding stages. In this example, this stage simply refers to
a Desmond .cfg file. The last line gives an example of how this .msj file could be run. You
would need to change the options for the job.

More examples can be found in the $SCHRODINGER/mmshare-vversion/data/desmond

directory. Detailed information on the syntax of .msj files is available in [Appendix].

6.2.3 Treatment of Intermediate Files

In general, the files from each stage are copied back to the launch directory as a gzipped tar
archive with the name jobname_stage-number.tgz. The production simulation stage is treated
differently and the files are directly copied back without being archived. This treatment would
be similar to a simple molecular dynamics simulation (i.e. not a multisim job). This allows
the final output CMS file to be incorporated and the trajectory to be played as a normal job.

6.3 Building a Model System

To prepare a model system from the command line, you can use the system builder utility,
system_builder,. You can choose to add solvent, ions, and insert the solute into a
membrane. The system_builder utility takes a composite system builder (CSB) file as
input. This file contains all the required information to convert a solute file into a solvated
system.

The syntax of the system_builder command is as follows:

$SCHRODINGER/utilities/system_builder [-rezero] [-minimize_volume]

[job-options] input.csb
The standard Job Control options, described in Table 2.1 of the Job Control Guide, are
accepted, as are the options -WAIT, -LOCAL, and -NOJOBID, described in Table 2.2 of the Job
Control Guide.

The possible steps involved in the conversion are:

• Reading the solute from a file

• Reading the water model
• Reading the positive ion data
• Reading the negative ion data
• Adding a membrane
• Solvating
• Neutralizing
• Writing out a composite molecular system (CMS) file.

Desmond 3.1 User Manual 75

 Chapter 6: Running Simulations from the Command Line

Two other global tasks are performed by setting the command-line options:

-rezero Set the coordinate origin to the centroid of the solute coordinates.
-minimize_volume Minimize the volume of the simulation box.

The order of the keywords in a CSB file matters. Lines beginning with # are considered to be
comments and are ignored. An example CSB file is shown below:

{
read_solute_structure mysolute_setup-in.mae # solute file name
solvent_desmond_oplsaa_typer {
input_file_name spc.box.mae
run
}
positive_ion_desmond_oplsaa_typer {
input_file_name Na.mae
run
}
negative_ion_desmond_oplsaa_typer {
input_file_name Cl.mae
run
}
membranize POPE.mae.gz 10.000000 10.000000
create_boundary_conditions orthorhombic 0.000000 0.000000 10.000000
exclude_ion_from { 1 2 } 10.0
solvate
neutralize
write_maeff_file chorus_setup-out.cms
}
The various sections of the file are described in the following subsections.

6.3.1 Reading the Structures

The first five sections read in all the structural information for the solute, solvent, and ions. To
locate the files that contain this information, the current directory is searched first, then the
directory $SCHRODINGER/mmshare-vversion/data/system_builder.

The first section reads the solute and solvent structure.

read_solute_structure mysolute_setup-in.mae # solute file name

solvent_desmond_oplsaa_typer {
input_file_name spc.box.mae
run
}

76 Schrödinger Suite 2012 Update 2

 Chapter 6: Running Simulations from the Command Line

The keyword read_solute_structure reads the solute structure from the specified file. If
the file is given as a relative path, it is copied to the temporary directory for the job. If it is
given as an absolute path, the file must exist at that location on the execution host. The struc-
tures can be a solute structure, structures for different stages of FEP simulations, or a
completely solvated system. The force-field information is obtained by either reading from an
existing ffio_ff block in the input file or running the force field server.

Note that alternate coordinates are removed from the solvent structure when it is imported.

The solvent_desmond_oplsaa_typer section describes the solvent system that is used to

solvate the structure. The keyword input_file_name specifies the file to be used to read the
solvent model.

The next two sections determine the data to be used for counter ions to neutralize (or in some
cases charge) the system.

positive_ion_desmond_oplsaa_typer {
input_file_name Na.mae
run
}
negative_ion_desmond_oplsaa_typer {
input_file_name Cl.mae
run
}
These two sections describe the ion systems to neutralize or to add ions to the current struc-
tures. In each section one input_file_name keyword is provided to determine the file to be
used.

The next two sections (not used in the example file above) similarly determine the ion systems
to add salts to the system.

salt_positive_ion_desmond_oplsaa_typer {
input_file_name Na.mae
run
}
salt_negative_ion_desmond_oplsaa_typer {
input_file_name Cl.mae
run
}
The syntax for this section is same as for the positive and negative ion sections.

Desmond 3.1 User Manual 77

 Chapter 6: Running Simulations from the Command Line

6.3.2 Adding a Membrane

The keyword membranize instructs the system builder to create a membrane patch around the
current solute structure.

membranize filename x-buf y-buf

The first input parameter is the name of the file that includes the membrane template and its
equilibrating solvent. The other two parameters x-buf and y-buf specify the minimum distance
between the solute and the box boundary in the plane of the membrane.

6.3.3 Setting the Box Shape and Dimensions

The box shape and dimensions can be specified either in terms of an absolute size or in terms
of a buffer distance.

The keyword boundary_conditions is used to define a box with specified absolute dimen-
sions. The command using this keyword can be one of the following:

boundary_conditions cubic a
boundary_conditions orthorhombic a b c
boundary_conditions triclinic a b c alpha beta gamma
The box shape is the first parameter, and can be cubic, orthorhombic, or triclinic. It is
followed by the absolute size of the box defined by a or by a, b, and c. For a triclinic box
shape, the angles alpha, beta, and gamma must also be given.

Alternatively, the keyword create_boundary_conditions can be used to specify a box in

terms of the distance between solute atoms and box edges. he command using this keyword
can be one of the following:

create_boundary_conditions cubic a
create_boundary_conditions orthorhombic a b c
create_boundary_conditions triclinic a b c alpha beta gamma
The parameters are similar to those for the absolute box size, but the distances are the
minimum distance any solute atom and the box boundary in the given direction. The distance
between two images of the solute structures is therefore twice the specified values. If a
membrane is used, the box shape must be orthorhombic and the a and b values must be set to
zero so that solvent molecules are not placed within the membrane layers.

You can minimize the size of the box by using the -minimize_volume option to the
system_builder command.

78 Schrödinger Suite 2012 Update 2

 Chapter 6: Running Simulations from the Command Line

6.3.4 Setting Force Field Information

The command set_oplsaa_version can be used to specify OPLSAA version. Currently
only OPLS_2005 (value: 2005) is supported.

set_oplsaa_version 2005
By default, force-field parameters that are already present in the solute CTs (the ffio_ff
block) are replaced. The following setting can be used to retain existing parameters:

remove_solute_ffio no

6.3.5 Setting the Number and Location of Ions

The addition of ions to the system is governed by several keywords: add_ion,
ion_location, exclude_ion_from, add_salt, and neutralize.

Two commands using the keyword add_ion can be used to add specific numbers of positive
and negative ions.

add_ion positive 5
add_ion negative 5
The keyword add_ion adds positive or negative ions to the system. The number of ions to be
added is specified as the second argument of this keyword.

The ion_location keyword can be used to specify the proximity of ions with reference to
certain atoms.

ion_location { { atom1 } { atom2 } }

The keyword ion_location is followed by the list of solute atom indexes. Ions are placed
near the listed atoms. For each ion, the atom index is given inside braces. Extra atom index
specifications are ignored. If there are fewer atom index specifications than there are ions, the
remaining ion locations are determined randomly.

In order to exclude ions near certain atoms, the exclude_ion_from keyword can be used:

exclude_ion_from { atom-list } distance

The keyword exclude_ion_from has two arguments. The first argument is a list of atom
indices of solute atoms, in braces. The second argument is the distance value. No ion, whether
a single ion or from a salt, is placed within the given distance in angstroms from the listed
atoms.

The salt concentration is determined using the add_salt keyword.

Desmond 3.1 User Manual 79

 Chapter 6: Running Simulations from the Command Line

add_salt conc
The keyword add_salt adds the positive and negative salt ions to the system defined above.
The argument specifies the salt concentration for the molecular system. Based on the volume
of the solvent and the concentration, the number of salt ions is calculated and rounded to an
integer value. The coordinates of salt ions are determined randomly.

The following command can be used to neutralize the system.

neutralize
The net charge of the current solute structures is calculated and the number of counter ions
necessary to neutralize the system is added.

6.3.6 Solvating the System

This command is used to initiate the solvation of the current solute structure, which includes
the ions and the membrane.

solvate
The keyword solvate is used to solvate the current solute structures using the solvent system
defined in the solvent_desmond_oplsaa_typer section. The solvent system is extended to
be consistent with the boundary condition defined by the boundary_conditions or
create_boundary_conditions section. Any solvent molecules that overlap the solute
structures are removed.

6.3.7 Writing the Output File

This command is used to write the output to a “maeff” (Maestro + force-field) file, otherwise
known as a composite model system (CMS) file, which has the extension .cms.

write_maeff_file filename
The keyword write_maeff_file writes the current composite structures and their force field
parameters into a file in CMS format. The first structure contains all of the molecules in the
system and is usually referred to as the “full system CT”. The structures (or CTs) that follow it
contain different components of the system. For instance there usually is a solute structure, a
solvent structure (containing all of the solvent molecules) and structures for different types of
ions. There may also be a structure that contains solvent molecules that are extracted from the
solute. The force field parameters are inserted as ffio_ff blocks within each CT block
(“structure”) except the first. The coordinate origin of the structures is by default the center of
mass of the solutes.

80 Schrödinger Suite 2012 Update 2

 Desmond User Manual

Chapter 7

Chapter 7: Using VMD for Desmond Trajectories

VMD [7] is a powerful program for visualizing molecular dynamics simulations that is avail-
able from the Theoretical and Computational Biophysics Group at the University of Illinois at
Urbana-Champaign. A plugin for Desmond is provided as part of VMD that makes it possible
for VMD to read and write Maestro files and read Desmond trajectories. The output Maestro
files are suitable for use in building model systems that can then be used to run Desmond.

This chapter contains information on reading a CMS file and a Desmond trajectory into VMD,
and writing a Maestro file from VMD. For information on installing VMD, see Section 3.10.4
of the Installation Guide.

To start VMD, enter the command:

vmd
Two windows are opened, VMD main and VMD version OpenGL Display. The former can be
used to control VMD while the latter can be used to display the molecular systems and to view
trajectories. VMD has extensive documentation which is available from the Help menu in the
VMD main window.

7.1 Reading a CMS File and a Desmond Trajectory

This section describes how to read a Maestro CMS file and a Desmond trajectory into VMD.
To view a Desmond trajectory in VMD, you must first read in the output structure CMS file
from a Desmond simulation. This file is usually named jobname-out.cms. Viewing trajecto-
ries from Desmond FEP simulations is not currently supported.

To read in a CMS file:

1. Choose New Molecule from the File menu of the VMD Main window.
The Molecule File Browser panel opens.
2. Click Browse.
A file selector opens.
3. Navigate to and select the output .cms file from a Desmond simulation, and click OK.
The file selector closes.

Desmond 3.1 User Manual 81

 Chapter 7: Using VMD for Desmond Trajectories

4. In the Molecule File Browser panel, set the file type to Maestro File (no timesteps).
5. Click Load.

You have just created a Molecule listing in VMD which should appear as a new a line in the
VMD Main window corresponding to the CMS file that you just read in.

To read in a trajectory for this system:

1. Ensure that the line for the CMS file that you just read in is the only one that is high-
lighted.
2. Choose Load Data into Molecule from the File menu in the VMD Main window.
The Molecule File Browser panel opens.
3. Click Browse.
A file selector opens.
4. Navigate into the trajectory directory from a Desmond simulation.
5. Select the file clickme.dtr and click OK.
The file selector closes. In the Molecule File Browser the file type should now be listed as
Desmond trajectory.

6. Click Load.
The trajectory should load into VMD and automatically start playing.

7.2 Writing a Maestro File

You can export individual configurations into a Maestro file from within VMD. This can be a
useful way to convert configurations produced by various programs into a format that will
function with Maestro.

To export a Maestro file from VMD:

1. Select the molecule that you want to export in the VMD main window.
2. Choose Save Coordinates from the File menu.
The Save Trajectory panel opens.
3. Set the File type to mae.
4. Click Save.

82 Schrödinger Suite 2012 Update 2

 Chapter 7: Using VMD for Desmond Trajectories

5. In the Filename text box, edit the directory, and append the name for the new file to the
directory.
This is a plain Maestro file, whose extension should be .mae.
6. Click OK.

Desmond 3.1 User Manual 83

84 Schrödinger Suite 2012 Update 2
 Desmond User Manual

Chapter 8

Chapter 8: Alternate Force Field Parameters and

Constraints

The Desmond installation includes two utilities, viparr and build_constraints that can
be used to add or adjust the force field parameters and the accompanying constraints for chem-
ical systems prior to simulating them with Desmond. Both programs read and write Maestro
structure files.

Viparr is a template-based force field assignment utility that comes with a number of built-in
force fields including some developed for Amber and CHARMM (see below for more infor-
mation). User-defined force fields are also supported by viparr, if they are provided in viparr’s
file format. Viparr can be used to specify different force fields for various components of the
system, provided that the force fields are compatible. This flexibility makes it possible to do
some things that may be useful in force-field development including:

• Using one force field for one part of the chemical system and another force field for
another part (this allows you, for example, to easily switch between water models)
• Using one or more components from one force field (e.g., the dihedral parameters) and
the remaining components from another force field
• Overriding some of the parameters (e.g., some but not all of the angle parameters) with
those from another force field

Some classes of constraints are often used with, and in some cases required by, various force
field representations of molecules. The utility build_constraints adds these constraints to
a structure file. You should run this utility after running viparr, to ensure that the force field is
ready for use.

8.1 The viparr Utility

To run viparr, use the following command:

$SCHRODINGER/run -FROM desmond viparr.py [options] input-file output-file

where input-file and output-file are the input and output structure (CMS) files, respectively.
This utility does not run under Job Control. It does not take much time, so it can be run locally.

The basic options are given in Table 8.1. The force field data files are located in the directory
$SCHRODINGER/desmond-vversion/data/viparr. When you have run viparr, you must run
build_constraints, to ensure that the force field constraints are set properly.

Desmond 3.1 User Manual 85

 Chapter 8: Alternate Force Field Parameters and Constraints

Table 8.1. Options for the viparr utility.

Option Description

-h Print usage message, including the names of the built-in force fields
-n name Specify force field name or other annotation to put into the output file. If not spec-
ified, a default name is used.
-f ffname Specify a built-in force field. The available force fields are listed in Table 8.2. You
can repeat this option to specify multiple force fields, one per instance of the
option. Parameters of force fields listed earlier override parameters of force fields
listed later. When multiple force fields are specified, the order is important if some
parameters are intended to override others.
-c ctnum Specify the index of a single structure (“CT block”) in the input file for processing
by viparr. Structures are numbered starting at 1. You can provide multiple -c
options to specify more than one structure to process. The default is to process all
structures.
-d ffdir Specify a user-defined force-field directory. You can repeat this option to specify
multiple directories. As for -f, the order is important.
-m mergedir Path to user-defined force field directory that is to be merged with previously spec-
ified force field. Multiple directories can be specified with multiple instances of
this option.
-nocompress Do not use ffio block compression.
-p pdir Specify the plugin directory. The default is to use the directory defined by the
environment variable VIPARR_PDIR, which contains the standard plugins. All
necessary plugins, including those for the built-in force fields, must be in the
directory specified by -p.
-v Verbose output.

The available force fields are listed in Table 8.2, with references.

86 Schrödinger Suite 2012 Update 2

 Chapter 8: Alternate Force Field Parameters and Constraints

Table 8.2. Force fields available with viparr.

Force Field Ref. Force Field Ref.

Amber OPLS-AA
amber94 8 oplsaa_impact_2001a 24-29
b,c
amber96 9 oplsaa_impact_2005 24-31
amber99 10 oplsaa_ions_Jensen_2005 32
amber99SB 10,11 Water models
amber99SB-ILDN 12 spc 34
amber03 13 spce 35
CHARMM tip3p 36
charmm22nocmap 14–16 tip3p_charmm 37
charmm27 14–21 tip4p 38
charmm32 14,15,20–22 tip4pew 39
charmm36_lipids 23 tip4p2005
tip5p 40

a. Parameter assignment as implemented in the OPLS_2001 force field in Impact.

b. Parameter assignment as implemented in the OPLS_2005 force field in Impact.
c. Note that the system_builder’s implementation of OPLS_2005 has more general coverage of
ligand-like molecules than viparr.

8.2 The build_constraints Utility

To create a constraint block use the following command:

$SCHRODINGER/run -FROM desmond build_constraints.py [options] input-file

output-file
where input-file is a structure file that has been previously processed with viparr and output-
file is a new structure file for the system with the constraints added. The options are given in
Table 8.3. The following constraint types are detected and automatically added:

• AHn, where n is a count of the number of hydrogen atoms (1, 2, 3, ...) in a group com-
posed of a heavy atom and the hydrogen atoms directly bonded to it
• HOH, oxygen bonded to two hydrogen atoms and no other atoms.

Desmond 3.1 User Manual 87

 Chapter 8: Alternate Force Field Parameters and Constraints

Table 8.3. Options for the build_constraints utility.

Option Description

-a, --angles Constrain AH2 and AH3 angles.

-h, --help Show usage message and exit.
-k, --keep Keep bonded terms that coincide with constraints.
-r, --revert Remove constraints and restore built terms.
-v, --verbose Print constraint terms.
-x exclude, Don't use the specified type of constraint (HOH, AH1, AH2, ...).
--exclude=exclude

8.3 Input and Output Files

The input structure file should contain all the atoms in the chemical system that are to be simu-
lated, including hydrogen atoms, water molecules, ions, and so on. The chemical system may
contain a number of structures (also called connection tables or CTs). Residues (including
water molecules and ions) in the chemical system are matched to templates in the force fields.
The CTs should also contain the following CT-level properties:

r_chorus_box_ax, r_chorus_box_ay, … r_chorus_box_bx,… r_chorus_box_cz

that specify the size and shape of the simulation box. The output from the system_builder
utility meets these conditions.

Residue Matching

Atomic numbers and the bonding pattern (graph isomorphism) are used to match residues to
templates. This methodology supports nonstandard atom or residue PDB names without modi-
fication. Atom and residue names in a force field need not be edited. In particular, viparr will
identifies the N- and C-terminus versions of the residues correctly, as well as protonated and
deprotonated versions of a residue, even if they are not explicitly mentioned as such in the
input file. Modification of atom and residue names for clarity is allowed.

Residue and Atom Ordering

The atom ordering in the input file is retained in the output structure file. The residue
numbering, which also remains unaltered, can begin with any integer (including negative inte-
gers) and does not need to be contiguous (viparr constructs a contiguous set of indices that it
uses internally). Residues with different chain names can have the same residue number. To aid
in diagnosing problems with the input structure file, messages involving residues have the form

88 Schrödinger Suite 2012 Update 2

 Chapter 8: Alternate Force Field Parameters and Constraints

<"chain-name",residue-number> (residue-name)

and are usually preceded by a structure number.

Output Format

A compressed force field representation is written when all the residues in a CT are the same.
For a CT that only contains water molecules, this means that force field parameters are written
only for a single water molecule.

A version number, which is associated with a particular version of viparr along with the
versions of the built-in force fields and their associated plugins, is written into the output struc-
ture file (in the ffio_version field). You are responsible for versioning your custom force
fields, using Perforce, for instance.

Potential Sources for Errors

If viparr reports that it cannot match a residue, please check the following:

• The template for the residue is really in the force field selected.
• Atom numbers for the residue are correct in the input structure file.
• Bonds for the residue are correct in the input structure file.

8.4 Specifying Multiple Force Fields

Multiple force fields can be specified using viparr. Common scenarios for this are outlined
below.

Different force fields for different parts of a chemical system

In this scenario, one force field is used for one part of a chemical system (e.g., the protein) and
another force field for another part (e.g., the water molecules). In this case, each residue in
your chemical system matches a template in exactly one of the specified force fields (warning
messages are printed otherwise).

Examples:

# use spc water model

$SCHRODINGER/run -FROM desmond viparr.py -f amber99 -f spc \
example.mae output.mae
# use tip3p water model
$SCHRODINGER/run -FROM desmond viparr.py -f amber99 -f tip3p \
example.mae output.mae

Desmond 3.1 User Manual 89

 Chapter 8: Alternate Force Field Parameters and Constraints

Combining components of two or more force fields

In this case, residues in the chemical system match templates in more than one of the specified
force fields (warning messages are printed). All matching force fields are applied. For
example, one force field provides the angle parameters for the residues, while another force
field provides the dihedral parameters. This can work if the force field components are disjoint
and there is no conflict in what parameters are assigned to each component.

Overriding parameters in a force field

Similar to the scenario mentioned above, residues in the chemical system match templates in
more than one of the specified force fields (warning messages are printed when this happens)
and all matching force fields are applied. However, if two or more force fields provide parame-
ters for the same term (e.g., two force fields provide parameters for the angle between atoms 1,
2, and 3) the conflict is resolved by using the parameters from the first force field listed on the
command line that matches the residue.

In all cases, if a bond exists between two residues that are not matched by the same force field,
viparr exits with an error message. You should correct the problem so that this bond is recog-
nized by one of the selected force fields. The force fields must have consistent van der Waals
mixing rules: viparr exits with an error message if they do not.

When using multiple force fields, viparr does the following:

• If any residue matches more than one template in a force field, viparr exits with an
error. No viparr force field should contain identical templates.
• If any residue name is matched to a force field template with a different name, a message
is printed. A maximum of 5 messages are printed per residue-template name pair.
• If there are any unmatched residues, viparr prints all unmatched residues and exits with
an error. A maximum of 5 messages are printed per unmatched residue name.
• If any residue is matched by more than one of the selected force fields, viparr prints a
warning message. You should be sure that you intended multiple force fields to match and
if so that the appropriate one was selected.

8.5 User-Defined Force Fields

A force field is composed of the following files:

• a template file
• a force-field parameter file, generally for each component of the force field. Angles,
proper dihedrals, van der Waals, etc. are examples of force field components

90 Schrödinger Suite 2012 Update 2

 Chapter 8: Alternate Force Field Parameters and Constraints

• a set of plugin programs that process the parameter files

• a rules file, which includes a list of plugin programs that viparr can call

A set of plugin programs has been provided for the built-in force fields. In most cases, user-
defined force fields can use these plugin programs. These plugin programs are located in
VIPARR_PDIR. VIPARR_PDIR is an environment variable that should be set to
$SCHRODINGER/desmond-vversion/lib/Linux-x86/viparr_plugins.

The other files, namely the templates file, parameter files, and rules file that specify a given
force field are placed in a force field directory. The force field directories for the built-in force
fields are located in the directory given by the environment variable, VIPARR_FFDIR, which
should be set to $SCHRODINGER/desmond-vversion/data/viparr.

The -d, -m and -p options, described in Table 8.1, are provided for working with user-defined
force fields.

8.6 Known Issues

Two or more geometrically identical hydrogen atoms (such as in CH2 or CH3) are treated
identically. If the force field needs to treat them differently, you must ensure that the residue
name exactly matches the force field template name.

When you import a structure into Maestro, make sure you correct any problems that Maestro
detects, especially those involving the atomic numbers of ions.

Desmond 3.1 User Manual 91

92 Schrödinger Suite 2012 Update 2
 Desmond User Manual

Chapter 9

Chapter 9: Utilities

The Desmond distribution contains a number of utilities for performing a range of specific
tasks, apart from those described previously. These utilities are described in this chapter.

9.1 solvate_pocket
The solvate_pocket utility is a tool for solvating subregions of a system with water, and in
particular to solvate buried regions in a protein or protein-ligand complex. To solvate such
regions in a thermodynamically consistent manner, solvate_pocket uses a grand canonical
Monte Carlo approach to sample both the water molecule positions and the number of water
molecules present, by attempting to introduce or remove water molecules in a Metropolis-like
manner.

9.1.1 Methodology
In grand canonical methods, a value of the chemical potential is set and the simulation samples
the number of water molecules in a manner consistent with the chemical potential and the
specified temperature. In principle, even the water in buried pockets is in equilibrium with bulk
water and so one should conduct the simulation using the excess chemical potential for bulk
water (for TIP4P this is about -6.95 kcal/mol).

The solvate_pocket utility samples the number and conformation of water molecules
within an orthorhombic region of the simulation cell using grand canonical Monte Carlo
(GCMC) in the μVT (constant chemical potential, volume and temperature) ensemble. As a
short-cut it does not use periodic boundary conditions other than to center the orthorhombic
cell in the simulation prior to simulating water molecules in the sampled region. As such, the
subregion sampled by solvate_pocket should be surrounded overall by bulk solution even if
the sampled region itself is not fully solvated. In its current form solvate_pocket expects a
CMS file as input and produces a CMS file containing the final conformation of the system
from the GCMC simulation.

Interactions are calculated using the real-space part of the Ewald sum only, an approximation
that still gives reasonable energetics. For instance, the chemical potential of bulk TIP4P water
using this approach is around -7.17 kcal/mol.

The solvate_pocket utility uses the concept of a “pass”, which is roughly 1 translation/rota-
tion move for each molecule being sampled. This can be useful because 1 pass is very roughly

Desmond 3.1 User Manual 93

 Chapter 9: Utilities

equivalent to a MD time step in terms of the amount of sampling for small molecules like
water. The passes can include insertion and deletion moves as well. Both types of moves are
required in order to equilibrate the number of water molecules in the system.

9.1.2 Command Syntax

The syntax of the solvate_pocket command is as follows:

$SCHRODINGER/utilities/solvate_pocket [options] [-NJOBS subjobs]

-spd command-file -icms input-cms-file -ocms output-cms-file
The options are given in Table 9.1.

Table 9.1. Options for the solvate_pocket command.

Option Description

-ChargeA n For FEP calculations, specify the lambda value for the charges for the origi-
nal molecule.
-ChargeB n For FEP calculations, specify the lambda value for the charges for the
mutated molecule.
-h[elp]|-HELP Print the usage message.
-lct N Use the Nth structure in the input CMS file to define the region sampled by
solvate_pocket. N must be greater than 1 (the full system structure).
Multiple instances of this option can be used to specify more than one struc-
ture for the region definition. Used to override the region specified in the
command file. Not compatible with -lmae.
-lmae ligand-file Specify the name of the ligand file used to define the region sampled by
solvate_pocket. Used to override the region specified in the command
file. Not compatible with -lct.
-seed number Provide a seed (an integer) for the random number generator.
-vdwA n For FEP calculations, specify the lambda value for the van der Waals poten-
tials for the original molecule.
-vdwB n For FEP calculations, specify the lambda value for the van der Waals poten-
tials for the mutated molecule.
-NJOBS jobs Run solvate_pocket the specified number of times, with the same input
but different random seeds. The output CMS file is the file with the number of
waters that is closest to the average over all the subjobs.

94 Schrödinger Suite 2012 Update 2

 Chapter 9: Utilities

Here, command-file is the name of the command file for solvate_pocket, which has the
extension .spd, and is described in Section 9.1.3; input-cms-file and output-cms-file are the
input and output CMS (composite model system) files and usually have a .cms extension. The
optional ligand-file is a file containing the ligand. If it is present, it is used to define the subre-
gion for solvation. By default, the maximum and minimum x, y, and z values in the command
file are used. If any of -ChargeA, -ChargeB, -vdwA, or -vdwB is used, all of them must be
used and it is assumed that this run is for a relative FEP calculation.

The solvate_pocket utility is run under Job Control by default.

Sampling can be improved by running multiple instances of solvate_pocket with the same
input file but different random seeds. These jobs can be run simultaneously by specifying more
than one processor with the -HOST option. The number of instances can be specified with the
–NJOBS option; by default it is set to the number of processors requested. When all subjobs are
complete, the output CMS file with the number of water molecules that is closest to the
average number across the subjobs is returned as the output structure file for the job as a whole.

A solvate_pocket run on a binding site that can contain around 15 water molecules (about
450 Å3) can take between 10 minutes and an hour. The cost of the simulation increases approx-
imately as the square of the number of water molecules sampled. It is unlikely that
solvate_pocket is needed to prepare systems for NPT simulations if there are no buried
pockets.

9.1.3 Command File Syntax

The solvate_pocket utility uses a keyword = value syntax that is at least nominally consis-
tent with Ark syntax. This means that you can use the same keywords in both the
solvate_pocket command file and in the solvate_pocket stage of the multisim input
file. Comments are lines that begin with a # character. Blank lines are ignored. Strings are
limited to 80 characters and must not contain a new line character. The keywords are described
in Table 9.2.

Table 9.2. Keywords for the solvate_pocket utility

Keyword Description

chargeA For FEP calculations, the lambda value for the Coulombic potentials for
the original molecule. Must be used with chargeB, vdwA, vdwB.
chargeB For FEP calculations, the lambda value for the Coulombic potentials for
the mutated molecule. Must be used with chargeA, vdwA, vdwB.
chemical_potential The chemical potential of water in kcal/mol. Required.
Recommended: –7.17 kcal/mol (for TIP4P).

Desmond 3.1 User Manual 95

 Chapter 9: Utilities

Table 9.2. Keywords for the solvate_pocket utility (Continued)

Keyword Description

cut_off The cutoff distance for calculating electrostatic and Lennard-Jones inter-
actions, in angstroms. Required. Recommended: 9.0 Å.
distribution_window The window used for calculating the distribution of the number of water
molecules. If this is larger than the value specified by num_passes it is
reduced to that value. If early termination upon convergence is used and
the number of passes carried out is smaller than specified by
distribution_window, that number of passes is used instead.
init_num_passes The number of Monte Carlo passes to perform in the equilibration prior
to the production Monte Carlo run. Recommended: 10000.
lig_ct_nums List of structures in the input CMS file to define the region sampled. The
value must be greater than 1, which is the full system CT.
max_dctheta The maximum change used for the cosine of theta (Euler angle) for a
combined translation/rotation move, in radians. Required. Recom-
mended: 0.0654.
max_disp The maximum change used in each of the x, y, and z directions for a
combined translation/rotation move. Required. Recommended: 0.105 Å.
max_dpsi The maximum change used for phi and psi (Euler angles) in radians for
a combined translation/rotation move. Required. Recommended: 0.318.
name A descriptive name for the simulation.
num_delete The number of attempts to delete a single water molecule per pass.
Required. Recommended: half the number of water molecules expected.
num_insert The number of attempts to insert a single water molecule per pass.
Required. Recommended: half the number of water molecules expected.
num_passes The number of Monte Carlo passes to perform in the production Monte
Carlo run. Required. Recommended: at least 20000.
num_trans_rot The number of water molecule translations to attempt per pass.
Required. Recommended: approximately the number of water mole-
cules expected to reside in the region being sampled.
output_structure Controls the output structure produced. Valid values are:
final—The last configuration sampled. This is the default.
average—The last structure from the trajectory that has a number of
dynamic water molecules closest to the average.
most_frequent—The last structure from the trajectory that has a num-
ber of dynamic water molecules closest to the most visited number.
If average or most_frequent are given then both
trajectory_freq and distribution_window must be given.

96 Schrödinger Suite 2012 Update 2

 Chapter 9: Utilities

Table 9.2. Keywords for the solvate_pocket utility (Continued)

Keyword Description

pass_term_window The number of passes over which the slope of the standard deviation of
the number of water molecules is calculated. This keyword may be used
to terminate the calculations before the number of passes specified by
num_passes has been completed. The simulation portions will run for
at least twice this duration before early termination occurs.
Default: 100000.
renum_waters If a nonzero value is given, renumber the water residues, starting from 1.
Otherwise retain the input water residue numbering.
sample_xmin Limits of the orthorhombic box in which the water molecules will be
sample_xmax sampled. For WaterMap jobs this box should enclose the binding site.
sample_ymin This box should be completely enclosed by atoms after it is centered in
sample_ymax the input periodic cell.
sample_zmin
sample_zmax
short_dist The closest acceptable approach for any two atoms, in angstroms.
Required. Recommended: 1.0 Å.
temperature The temperature used in the Monte Carlo simulation. Required.
term_std_slope Threshold for the standard deviation of the number of water molecules.
The calculation is terminated if the standard deviation falls below this
value. Default: 0.00001.
trajectory_freq If present, specifies the frequency at which structures are saved to a
Maestro pose viewer file (_pv.mae). The first structure contains the
fixed portion of the system. Structures containing only the water mole-
cules are written every n passes, where n is the value assigned to
trajectory_freq.
update_frequency The number of passes between updates in the log file. Default: 10.
vdwA For FEP calculations, the lambda value for the van der Waals potentials
for the original molecule. Must be used with vdwB, chargeA, chargeB.
vdwB For FEP calculations, the lambda value for the van der Waals potentials
for the mutated molecule. Must be used with vdwA, chargeA, chargeB.

Below is an example solvate_pocket command file.

name= solvate_pocket example command file

temperature= 298.15
init_num_passes= 10000
num_passes= 100000

Desmond 3.1 User Manual 97

 Chapter 9: Utilities

update_frequency = 10
pass_term_window = 10000
term_std_slope = 0.00001

num_trans_rot= 20
num_delete= 5
num_insert= 5

max_disp= 0.105
max_dpsi= 0.318
max_dctheta= 0.0654

sample_xmin= -8.0
sample_xmax= 8.0
sample_ymin= -8.0
sample_ymax= 8.0
sample_zmin= -8.0
sample_zmax= 8.0

cut_off= 9.0
short_dist= 1.0
chemical_potential= -7.17
This file instructs solvate_pocket to sample water within a cube, 16 Å on a side, centered at
0.0. 10,000 passes will be used to equilibrate the system before the production simulation
starts. The production simulation will run up to 100,000 passes but will terminate before that
number is reached if the slope of the standard deviation for the number of molecules drops
below 0.00001 over a window of 10,000 passes.

9.2 manipulate_trj.py
This script can be used to generate a new trajectory from a list of input trajectories. The trajec-
tory includes the .cms and .idx files that are needed to display the trajectory in Maestro. The
command syntax is as follows:

$SCHRODINGER/run -FROM desmond manipulate_trj.py [-h|--help]

[--mode={merge|concat [--dt=time]}] [-icms=cmsfile ] output-trj
input-trj1 [input-trj2 ... ]
This script currently supports two modes of operation, specified by the --mode option:

• merge—Multiple input trajectories are merged into one new trajectory based upon the
chemical time. For example, if the trajectories A = [a0 , ... , an ] and B = [b0 , ... ,bn ] are
merged, all frames from trajectory A whose chemical time is larger than that for b0 are

98 Schrödinger Suite 2012 Update 2

 Chapter 9: Utilities

discarded. Here, the trajectories are represented as a list of frames ai and bi. This is the
default mode, and is useful for merging trajectories that are continued in a new run.
• concat—Frames from the input trajectories are simply concatenated, and the time for
each frame is reset to account for the new ordering. You can specify the time between two
adjacent frames in ps with the --dt option.

You can select a subset of the frames present in each input trajectory with a syntax similar to
that used for Python lists. If a list is used, the entire trajectory specification must be quoted.
The examples below illustrate the syntax:

in_trj include all frames from in_trj

"in_trj[::]" include all frames from in_trj
"in_trj[0,6,8,10]" include frames 0, 6, 8 and 10 from in_trj
"in_trj[1:3:, 5]" include frames 1, 2 and 5 from in_trj
"in_trj[0, 4:11:2, 20]" include frames 0, 4, 6, 8, 10 (4–11 in increments of 2), and
20 from in_trj

The frame index is sorted in ascending order for each trajectory before any subset is selected.
The extension used in the input CMS file is preserved in the output CMS file (e.g. -out.cms).

Some examples are given below.

• To merge in1_trj and in2_trj:

$SCHRODINGER/run -FROM desmond manipulate_trj.py out_trj in1_trj
in2_trj
• To concatenate frame 0, 3, 6, 13, 5 of in1_trj and all frames of in2_trj:
$SCHRODINGER/run -FROM desmond manipulate_trj.py --mode concat
out_trj "in1_trj[0:7:3, 13, 5]" "in2_trj[::]"

9.3 amber_prm2cms.py
The utility amber_prm2cms.py can be used to convert Amber input files (a prmtop file plus a
prmcrd file) into a Desmond .cms file. The syntax for this command is:

$SCHRODINGER/run amber_prm2cms.py -c prmcrd -p prmtop -o cms-file

where prmcrd is the Amber prmcrd coordinate or restart file, prmtop is the Amber parameter
and topology file, and cms-file is the output CMS file.

The Amber parameter and topology file can be generated by the PARM or LeaP programs in the
Amber package. As well, AnteChamber, which is available for free under a GNU general

Desmond 3.1 User Manual 99

 Chapter 9: Utilities

public license from http://ambermd.org/antechamber/antechamber.html can also be used to

construct Amber input files.

9.4 mold_gpcr_membrane.py
The utility mold_gpcr_membrane.py can be used to embed a GPCR in a membrane.

With more GPCR X-ray structures being determined over the last few years and the inherent
flexibility of GPCR structures, there is great interest in simulating them, particularly within a
membrane. Properly constructing the system to simulate can be tedious, time consuming and
error prone. Proper alignment of the GPCR in the membrane can be difficult to attain. As well,
the relaxation of the membrane around the protein can take a very long time.

Most GPCR proteins are simulated based on either b2 or rhodopsin structures. As a result
membranes that are equilibrated around b2 or rhodopsin structures are likely to be able to
accommodate another GPCR structure with only relatively mild clashes. The utility
mold_gpcr_membrane.py, can automatically align a GPCR model to a2a, b1ar, b2ar or
rhodopsin and use the pre-equilibrated membrane structure from the latter to reduce the
membrane relaxation time and the potential structural problems that can arise during
membrane equilibration.

To run mold_gpcr_membrane.py, use the following command:

$SCHRODINGER/run mold_gpcr_membrane.py [options] {-j|-JOBNAME} jobname

-r[eference] reference-structure -l[ipid] lipid-type input-mae-file
The input Maestro file contains one or more model GPCR systems as separate structures. An
output structure file is generated for each structure present in the input file. The files are named
jobname_n-out.cms, where n is the index of the structure in the input file, starting from 1. If
there is only one input GPCR structure then _n is omitted.

The arguments and options are given in Table 9.3.

Table 9.3. Options for the mold_gpcr_membrane.py script.

Option Description

-j|-JOBNAME jobname Job name.

-f|-reference_file User reference template file. File that contains equilibrated template/
filename membrane systems. Contact Schrödinger if you want to set up your own
template file.
-l[ipid] lipid-type Specify the type of lipid in the membrane. Valid values are dppc or
popc.

100 Schrödinger Suite 2012 Update 2

 Chapter 9: Utilities

Table 9.3. Options for the mold_gpcr_membrane.py script.

Option Description

-r[eference] reference Specifies the GPCR template. Valid values are a2a, b1ar, b2ar, or rho
(rhodopsin).
-z[_dist] distance Buffer distance for solvation of the system. Default: 10 Å.
-h[elp] Show usage message and exit
-v[ersion] Show program version and exit

9.5 trajectory_extract_frame.py
The trajectory_extract_frame.py script can be used to extract selected frames from a
trajectory into a series of output structure files. To run this script, use the following command:

$SCHRODINGER/run -FROM desmond trajectory_extract_frame.py [options]

cmsfile trjdir
The options are given in Table 9.4. The input CMS file and trajectory directory must be speci-
fied; these have related names by default: myjob.cms and myjob_trj. The output files are
named with a base name and a sequential index.

Table 9.4. Options for the trajectory_extract_frame command.

Option Description

-h|--help Show usage message and exit.

-a asl| ASL expression to specify which part of the structure is to be extracted.
--asl=asl Not valid with -o. Default: all.
-b basename| Base name of the output structure file. The file name is constructed as
--basename=basename basename_index.outext, where index is a sequential index (not the
frame index). The default is the base name of the input .cms file.
-o outext | Format of the output structure file, given as the file extension. Valid val-
--output=outext ues are mae, cms and pdb. Default: mae.
-f frames | List of frames to extract. The list is comma-separated and can include
--frames=frames ranges specified as n:m. The default is to extract all frames.

Desmond 3.1 User Manual 101

 Chapter 9: Utilities

9.6 desmond_restraints.py
This command-line script allows you to add or remove distance, angle and dihedral angle
restraints to a CMS file for use in a Desmond simulation. The restraints can be flat-bottomed
(i.e. 0 for some range around the equilibrium value) if the width parameter is specified. The
syntax of the command is:

$SCHRODINGER/run -FROM desmond desmond_restraints.py [options]

-c restraint.cfg input-cms-file output-cmsfile
The options are given in Table 9.5.

Table 9.5. Options for the desmond_restraints.py script.

Option Description

-v[ersion] Show the program version and exit.

-h[elp] Show the usage message and exit.
-c input-file Specify the input config file that contains the restraint definitions.
-r rcut Cutoff distance for constraints. To correctly compute angle and improper restraint
interactions, each group of restrained particles should fit within a sphere of a given
radius, which can be set with this option. If they do not, Desmond halts with an
error. Default: 9.0 Å.
-d delete restraints

Example of adding restraints:

desmond_restraints.py -c restraint.cfg input_cms_file out_cms_file

Example of deleting restraints:

desmond_restraints.py -c restraint.cfg -d input_cms_file out_cms_file

The syntax of a restraint in the restraint config file is as follows:

[ ct atoms k s r0 ]
Here, ct is the index of the structure (CT) in the Maestro input file, atoms is 2, 3 or 4 atom
indices in the CT (2 for a distance, 3 for an angle, 4 for a dihedral), k is the force constant in the
same units as normal bonded interactions, s is the half-width of the flat bottom (optional) and
r0 is the reference value of the coordinate (optional).

102 Schrödinger Suite 2012 Update 2

 Chapter 9: Utilities

Example of a restraint config file:

# reference value (r0) is obtained from a current structure when

# reference value is not given.
# sigma value (s) is set to be zero if there is no input value.
ffio_stretch_fbhw = [
# ct atom1 atom2 k [s r0]
[2 1 2 0.1 ] # without reference value
[2 2 3 0.1 0.0 1.0] # with reference value
]
ffio_angle_fbhw = [
# ct atom1 atom2 atom3 k [s r0]
[2 2 3 40 0.1 ] # without reference value
[2 1 3 40 0.1 0.0 120.0] # with reference value
]
ffio_improper_fbhw = [
# ct atom1 atom2 atom3 atom4 k [s r0]
[2 2 3 4 5 0.1 ] # without reference value
[2 1 2 3 4 0.1 0.0 180.0] # with reference value
]

9.7 rebuild_cms.py
This utility creates a full model system from either the component CTs (solute, membrane, and
solvent) or a full system CT. To generate the full model system with all CTs from a set of
components, run the following command:

$SCHRODINGER/run -FROM desmond rebuild_cms.py -make_full_ct input.cms

output.cms
To generate the full model system from just the full system CT, run the following command:

$SCHRODINGER/run -FROM desmond rebuild_cms.py -make_comp_ct

[-solvent_asl ASL] [-membrane_asl ASL] input.mae output.cms
The solvent and the membrane are automatically detected, but you can specify them by using
ASL expressions. If the box vector is not present, the box size is estimated from the system
size.

Desmond 3.1 User Manual 103

 Chapter 9: Utilities

9.8 Other Scripts

In addition to the utilities described above, which are provided with the distribution, there are
some other useful scripts available from the Script Center on the Schrödinger web site, or in
the distribution, listed below. These should be run with $SCHRODINGER/run.

• trajectory_cluster—cluster frames from a trajectory (Script Center)

• trajectory_delete_water—create a new trajectory without any water molecules
present in the trajectory provided. (Script Center)
• predict_remd_temp.py—suggests temperatures for a replica exchange simulation
given a CMS file. (distribution)
You can run this script with the following command (without line breaks):
$SCHRODINGER/run $SCHRODINGER/mmshare-vversion/lib/Linux-x86_64/
lib/python2.7/site-packages/schrodinger/application/desmond/
predict_remd_temp.py
• assign_custom_charge.py—assign custom charges defined in a reference file to the
structure in a Maestro file. Intended for use with FEP simulations. (distribution)

104 Schrödinger Suite 2012 Update 2

 Desmond User Manual

Appendix A

Appendix A: The multisim Utility

The multisim utility is useful for running tasks that consist of a sequence of steps, such as
relaxation of a system followed by a production simulation. Maestro runs multisim for jobs
that include multiple stages. This appendix provides detailed information on how to run
multisim and the format of multisim (.msj) files.

Multisim is used by both Desmond and MCPRO+ for running jobs.

A.1 Running multisim

The basic usage information for multisim, including examples, is given in Section 6.2 on
page 70. This section describes more advanced multisim features such as template
multisim commands, node locking, restarting multisim jobs, and obtaining information
from multisim checkpoint files.

A.1.1 Template multisim Commands

Most .msj files produced by Maestro end with a commented out example of the command
needed to run the job. For example:

$SCHRODINGER/utilities/multisim -JOBNAME example -HOST master-job-host

-SUBHOST subjob-host -maxjob 0 -cpu "2 2 2" -i example.cms
-m example.msj -c example.cfg -o example-out.cms
To use this command you would replace the following text with the values for your own setup:

• master-job-host—the host on which to run the master job (the Python script that manages
the multisim job).
• subjob-host—the host on which to run the subjobs, which are often cpu intensive.
• example—the job name and the stem of various file names, which are constructed in a
standard way from the job name.
• the cpu specification "2 2 2"—the specification of the CPUs used by the job. This spec-
ification can either be a triplet of numbers that defines the spatial decomposition of the
system, with the total number of processors being the product of these three numbers; or
a single integer, which is the total number of processors, and must have only the factors 2,
3 and 5.

Desmond 3.1 User Manual 105

 Appendix A: The multisim Utility

• the value for -maxjob, which is used to specify how many subjobs can be queued at the
same time. The value 0 means “all subjobs” normally, but is the same as the value 1 for
umbrella mode.

A.1.2 Node Locking

Node locking involves requiring all subjobs to run on the same nodes (CPUs) as the master job.
Node locking can be turned on by adding -mode umbrella to the command line. This feature
can be useful on busy computer systems that are controlled by a queuing system, because only
the master process is submitted to the queuing system. All subjobs will then run on the nodes
allocated to the master process without being resubmitted to the queuing system. When this
option is given the -SUBHOST option is ignored.

Node locking is supported for FEP jobs. You can run these jobs in a serial queue if you use
–maxjob 1.

A.1.3 Restarting multisim Jobs

The multisim job periodically writes out a checkpoint file, which records the current state of
the workflow. The checkpoint file is named jobname-multisim_checkpoint. This file does
not include data produced by subjobs. The checkpoint file is copied back to the job launch
directory when the master job stops.

In most cases the multisim job can be restarted with a command similar to the following:

$SCHRODINGER/utilities/multisim -RESTART myjob-multisim_checkpoint

-d myjob_stage-out.tgz -JOBNAME myjob
where you specify the checkpoint file with the -RESTART option and the -out.tgz file of the
last completed stage with the -d option. The -d option is usually necessary because the subse-
quent stages need access to the data produced by the last successfully completed stage. For
example, if stage 5 is partially done (some subjobs finished, some did not), then you would
also have to include myjob_4-out.tgz, and the command to use would be:

$SCHRODINGER/utilities/multisim -RESTART myjob-multisim_checkpoint

-d myjob_5-out.tgz -d myjob_4-out.tgz -JOBNAME myjob
This command restarts the job on the same compute hosts as before. Unless the -HOST option
is given the master job runs on the local host. The subjobs will run using the same number of
CPUs as before. If you want to use different hosts to restart a job you can use the -HOST and
–SUBHOST options. The number of CPUs used by the subjobs can also be changed using the -
cpu option, for example,

106
 Appendix A: The multisim Utility

$SCHRODINGER/utilities/multisim -HOST another_master_host

-host another_subjob_host -RESTART myjob-multisim_checkpoint
-d myjob_5-out.tgz -JOBNAME myjob -cpu 1
You can modify the characteristics of some stages when you restart a job, by using the –set
option. For example, if you use the command:

$SCHRODINGER/utilities/multisim -RESTART myjob-multisim_checkpoint

-d myjob_4-out.tgz -JOBNAME myjob
-set "stage[5].time = 1200.0 stage[5].cutoff_radius = 16.0"
then stage 5 is modified so that its subjobs run with the time set to 1200.0, and the
cutoff_radius set to 16.0. However, if you also provide the stage 5 .tgz file and some subjobs
from stage 5 have finished, only the unfinished jobs will run with the new settings.

You can restart a job with a different .msj file. For instance, the command:

$SCHRODINGER/utilities/multisim -RESTART myjob-multisim_checkpoint

-d myjob_5-out.tgz -JOBNAME myjob -m newworkflow.msj
runs the uncompleted stages according to newworkflow.msj. This new .msj file must
contain the same completed stages as the original job. Any of the remaining stages can be
modified or deleted, and new stages may be inserted.

To restart a job from a completed stage, you can specify the stage number right after the check-
point file name. For instance, the command:

$SCHRODINGER/utilities/multisim -RESTART myjob-multisim_checkpoint:4

-d myjob_3-out.tgz -JOBNAME myjob
reruns the rest of the workflow starting from stage 4. The archive specified with -d must be the
one for the prior stage (which is stage 3 in this example).

Some other changes that you can or cannot make are:

• You can restart the job with a different .cfg file by specifying the new .cfg file with the
-c option.
• You can restart the job with a different maximum number of simultaneously running sub-
jobs by using the -maxjob option.
• You cannot change the node-locking mode.

In most cases, multisim automatically detects and uses additional input files needed for a job
when it is restarted. If the job fails because an existing input file was not detected when
restarting a job you can specify that file by using the -ADD_FILE option.

Desmond 3.1 User Manual 107

 Appendix A: The multisim Utility

A.1.4 Obtaining Information from multisim Checkpoint Files

To obtain information on the contents of a multisim_checkpoint file, you can use the
multisim command with the -probe option:

$SCHRODINGER/utilities/multisim -probe myjob-multisim_checkpoint

This command summarizes the contents of the checkpoint file without submitting a job to
continue the calculation. Below is an example of the output from a successful relative solvation
free energy calculation:
Probing checkpoint file: biphenyl_to_benzene-multisim_checkpoint
multisim version: 3.8.3.14
mmshare version: 19104
Jobname: biphenyl_to_benzene
Username: me
Master job host: rosaleen
Subjob host: myhost
CPUs per subjob: "1"
Original start time: Sat May 1 09:14:46 2010
Checkpoint time: Sat May 1 19:55:35 2010
Master job ID: rosaleen-0-4bdc2939
Structure input file: biphenyl_to_benzene_lig.mae
Original *.msj file: biphenyl_to_benzene_solvent.msj

Stages:
Stage 1 completed.
Stage 2 completed.
Stage 3 completed.
Stage 4 completed.
Stage 5 was skipped.
Stage 6 completed.
Stage 7 completed.
Stage 8 completed.

Current version of multisim is 3.8.3.15

This checkpoint file can be restarted with the current version of multisim.

Note the multisim version: checkpoint files generated by older versions of multisim might
not be compatible with the current version. The probe option of the multisim command can
be used detect whether the checkpoint file is supported.

If the job had failed in Stage 5 you would see something like:
...
Stage 5 partially completed. 1 subjobs failed, 11 subjobs done.
Jobname of failed subjobs:
biphenyl-to-benzene_5_lambda2
Stage 6 not run.

108
 Appendix A: The multisim Utility

To resume the job as described earlier, this output tells you that you need to provide the .tgz
file from stage 4, and that you might want to provide the .tgz file from the partially completed
stage 5. In this case only one subjob failed in stage 5 so rerunning just that subjob might
require significantly less computer time than rerunning the entire stage.

A.2 The multisim File Syntax

The multisim input file (.msj file) generally adheres to the Ark format from D. E. Shaw
Research. This is the same syntax that is used in Desmond configuration (.cfg) files, which
are described in brief in Appendix B and in full in the Desmond User’s Guide. Ark format is
completely compatible with the syntax used in multisim files in Schrödinger Suite 2008, but
supports other features.

The syntax rules of the Ark format can be summarized as follows.

• Values are assigned to a keyword with the keyword = value syntax.

• Values can be numbers, strings, lists, or blocks. Numbers can be integers or real numbers.
Strings do not need to be enclosed in quotes unless they contain embedded blanks or span
multiple lines.
• A block is one or more settings within a pair of braces, { }. For example: a = { b = 3
c = 4 }. The block in this example contains two keyword-value assignments.
• A list is a sequence values that is enclosed by a pair of square brackets [ ]. Elements of a
list do not have to be of the same type. An example is [ 1 { a = 1 } 5 [ 2 3 ] ].
• Hierarchical expressions involving blocks and lists can be created: there is no limit on
nesting of lists and blocks.
• Individual elements of a hierarchy can be set by joining the names of the parents with
periods to form a compound key. For example, a.b.c = 3 sets element c of block b in
block a to 3.
• If a keyword is assigned a value twice, the second takes precedence. If the value is a
block the blocks are merged and keywords present in each block are assigned the values
from the latter block.
• The = sign can be omitted if the value is a block value. This means that a = { b = 1 }
and a { b = 1 } are equivalent. This syntax is used to specify the multisim stages.

Multisim processing deviates from the Ark standard in the following ways:

• Multisim stages with the same name remain separate, otherwise stages that appear more
than once in an .msj file (such as simulate or minimize) would be combined into one
stage.

Desmond 3.1 User Manual 109

 Appendix A: The multisim Utility

The multisim input file consists of a sequence of stages, each of which specifies a particular
calculation to be run. A stage begins with a label identifying the type of stage followed by
braces enclosing parameters for that stage. A msj file will in general look something like:

#an outline for a msj file.

first_stage_name {
parameter1 = 3000.0
parameter2 = "this is a string"
parameter3 = [ "list element 1" 2 "list element 3" ]
}
second_stage_name{
parameter5 = 20
}
The types of stages that are supported for Desmond include the following:

• task—describes the type of job

• minimize—minimize the system
• simulate—run an MD simulation on the system
• system_builder—run the system builder
• replica_exchange—run a replica exchange MD simulation on the system
• lambda_hopping—run a lambda hopping simulation on the system
• solvate_pocket—add or remove water molecules in buried pockets
• analysis—analyze MD simulation trajectories
• fep_analysis—analyze FEP calculations
• extern—custom Python stage
• trim—remove unwanted stage data files
• stop—stop the workflow at this stage
The keywords supported for each of these stages and their default values are described in the
sections that follow.

110
 Appendix A: The multisim Utility

A.2.1 General Keywords

The keywords that can be used in any stage of the multisim job are listed in Table A.1.

Table A.1. General keywords that can be used in any stage.

Keyword Description

compress File name pattern of the stage data file. If it is set to an empty string, then the
data of this stage is not packaged and compressed. Default:
$JOBNAME_$STAGENO-out.tgz.
dir Pattern for the names of the directories used by subjobs of this stage. Default:
[$JOBPREFIX/][$PREFIX/]$MASTERJOBNAME_$STAGENO
[_lambda$LAMBDA]
jlaunch_opt Options to add to the command to run Desmond. The options are listed in
Table 6.1 on page 68. Example: ["-dp"] turns on use of double precision.
jobname Jobname pattern for subjobs of this stage. Default:
$MASTERJOBNAME_$STAGENO[_lambda$LAMBDA]
prefix Value of the PREFIX macro, which by default is used to specify the sub-direc-
tory name of subjobs (see the dir keyword). Default is an empty string.
should_skip Skip this stage. Allowed values: true, false. Default: false.
should_sync Do not start this stage until all subjobs of the previous stage finish successfully.
Allowed values: true, false. If it is set to false, this stage is started as soon as
any subjob of the previous stage finishes successfully. For FEP jobs, setting this
keyword to false can result in earlier completion of the job. Default: true.
struct_output File name of the final output structure file. This keyword is only effective when
set in the last stage, and the setting can be overwritten by the -o option of
multisim.
title The title for the stage. Default: ?. which stands for none.

A.2.2 Desmond-Specific Common Keywords

The stages that run Desmond directly (minimize, replica_exchange, lambda_hopping,
and simulate) accept a common set of keywords in addition to keywords specific to the
Desmond task. Keywords that are supported in the front-end config file (see Appendix B) can
be used directly in the corresponding stage.

Desmond 3.1 User Manual 111

 Appendix A: The multisim Utility

A.2.3 The restrain Keyword

The restrain keyword specifies the atom sets to be restrained. Restraints set in all stages
except the system_builder stage apply only to the current stage. Restraints set in the
system_builder stage are “permanent”—they are inherited by all subsequent stages. Perma-
nent restraints can be overridden in a particular stage, but only to increase the restraint force
constants. They will still be applied in subsequent stages.

The restrain keyword supports the specific values listed in Table A.2, and also supports a
block syntax and a list syntax.

Table A.2. Specific allowed values of the restrain keyword.

Value Description

heavy_atom restrain all heavy (non-hydrogen) atoms

none Remove the temporary restraints (those set in the previous simulate or
minimize stage), but do not change permanent restraints such as those set
in the system_builder stage or present in the original .cms file given to
the multisim job. If you use this value in a system_builder stage, it
removes all restraints. This is the default.
retain Keep the restraints as set in the previous stage.
solute Restrain all solute atoms
solute_heavy_atom Restrain all heavy atoms in the solute
solvent Restrain all solvent atoms
solvent_heavy_atom Restrain all heavy atoms in the solvent

Restraints can be specified using blocks, and ASL expressions can be used to define the atoms
that are restrained within restraint blocks. The syntax is:

restrain = {atom = atoms fc|force_constant = value

ref|reference_position = retain|reset|refvalue sigma = sigmavalue}
where atoms can be solute, solvent, solute_heavy_atom, or heavy_atom as before, an
atom list, or an ASL expression prefixed by asl:. For example,

restrain = {atom = asl:all force_constant = 10}

restrains all atoms with a force constant of 10 kcal mol-1 Å-2. Restraints can be applied with
respect to the positions of the atoms at the start of the current stage by setting
reference_position to reset or the positions from the previous stage by setting
reference_position to retain. The default is reference_position=reset.

112
 Appendix A: The multisim Utility

Relative restraints can be included by setting atoms and refvalue appropriately. For distance
and NOE restraints, atoms should specify exactly 2 atoms; for angle restraints, atoms should
specify exactly 3 atoms; and for dihedral restraints, atoms should specify exactly 4 atoms. All
atoms in the relative restraint must be in the same CT. For NOE constraints, refvalue should be
a list with two elements, the upper distance then the lower distance; for all other types refvalue
is the desired distance or angle value, in angstroms or degrees.

Flat-bottomed restraints can be included by setting sigmavalue to a distance or an angle toler-

ance, as appropriate. The restraint potential is set to zero in the range refvalue±sigmavalue. For
NOE restraints, sigmavalue can be a list with two elements, the width of the flat bottom, and
the coefficient of the linear term in the restraint potential at large distances.

Different groups of atoms in the system can be restrained with different force constants by
listing the restraint blocks for each set of atoms within a list using the syntax:

restrain = [{atom = "asl:ASL1" force_constant = value1}

{atom = "asl:ASL2" force_constant = value2} ... ]
If the ASL expression contains quotes, they must be escaped, either by using a backslash, or by
using double quotes in the ASL expression and surrounding the entire expression in single
quotes. You can include anisotropic constraints by assigning the force constant in any of these
expressions as follows:

force_constant = [x-value y-value z-value]

A.2.4 The atom_group Keyword

The atom_group keyword can be used to define atoms groups within the .cms file. Atom
groups can be frozen, restrained or associated with particular thermostats. The atom group is
defined by the atom-level property i_ffio_grp_name. This keyword can be set to the
following values:

• none—Remove all atom groups.

• retain—Keep all atom groups from the previous stage.
• { atom = atoms index = i name = name }—atom group block. Put the specified
atoms in the atom group i that has the name name. The atom keyword accepts the same
values for atoms as the atom keyword for restrain, described above. The index i is the
value for the i_ffio_grp_name property. Desmond only supports index numbers from
0 to 7.
• [ group1 group2 ... ]—Specify multiple atom groups. Each group can be specified as
a block, in the above format.

Desmond 3.1 User Manual 113

 Appendix A: The multisim Utility

An atom group set in the system_builder stage is persistent, which means that it remains
defined in all subsequent stages, until the next system_builder stage.

A.2.5 The task Stage

The .msj file should start with a task stage to specify the type of job. Although Multisim can
determine the type of job based on the input structure file provided, this can lead to unpredict-
able behavior if the file was previously used by a different type of job. Explicitly stating the job
type avoids this problem, and also permits Multisim to ensure that the input structure file
provided is appropriate for the type of calculation requested.

The task stage has two keywords. The first is task, whose allowed values are listed in
Table A.3.

Table A.3. Values of the task keyword.

Value Description

desmond:regular Non-FEP Desmond job

desmond:fep Desmond absolute or relative free energy (FEP) job
desmond:afep Desmond absolute free energy (FEP) job
desmond:auto Desmond job whose type is determined from the input structure file
mcpro:auto MCPRO+ job whose type is determined from the input structure file
mcpro:fep MCPRO+ FEP job

The second is set_family which can be used to sets parameters for a family of stages. For
example:

task {
task = "desmond:fep"
set_family = {
desmond = {
checkpt.wall_interval = 7200.0
checkpt.write_last_step = no
}
}
}
The family of stages here is desmond. This family includes minimize, simulate,
replica_exchange, and lambda_hopping stages. The settings in the desmond = {...}
block are effective for all subsequent stages belonging to the desmond family. If there is a

114
 Appendix A: The multisim Utility

conflict between the settings in the set_family block and the explicit settings in a stage, the
latter take precedence. For example:

task {
task = "desmond:fep"
set_family = {
desmond = {
checkpt.wall_interval = 7200.0
checkpt.write_last_step = no
}
}
}
simulate {
checkpt.wall_interval = inf
}
The value of the checkpt.wall_interval parameter is set to inf, overriding the setting
checkpt.wall_interval = 7200.0 within the set_family block, because explicit
settings in the stage have higher precedence.

Valid stage family names include desmond and all stage names. If a stage name is used, the
family consists of all stages of that name. For example:

task {
task = "desmond:fep"
set_family = {
simulate = {
checkpt.wall_interval = 7200.0
checkpt.write_last_step = no
}
}
}
Here all subsequent simulate stages will be affected, but minimize, replica_exchange,
and lambda_hopping stages will not.

Desmond 3.1 User Manual 115

 Appendix A: The multisim Utility

A.2.6 The system_builder Stage

Keywords for the system_builder stage are listed in Table A.4. The keywords restrain
(Section A.2.3 on page 112) and atom_group (Section A.2.4 on page 113) can also be used in
this stage. When they are used, the settings are persistent: they apply to all subsequent stages
unless explicitly overridden by another system_builder stage.

Table A.4. Keywords for the system_builder stage.

Keyword Description

assign_forcefield Reassign the force field.If set to false, the system_builder

stage can be used to only rebuild the permanent restraints.
Default: true.
box_shape Specifies the box shape. Allowed values are cubic,
orthorhomibic, triclinic. Default: cubic.
buffer_width Minimum distance between the box edge and solute in Ang-
stroms. Default: 10.0 Å.
csb_file Input composite system builder (.csb) file. This option is not
required and is usually used to customize system building beyond
the keywords supported by multisim.
distil_solute Move water molecules from the solute entry (CT) to a separate
entry (CT). Allowed values are true, false, For non-aqueous
solvents, use 'asl:ASL-expression'to specify the solvent mole-
cule. Default: true.
ion_awaydistance Minimum distance between the added ions and the atoms that the
ions should not be placed near, specified by ion_awayfrom.
Valid values are non-negative real numbers. Default: 10.0
ion_awayfrom Do not place ions close to the specified atoms. Valid value is a list
of atom indices. Default: no atoms.
minimize_volume Reorient the solute to minimize the volume of the simulation box
when adding solvent molecules. Default: false.
neutralize_system Add counter ions to neutralize the system. Allowed values are
true, false. Default: true.
rezero_system Reset the origin of the coordinates to the center of mass of the
solutes. Allowed values are true, false. Default: true.
solvate_system Solvate the system. Allowed values are true, false.
Default: true.
solvent Solvent type. Allowed values are SPC, TIP3P, TIP4P, and
TIP4PEW. Default: SPC.

116
 Appendix A: The multisim Utility

A.2.7 The build_geometry Stage

This stage can be used to build the geometry of a simulation system from its components. It
provides a greater degree of flexibility in building a system than the system_builder stage, but
it only operates on the geometry, not the force field. All keywords for the system_builder stage
are accepted except for assign_forcefield, and the solvent keyword has more values,
listed below. The keywords specific to this stage are given in Table A.5.

Table A.5. Keywords for the build_geometry stage.

Keyword Description

only_merge_ct Merge the specified CTs into a single CT, and ignore all other settings
for the stage. This is useful for building a full system CT from the sepa-
rate CTs. Valid value is a list of CT indices. Default: [].
preserve_box When building the geometry of the system, don't change the box size if
box information exists in the given solute CT. Valid values are true and
false. Default: false.
preserve_ffio When building the geometry of the system, do not delete the existing
ffio_ff blocks in the original CT. Valid values are true and false.
Default: true.
rebuild_cms Extract the component CTs from a full system CT. All other settings are
ignored. Valid values are true, false, or a block like:
{ membrane = ASL-expression | ""
solvent = ASL-expression | ""
}
If the value is not false, the component CTs are extracted from the full
system CT. The block value allows you to you can specify which atoms
belong to the membrane component CT and which belong to the solvent
component CT. Default: false.
solvent Solvent type. Allowed values are water, SPC, TIP3P, TIP4P, TIP4PEW,
methanol, octanol, DMSO. Values are case sensitive. water is a syn-
onym for SPC. Default: water.

Desmond 3.1 User Manual 117

 Appendix A: The multisim Utility

A.2.8 The assign_forcefield Stage

This stage can be used to assign the force field to a model system. You can either keep or
replace existing assignments. The keywords are given in Table A.6.

Table A.6. Keywords for the assign_forcefield stage

Keyword Description

forcefield Specify the force field to use. Valid values are: OPLS_2005, CHARMM, AMBER,
amber03, amber99, amber94, amber96, amber99SB, amber99SB-ILDN,
charmm22nocmap, charmm36_lipids, charmm27, charmm32,
oplsaa_ions_Jensen_2006, oplsaa_impact_2001,
oplsaa_impact_2005, oplsaa_impact_2001, oplsaa_impact_2005.
These force fields are the ones available from viparr—see Chapter 8. Values are
case-sensitive. Default: OPLS_2005.
water Specify the water model to use. Valid values are: SPC, SPCE, TIP3P,
TIP3P_CHARMM, TIP4P, TIP4PEW, TIP4P2005, TIP5P, none. Default: SPC
humble Do not overwrite the existing ffio_ff block in the input. Valid values are true,
false. Default: false.

A.2.9 The simulate, replica_exchange, and lambda_hopping Stages

The keywords that are specific to the simulate, replica_exchange, and
lambda_hopping stages are listed in Table A.7. In addition, the keywords as described in
Section B.6 on page 140 can be used in the replica_exchange stage. The keywords
restrain (Section A.2.3 on page 112) and atom_group (Section A.2.4 on page 113) can
also be used in this stage. When they are used, the settings are temporary: they apply only to
the current stage.

Table A.7. Keywords for the simulate, replica_exchange, and lambda_hopping stages.

Keyword Description

fep.model_file List of structure files, each for use with different lambda windows.
This keyword should only be used for FEP jobs.
jin_file List of auxiliary input files for the stage. Usually only needed when
custom plugins are used.
jout List of auxiliary output files for the stage. Usually only needed
when custom plugins are used.

The fep.model_file keyword makes it possible to use different input files for different
lambda windows. For instance:

118
 Appendix A: The multisim Utility

task {
task = "desmond:fep"
}

simulate {
fep.model_file = ["file1.cms" "file2.cms"]
}
Here, the first two windows use the file1.cms and file2.cms respectively. The rest of the
windows use the default input file.

Metadynamics simulations can be run by including a meta block in the simulate stage. See
Section B.7 on page 141 for the syntax of this block. However, in the multisim file, you must
use an ASL expression for the center of mass of a set of atoms, not a list. For example,

atom = ['atom.num 2379 2380 2382 2384 2385' 'atom.num 631 642 651']

A.2.10 The minimize Stage

The keywords that are specific to the minimize stage are described in Section B.5 on page 139.
The keyword restrain (Section A.2.3 on page 112) can also be used in this stage. When it is
used, the settings are temporary: they apply only to the current stage.

A.2.11 The solvate_pocket Stage

The keywords for the solvate_pocket stage are listed in Table A.8. The only keywords that
can be set directly within the solvate_pocket stage are spd_file, ligand_file,
fep.lambda, and backend. The other keywords in the table should be set in the backend
block. This stage runs the solvate_pocket utility, which is described in Section 9.1 on
page 93.

Table A.8. Keywords for the solvate_pocket stage

Keyword Description

spd_file The name of the solvate_pocket command file. If omitted, the default
settings are used.
spd_overwrite Block that provides settings in Ark syntax to overwrite specific commands
in the command file. All keywords other than spd_file and
ligand_file must be inside this block.
ligand_file The name of the ligand file used to define the region sampled. If the name
is set to an empty string or omitted, the name jobname-ligand.mae is
used. A ligand must be used to define the region.

Desmond 3.1 User Manual 119

 Appendix A: The multisim Utility

Table A.8. Keywords for the solvate_pocket stage

Keyword Description

name A descriptive name for the simulation. Default: name of standard

solvate_pocket command file.
temperature The temperature used in the Monte Carlo simulation. Default: 300 K
init_num_passes The number of Monte Carlo passes to perform in the equilibration prior to
the production Monte Carlo run. Default: 10000.
num_passes The number of Monte Carlo passes to perform in the production Monte
Carlo run. Default: 100000.
update_frequency The number of passes between updates in the log file. Default: 100.
pass_term_window The number of passes over which the slope of the standard deviation of the
the number of water molecules is calculated. This keyword may be used to
terminate the calculations before the number of passes specified by
num_passes has been completed. The simulation portions will run for at
least twice this duration before early termination occurs. Default: 100000.
term_std_slope Threshold for the standard deviation of the number of water molecules.
The calculation is terminated if the standard deviation falls below this
value. Default: 0.00001.
num_trans_rot The number of water molecule translations to attempt per pass. This
should be set to approximately the number of water molecules expected to
reside in the region being sampled. Default: 100.
num_delete The number of attempts to delete a single water molecule per pass.
Default: 25.
num_insert The number of attempts to insert a single water molecule per pass.
Default: 25.
max_disp The maximum change used in each of the x, y, and z directions for a com-
bined translation/rotation move. Default: 0.105 Å.
max_dpsi The maximum change used for phi and psi (Euler angles) in radians for a
combined translation/rotation move. Default: 0.318.
max_dctheta The maximum change used for the cosine of theta (Euler angle) for a com-
bined translation/rotation move, in radians. Default: 0.0654.
cut_off The cutoff distance for calculating electrostatic and Lennard-Jones inter-
actions. Default: 9.0 Å.

120
 Appendix A: The multisim Utility

Table A.8. Keywords for the solvate_pocket stage

Keyword Description

short_dist The closest acceptable approach for any two atoms. Default: 1.0 Å.
chemical_potential The chemical potential of water in kcal/mol. Default: -7.17 kcal/mol (for
TIP4P).
fep.lambda For FEP calculations, turn on the solvate_pocket calculation and spec-
ify the lambda schedule that is in use. This keyword should only be used
for an FEP job, and the schedule must be the same as for the FEP calcula-
tion itself (as specified in the config file). The value true specifies the
default:12 schedule, which is the default for FEP calculations. See
Appendix B for more information.

A.2.12 The extern Stage

The extern stage provides an extremely flexible way to include your own Python code in a
multisim run. The code can be embedded in the .msj file as a string value assigned to the
command keyword. For example:

extern {
command = "
import os;
def main( current_stage, job ):
os.system( ’ls’ )
"
}
In the embedded Python code, you can import and use modules from your Schrödinger Python
installation. If you need extra modules, you can pass their file names to multisim by setting
the auxiliary_file parameter, and multisim transfers them to the scratch directory of the
master job at run time. For example:

extern {
auxiliary_file = [mod1.py mod2.py]
command = "
import os
import mod1
import mod2
def main( current_stage, job ):
// does something with mod1 and mod2
os.system( ’ls’ )
"
}

Desmond 3.1 User Manual 121

 Appendix A: The multisim Utility

The code given above is run once for each subjob in the previous stage. If you want to run it
only once, use command_once instead of command. For example:

extern {
auxiliary_file = [mod1.py mod2.py]
command_once = "
import os
import mod1
import mod2
def main( current_stage ):
// does something with mod1 and mod2
os.system( ’ls’ )
"
}
For scripts that are stage-specific, your code must provide a main function that takes two argu-
ments for command or one for command_once. The first argument in both cases corresponds
to information for the current stage, while the second argument for command corresponds to
information from the previous stage.

For very simple scripts, your code for command or command_once does not need to provide a
main function. The following example removes a temporary file if it exists:

extern {
command = "
import os
# Removes a temporary file.
if (os.path.isfile( ’my_temporary_file’ )) :
os.remove( ’my_temporary_file’ )
"
}
Without the main function, multisim cannot pass the current stage and the current job objects
to the Python code, but that is presumed to be not needed for simple operations.

The extern stage is an advanced feature. Please do not hesitate to contact us for additional
information on its use. The keywords for this stage are listed in Table A.9.

122
 Appendix A: The multisim Utility

Table A.9. Keywords for the extern stage.

Keyword Description

auxiliary_file List of files containing extra modules to be transferred to the runtime direc-
tory.
command Command to execute once for each subjob of the previous stage. The com-
mand specifies Python code that can span multiple lines.
command_once Command to execute once for the previous stage. The command specifies
Python code that can span multiple lines.

A.2.13 The analysis Stage

Keywords for the analysis stage are listed in Table A.11.

Table A.10. Keywords for the analysis stage.

Keywords Description
meta.range For metadynamics simulations, specify the range for each collective
variable, in the format [[min1 max1] [min2 max2] ... ].
meta.nbin For metadynamics simulations, specify the number of bins for each col-
lective variable in the analysis.
plotter Specify the plotter to use to draw images for figures. Allowed values:
mplchart, gchart. mplchart generates static .png file for each fig-
ure. gchart generates a URL for each figure, and the URLs are saved in
the report file. When you open the report file in a browser, the figure
images are generated using Google chart APIs. The data for the figures
is sent to the Google web site to create the images.
Default: mplchart.
prob_profile Specify how to calculate the probability profiles for the time series given
by the time_series parameter. Example: prob_profile = [2 0
360 true], which means the bin size is 2, the data range is [0, 360),
and the range is periodic (true).
report Specify the name of the report file. Default: $JOBNAME_report.html.
SEA Block for specifying the configuration of the simulation event analysis.
This block can be the contents of an .st2 file, which can be written
from the Simulation Event Analysis panel.
time_series Specify the variables for which to analyze the time series. Example:
time_series = [[dihedral 1 2 3 4] [dihedral 5 6 7 8]],
which specifies two dihedral angles to analyze.

Desmond 3.1 User Manual 123

 Appendix A: The multisim Utility

A.2.14 The fep_analysis Stage

Keywords for the fep_analysis stage are listed in Table A.11.

Table A.11. Keywords for the fep_analysis stage.

Keywords Description

bennett.random_seed Random seed for the Bennett method. Default: 2111839.

correct_vdw Calculate the long range dispersion correction for absolute free energy
jobs. This keyword has no effect for relative free energy jobs. Default:
true.
correct_restr Calculate the enthalpic correction for position restraints. This parameter
does no harm if the production simulations have no restraints. Default:
true.
plotter Specify the plotter to use to draw images for figures. Allowed values:
mplchart, gchart. mplchart generates static .png file for each fig-
ure. Choosing gchart generates a URL for each figure, and the URLs
are saved in the report file. When you open the report file in a browser,
the figure images are generated using Google chart APIs. The data for
the figures is sent to the Google web site to create the images.
Default: mplchart.
report Specify the name of the report file. Default: $JOBNAME_report.html.

The free energy can be plotted as different functions of the simulation time. The following
settings are added to control these new functionalities:

bennett = {
forward_time = {
name = "freeenergy_time"
begin = 100.0
end = inf
dt = 30.0
}
reversed_time = {
name = "freeenergy_rtime"
begin = 100.0
end = inf
dt = 30.0
}
sliding_time = {
name = "freeenergy_stime"
begin = 100.0

124
 Appendix A: The multisim Utility

end = inf
dt = 30.0
window = 500.0
}
}
The forward_time, reversed_time, and sliding_time settings control calculations of
the free energy over different ranges of time.

The forward_time keyword provides the free energy as a function of the simulation course.
In other words, each free energy value of this function is calculated using the trajectory frag-
ment from the time as given by the begin setting to the time point. The dt keyword indicates
the interval between consecutive estimates of the free energy (in this case, 30 ps).

The reversed_time keyword functions similarly to the forward_time keyword except that
the free energy is calculated from each time point to the end of the trajectory.

For the sliding_time keyword, the fragment length has a fixed length, specified by the
window keyword (in this case, 500.0 ps). The beginning of the fragment slides along the trajec-
tory with the step size given by the dt keyword (30.0 ps in this example).

A.2.15 The vrun and fep_vrun Stages

This stage runs vrun with the Desmond trajectory from the last simulate stage. The vrun
stage should only be used with non-FEP simulations. For FEP simulations, you should use the
fep_vrun stage, which has the same syntax. All the parameters from the simulate stage can be
used with these stages, but the default settings for the following keywords are different:

trajectory = false
checkpt = off
maeff_output = off
eneseq.interval = 0.0
Setting eneseq.interval to zero means that the energy is computed for every frame of the
trajectory.

A.2.16 The trim Stage

The purpose of the trim stage is to automatically delete unwanted stage data files. A common
scenario is that most of the relaxation stage files are not needed if the job is successful. Adding
a properly-set trim stage at the end of the job automatically removes the these files before they
are transferred back to the launch directory. The following example illustrates how a trim stage
might be used:

Desmond 3.1 User Manual 125

 Appendix A: The multisim Utility

task {
task = "desmond:auto"
}
minimize {
# relaxation stage
}
simulate {
# relaxation stage
}
simulate {
# production stage
}
trim {
save = [-1]
}
The save keyword is the only trim-specific keyword and specifies stages for which the output
files should be kept. Its value is a list of integers. The integers can be either positive or nega-
tive. Positive integers are the stage index, negative integers are relative to the trim stage, i.e.
they refer to stages counting backwards from the trim stage. In this example, -1 means the
first stage before this trim stage, i.e. the production stage or the 4th stage. So the trim stage
removes all .tgz files that were generated by the earlier stages except for the one generated by
the production stage.

The following setting for the above example saves the .tgz files generated by the 3rd and 4th
stages (the two stages before the trim stage):

save = [-1 -2]

The following setting saves the .tgz files generated by the 2nd, 3rd, and 4th stages

save = [2 3 -1]
Note that the task counts as a stage.

126
 Desmond User Manual

Appendix B

Appendix B: The Configuration File

The configuration file (hereafter referred to as a config file) describes the nature of the calcula-
tion that the Desmond backend should perform. Starting with the 2010 release, there are two
styles of config files: a user facing, “front-end” style that is designed to make it easier for users
to work with, and a “back-end” style, which provides a very detailed and specific description
of the calculation to the back-end itself.

For most Desmond jobs a CMS file and a front-end config file are specified on the command
line. Then the software automatically translates the front-end config file into a back-end config
file in a manner that depends on, or rather, takes into account, the nature of the system (in the
cms file). The appropriate Desmond back-end executable for the job (e.g. mdsim, minimizer)
is determined automatically based upon the contents of the front-end config file. This appendix
is mainly focused on describing the syntax of the front-end config file. For information on the
back-end config file see the Desmond User’s Guide.

The Ark syntax from D. E. Shaw Research is used for both the back-end and front-end config
files. The definitive description of the Ark format is available in the Desmond User’s Guide.
This format is also used by multisim (.msj) files. Below is a brief recap of the Ark syntax,
followed by detailed documentation of the parameters in the front-end config file.

B.1 Ark Format Syntax Summary

Ark uses a free format syntax, in that is there are no indentation or alignment requirements.
Spaces (including some control characters, e.g., tab, return, new line.) are useful only for sepa-
rating keys and values. The Ark format syntax is case-sensitive, so keys and values with
different capitalization patterns are treated as distinct.

At the most basic level the syntax is the familiar key-value pair syntax of the form:

key = value
where key must be a non-empty string, and value must be one of the following three objects:
atom, list, and map. A value of the atom type is one of the following: integer, floating number,
string, boolean (true, false, yes, no, on, off are recognized as boolean values), or none (?
is recognized as none). A value of the list of type is of the form: [ value1 value2 ... ], in
other words a list of one or more values wrapped within brackets [ ], and the elements value1,
value2, ... are values of any type and are separated by spaces. Map type values generally look
like:

Desmond 3.1 User Manual 127

 Appendix B: The Configuration File

{ key1 = value1
key2 = value2
...
}
where the value is wrapped with braces { }, and elements in the map are key-value pairs.

The form describe above is the canonical form of the syntax. Here is an example of this syntax
with nested values:

a = {
b = {
c = [1 2 4]
}
d = string_value
}
Ark also supports the so-called “pathname” syntax, where nested keys are separated by a
period. The pathname equivalent of the canonical example given above is:

a.b.c = [1 2 4]
a.d = string_value

B.2 Units
The units used in the configuration file are:

• time: ps
• distance: angstroms
• energy: kcal/mol
• pressure: bar
• surface tension: bar angstroms
• temperature: kelvin
Time can be specified as a positive real number or as the string inf, meaning infinity, or
“never”.

B.3 Macros in String Values

A macro works as a place-holder, and is replaced by a corresponding value after macro expan-
sion is performed (during conversion of the front-end config file into the back-end config file).
The name of a macro starts with a $ character, followed by a sequence of alphanumeric charac-
ters. A macro can be expanded only when its value is defined. If its value is undefined, the
macro is left in the string unchanged.

128
 Appendix B: The Configuration File

For Desmond, the supported macros are listed in Table B.1.

Table B.1. Supported macros

Macro Name Value defined Value

$JOBNAME Always The actual name of the job

$MASTERJOBNAME Only in Multisim The actual name of the master job
$REPLICA Only for REMD jobs. The index of the replica. (Index starts from 0.)

Optional string fragments that contain macros can also be defined. An optional fragment is a
portion of a string enclosed by $[ and $]. Such a fragment, together with the $[ and $] char-
acters, is deleted if there are one or more $ characters within the fragment. If there is no $ char-
acter within the fragment, only the $[ and $] are deleted. This operation is performed after
macro expansion. To see its effect, consider the following example.

name = "$JOBNAME$[_replica$REPLICA$].dE"
If we have defined the values of $JOBNAME and $REPLICA to be myjob and 0, respectively,
then the actual value for name will be myjob_replica0.dE. If we do not give a value to
$REPLICA, macro expansion yields myjob$[_replica$REPLICA$].dE. Because there is a
$ within the fragment, this fragment is deleted, so that the final value for name is myjob.dE.

B.4 Common Front-End Parameters

All currently supported parameters are listed and explained below. Each parameter is in a sepa-
rate subsection.

B.4.1 fep
Common usage:

fep = {
lambda = "default:12"
i_window = ?
output = {
name = "$JOBNAME$[_replica$REPLICA$].dE"
first = 0.0
interval = 1.2
}
}

Desmond 3.1 User Manual 129

 Appendix B: The Configuration File

B.4.1.1 fep.lambda

This parameter sets the lambda schedule for FEP calculations. You can set it to a string like
default:12, which means “use the default scheme with 12 windows”. To use 20 windows
with the default scheme, set the parameter to default:20. Available lambda schedule
schemes are: default, quickcharge, and superquickcharge. You can define your own
schedule. For instance, for an alchemical free energy calculation, the following syntax can be
used:

fep.lambda = {
bondedA = [1.0 0.9 ...]
bondedB = [0.0 0.1 ...]
chargeA = [1.0 0.75 ...]
chargeB = [0.0 0.25 ...]
vdwA = [1.0 1.0 ...]
vdwB = [0.0 0.0 ...]
}
Here, A refers to the original molecule while B refers to the mutated molecule, and the ellipsis
represents additional values that you would explicitly include. The number of values (lambda
stages) must be consistent in all 6 lists.

Similarly for a total free energy calculation the following syntax may be used:

fep.lambda = {
coulomb = [0.0 0.1 ...]
vdw = [1.0 1.0 ...]
}

B.4.1.2 fep.i_window

This parameter is automatically set by multisim to the index of the lambda window. You do
not normally set or adjust this parameter.

B.4.1.3 Default Lambda Schedules

The default 12 step lambda schedule for relative free energy calculations is:
lambda = {
bondedA = [1.0 0.916666666667 0.833333333333 0.75 0.666666666667 0.583333333333
0.416666666667 0.333333333333 0.25 0.166666666667 0.0833333333333 0.0]
bondedB = [0.0 0.0833333333333 0.166666666667 0.25 0.333333333333 0.416666666667
0.583333333333 0.666666666667 0.75 0.833333333333 0.916666666667 1.0]
chargeA = [1.0 0.75 0.5 0.25 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0]
chargeB = [0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.25 0.5 0.75 1.0]
vdwA = [1.0 1.0 1.0 1.0 1.0 0.67 0.46 0.33 0.25 0.19 0.12 0.0]

130
 Appendix B: The Configuration File

vdwB = [0.0 0.12 0.19 0.25 0.33 0.46 0.67 1.0 1.0 1.0 1.0 1.0]
}

The default 12 step lambda schedule for absolute free energy calculations is:
lambda = {
coulomb = [0.0 0.118514957434 0.189778434985 0.24741405481 0.325045439067
0.456296209913 0.674789989504 1.0 1.0 1.0 1.0 1.0 ]
vdw = [1.0 1.0 1.0 1.0 1.0 0.674789989504 0.456296209913 0.325045439067
0.24741405481 0.189778434985 0.118514957434 0.0 ]
}

B.4.2 cutoff_radius
This parameter sets the cutoff radius for the non-bonded interactions. If the particle mesh
Ewald (PME) method is used for electrostatic interactions, this sets the cutoff radius for the
real space part of the electrostatic interaction calculations.

cutoff_radius = 9.0

B.4.3 taper
The taper parameter is used to specify how interactions are truncated at the cutoff. For
example,

taper = off
turns off any special treatment of interactions at the cutoff (i.e. they are just truncated). There
are several ways to turn on tapering. The simplest way is to set the taper parameter to on, and
the default, shift, tapering scheme will be used. In this scheme the tapering consists of
shifting the potential to 0 at the cutoff_radius. More elaborate tapering schemes, requiring
more detailed information, are supported:

taper = {
method = potential
width = 1.0
}
where the method parameter specifies the type of tapering and has one of the following values:
potential, c1switch, c2switch, or shift. The width parameter controls the range of
distances, between (cutoff_radius - width) and cutoff_radius, over which tapering is
applied. For more information on tapering methods, see the Desmond User’s Guide.

Desmond 3.1 User Manual 131

 Appendix B: The Configuration File

B.4.4 coulomb_method
This parameter sets the method for treating Coulombic interactions in the simulation. For
example,

coulomb_method = pme
requests that the particle mesh Ewald (PME) method be used.

Supported values are: pme, cutoff.

The computational accuracy of the PME method can be controlled by the so-called Ewald
tolerance. The smaller the tolerance is, the more accurate but slower the computation is. The
default value of the tolerance is 1E-9, which is very accurate for most simulations. If you want
to use a different tolerance value, set the coulomb_method parameter explicitly in the
following form:

coulomb_method = [pme 1E-9]

where the first element of the list value must be pme, and the second element is the desired
tolerance value.

B.4.5 temperature
This parameter sets the temperature of the system. For instance:

temperature = 300.0
requests that a temperature of 300K be used. For multiple thermostats, you must explicitly set
the temperature for each of the thermostats using the following syntax:

temperature = [[temp1 group-index1] [temp2 group-index2] ...]

where the value of the temperature parameter is a list. Each element of the list is a pair, in
which the first value of the pair is the temperature, and the second value is the index of the
atom group that should have this temperature.

For simulated annealing jobs, the value of the temperature parameter specifies the temperature
schedule of the annealing process in the following form:

temperature = [[temp1 time-point1] [temp2 time-point2] ...]

where the value of the temperature parameter is a list. Each element of the list is a pair, in
which the first value of the pair is the reference temperature, the second value is the time point.
Between time-point1 and time-point2, the value of the reference temperature is linearly inter-
polated from temp1 to temp2, and so on for the remaining points.

132
 Appendix B: The Configuration File

B.4.6 annealing
The annealing parameter indicates whether this is a simulated annealing simulation, and takes
values of off (not a simulated annealing simulation) or on. If annealing is set to on then the
temperature parameter should be set to an appropriate temperature schedule.

B.4.7 pressure
The target pressure for simulations involving a barostat can be set as follows:

pressure = 1.01325
The default coupling scheme of the barostat to the system is isotropic. Use the following
format to use another coupling scheme:

pressure = [1.01325 anisotropic]

where the second element of the list value sets the coupling style. The valid values for coupling
style are the following: isotropic, anisotropic.

B.4.8 surface_tension
This parameter sets the surface tension, e.g.

surface_tension = 4E3
It is ignored unless the ensemble parameter is set to NPgT (or the ensemble.method param-
eter is set to NPgT).

B.4.9 ensemble
This parameter is used to select the ensemble to use for the simulation. For instance,

ensemble = NPT
requests that the simulation be done in the constant pressure / constant temperature ensemble
with a constant number of particles (atoms).

In the simplest form, you can set the parameter to different ensemble classes, and the default
setup for that class is used. Valid values for ensemble class are the following: NPT, NVT, NVE,
NPgT, NPAT, NPT_Ber, NVT_Ber, NPT_L, NVT_L. In all of these ensembles the number of
particles is constant. NPgT represents the ensemble where the pressure, temperature, and
surface tension (x-y plane) are also held constant. In the NPAT ensemble, in addition to the
number of particles, the pressure, temperature, and area of the system (in the xy plane) are held
constant.

Desmond 3.1 User Manual 133

 Appendix B: The Configuration File

The ensembles represented by NPT, NVT, NPgT, and NPAT are controlled by a Nosé-Hoover-
based algorithm. Those ending with _Ber or _L use alternative algorithms, namely those based
upon the work of Berendsen or Langevin.

The ensemble can be described in more detail using the following form:

ensemble = {
class = NPT
method = MTK
thermostat.tau = 1.0
barostat.tau = 2.0
}
where the valid values for class are: NPT, NPgT, NPAT, NVT, and for method are: MTK, NH,
Berendsen, Langevin. The thermostat.tau and barostat.tau parameters set the relax-
ation times for the thermostat and barostat, respectively.

For the NVE ensemble, the ensemble parameter can be set as:

ensemble = NVE
or as:

ensemble.class = NVE

B.4.10 time
The time parameter can be used to set the total simulation time. e.g.,

time = 1200.0

B.4.11 elapsed_time
The elapsed_time parameter controls the starting time for the simulation. Normally this is 0.
However, if the simulation is a continuation of an earlier one the current time can be set using
this parameter, e.g.

elapsed_time = 10.0

B.4.12 timestep
The timestep parameter specifies values for the bonded, near, and far time steps, respec-
tively. e.g.,

timestep = [0.002 0.002 0.006]

134
 Appendix B: The Configuration File

B.4.13 cpu
The cpu parameter specifies the total number of processors and, optionally, the domain-
decomposition of the system in the x, y, and z directions. For instance,

cpu = [1 2 4]
species that 1 × 2 × 4 = 8 processors should be used and that the system should be decomposed
into to 1, 2 and 4 domains in the x, y, and z directions. Alternatively, you can set this parameter
to a single integer value, which specifies the total number of CPUs, and then the decomposition
is automatically done based on the shape of the simulation box for the system. For example:

cpu = 8
sets the total number of CPUs to 8. The values specified for the domain decomposition must be
powers of 2, 3, or 5, or products of these powers.

B.4.14 glue
Glue tries to ensure that the closest images for closely associated molecules are recorded in the
output cms and trajectory files. This parameter specifies the molecule set. Only one value is
supported at present: solute. The parameter can be set with:

glue = solute

B.4.15 trajectory
The trajectory parameter is used to control how the trajectory is written.

trajectory = {
name = "$JOBNAME$[_replica$REPLICA$]_trj"
first = 0.0
interval = 4.8
periodicfix = true
write_velocity = true
frames_per_file = 25
}
The periodicfix parameter, if set to true, instructs Desmond to wrap molecules as a unit
within the periodic boundary conditions rather than to wrap individual atoms (potentially split-
ting molecules across the unit cell in the recorded trajectory). The first parameter is used to
indicate at what time to start recording the trajectory, while interval controls how often to
write trajectory frames. The frames_per_file parameter is used to control the number of
frames written to each file in the trajectory. Large numbers of files can cause very slow
copying and reading of the trajectory, and can even cause failure. You should set this parameter

Desmond 3.1 User Manual 135

 Appendix B: The Configuration File

so that the total number of files is small enough for the file system to handle efficiently (no
more than a few thousand). Fewer files also reduces the IO overhead when reading the files.

B.4.16 eneseq
The eneseq parameter controls when to start recording the .ene file (first), how often to
update it (interval), and the precision of the energies (precision). The minimum precision
is 8 figures.

eneseq = {
name = "$JOBNAME$[_replica$REPLICA$].ene"
first = 0.0
interval = 1.2
precision = 9
}

B.4.17 checkpt
The checkpt parameter specifies the name of the checkpoint .cpt file, when to start
recording (first) and how often to overwrite it (interval).

checkpt = {
name = "$JOBNAME.cpt"
first = 0.0
interval = 240.0
write_last_step = yes
}
To let Desmond periodically write out a .cpt file at certain wall time interval, you can set
checkpt in this form:

checkpt = {
name = "$JOBNAME.cpt"
wall_interval = 3600.0
write_last_step = yes
}
where wall_interval sets the wall time interval in seconds. You can turn off wall time
interval by either setting it to inf (wall_interval = inf) or not including it at all. If the
wall_interval parameter is set to a finite value, then the first and interval parameters
have no effect.

You can turn off recording the .cpt file altogether using:

checkpt = off

136
 Appendix B: The Configuration File

B.4.18 maeff_output
The maeff_output parameter specifies the name of the output .cms file, when to start
recording this file (first), how often to overwrite it (interval), and the number of signifi-
cant figures given in floating-point numbers (precision). It also specifies the name of the
.idx file (used for recording the trajectory name).

maeff_output = {
name = "$JOBNAME$[_replica$REPLICA$]-out.cms"
trjidx = "$JOBNAME$[_replica$REPLICA$]-out.idx"
first = 0.0
interval = 120.0
precision = 8
}

B.4.19 randomize_velocity
Sometimes it is useful to randomize the velocities of the atoms at the start of or during a simu-
lation. The randomize_velocity parameter specifies when to start randomizing velocities
(first), how often to do it (interval) and the random number seed used for generating the
velocities (seed). The temperature parameter sets the target temperature for the randomized
velocities.

randomize_velocity = {
first = 0.0
interval = inf
seed = 2007
temperature = '@*.temperature'
}
The value '@*.temperature' means the value of the temperature parameter that is in the
parent map of the randomize_velocity parameter. The @ symbol in the value is referencing
operator, * means the parent map of the current map, so *.temperature means the tempera-
ture in the parent map. This syntax is a general feature in the sense that every parameter can be
set syntactically in this way. For example:

randomize_velocity.first = '@*.trajectory.first'
sets randomize_velocity.first to the value of trajectory.first. Circular references
(which are possible syntactically) are caught during the initial job-launch processing.

Desmond 3.1 User Manual 137

 Appendix B: The Configuration File

B.4.20 simbox_output
The simbox parameter controls how Desmond records information about the simulation box.
This parameter specifies the name of the file, when to start recording the file (first) and how
often to update it (interval).

simbox = {
name = "$JOBNAME$[_replica$REPLICA$]_simbox.dat"
first = 0.0
interval = 1.2
}

B.4.21 energy_group
Desmond can record a break-down of the energy components during the simulation. The
energy_group parameter controls this functionality and specifies the name, time to start
recording (first) and frequency to record (interval) this information, e.g.,

energy_group = {
name = "$JOBNAME$[_replica$REPLICA$]_enegrp.dat"
first = 0.0
interval = 1.2
self_energy = false
corr_energy = true
}
Setting corr_energy to true allows the correction energy to be printed to the output file.
Setting self_energy to true includes the self-energy term of the Ewald summation into the
correction energy.

B.4.22 backend
The descriptions above handle most (if not all) of the parameters that you would typically use
with Desmond. However, if additional parameters available in the back-end config file are
needed then the backend parameter can be used to specify them in a front-end config file.

For example:

backend = {
force.nonbonded.r_lazy = 12.0
}
adds force.nonbonded.r_lazy = 12.0 to the back-end config file.

138
 Appendix B: The Configuration File

Settings within the backend maps have the highest precedence and thus are honored uncondi-
tionally. Incorrect settings within this map are not detected by the driver.

B.5 Parameters for Minimization

All common parameters can be set in a config file used for a minimization job without causing
errors, but understandably, not every common parameter makes sense for minimization, e.g.,
annealing, temperature, timestep. Such parameters, if set in the config file, are ignored
in minimization calculations.

B.5.1 max_steps
The max_steps parameter specifies the maximum number of steps (or iterations) that the
Desmond minimizer runs, e.g.

max_steps = 2000

B.5.2 convergence
The convergence parameter sets the convergence criterion in kcal/(mol angstrom), e.g.,

convergence = 1.0

B.5.3 steepest_descent_steps
It is often good to use a few steepest descent minimization steps prior to using more elaborate
minimization approaches. The keyword steepest_descent_steps sets the number of
steepest descent steps to use, e.g.,

steepest_descent_steps = 10
After these steps, the LBFSG method is used to further minimize the system.

B.5.4 num_vector
This parameter sets the number of vectors to use for LBFSG method, e.g.

num_vector = 3

Desmond 3.1 User Manual 139

 Appendix B: The Configuration File

B.6 Parameters for Replica Exchange Simulations

All common parameters can be set in a config file used for a replica exchange simulation job
without causing errors.

B.6.1 replica
The replica parameter sets the simulation configuration for each replica. There are two
forms for setting this parameter. The first is for replica exchange calculations in which the
temperature is set in each replica for all atoms in the system. Here, the replica parameter is a
list, and each element of the list is a map value, which in turn specifies the simulation configu-
ration for the corresponding replica. This example

replica = [ { temperature = 300.0 }

{ temperature = 302.0 }
{ temperature = 305.0 }
{ temperature = 308.0 }
{ temperature = 312.0 }
]
specifies that 5 replicas that will run, one at each of 300.0, 302.0, 305.0, 308.0, and 312.0 K.
Within each of these map values, additional common parameters may be set. Note that there is
no default value for replica parameters, since the settings often strongly depend on the model
system.

Sometimes, different CMS files are used for different replicas. The CMS files can be specified
in this form:

replica = [ { model_file = file1.cms temperature = 300.0 }

{ model_file = file2.cms temperature = 302.0 }
{ model_file = file3.cms temperature = 305.0 }
{ model_file = file4.cms temperature = 308.0 }
{ model_file = file5.cms temperature = 312.0 }
]
The second form of the replica parameter is used for the REST (replica exchange with solute
tempering) method [46], in which the temperature is set for only a subset of the atoms (the
solute) in each replica. Here, replica is specified as a map, for example

replica = {
generator = solute_tempering
atom = "asl:atom.num 1-128"
temperature = [300.0 400.0 500.0 600.0]
}

140
 Appendix B: The Configuration File

The generator keyword specifies that solute tempering is used to generate the replicas. The
atom keyword specifies the atoms involved, and the temperature list provides the tempera-
tures to be used.

The precedence of replica settings over other settings is:

• backend for replica

• backend
• replica
• common settings

B.7 Parameters for Metadynamics Simulations

All common parameters can be set in a config file used for a replica exchange simulation job
without causing errors.

B.7.1 meta
The meta block is used to set up a metadynamics simulation. The parameters that apply to all
collective variables (CVs) are the height of the repulsive Gaussian potential, in kcal/mol, the
time in ps at which they are first added, and the interval in ps at which they are subsequently
added. The collective variables are defined by the cv keyword.

meta = {
cv = cv-map
cv_name = $JOBNAME.cvseq
first = 0.0
height = 0.03
interval = 1.2
name = kernels.kerseq
}
Two output files are defined: cv_name sets the name of the metadynamics log file (default
$SJOBNAME.cvseq), and name sets the name of the metadynamics output file, which contains
the Gaussian widths and height at each step, and is used to calculate the free energy. You can
specify the time of the last addition with the last keyword; the default is the simulation time.

B.7.2 cv
The cv parameter defines the collective variables in a metadynamics simulation. Each variable
is defined in a map that includes the type (dist, angle, dihedral, rmsd), the atoms that are
in the collective variable, and the RMS width of the repulsive Gaussian potential. For distance
variables, a wall can be placed at a specified distance, which prevents the system from moving

Desmond 3.1 User Manual 141

 Appendix B: The Configuration File

too far in the direction defined by the collective variable. For the analysis, the range of coordi-
nate values can be defined. This is useful if the interesting phenomena occur at the end of the
range, e.g. at 180° for a dihedral angle. An example with two CVs is as follows:

cv = [ { type = dist
atom = [1 3]
width = 0.4
wall = 10.0
{ type = angle
atom = [1 3 5 6]
width = 0.4
range = [0 360] } ]
To define a collective variable in terms of the center of mass of a set of atoms, you can provide
a list of atoms inside the atom list, e.g.

atom = [[2379 2380 2382 2384 2385] [631 642 651]]

type = "dist"
In this example, the first “atom” in the distance variable is the center of mass of six atoms, and
the second “atom” is the center of mass of three atoms. Distance variables are in angstroms,
angles are in degrees.

If you define the collective variable in the config file, you can only use a list of atoms to define
the center of mass. If you want to use ASL to define the center of mass, you can do so in a
meta block in the .msj file.

B.8 Parameters for vrun

All common parameters that can be set in a config file can be used for a vrun job without
causing errors.

B.8.1 vrun_frameset
The vrun_frameset parameter is used to set the path of the MD simulation trajectory that
should be analyzed. e.g.,

vrun_frameset = /home/joeuser/Desmond_files/my_study_trj
There is no default value for this parameter.

142
 Desmond User Manual

Appendix C

Appendix C: Analyzing a Simulation from the

Command Line

This appendix documents the command line usage of two related Python scripts and the syntax
of two file formats used for analysis of simulation quality. The two Python scripts are
simulation_block_data.py and simulation_block_test.py, and are located in the
directory $SCHRODINGER/mmshare-vversion/lib/Linux-x86/lib/python2.6/site-
packages/schrodinger/application/desmond/. These scripts can be run with the
$SCHRODINGER/run command.

C.1 simulation_block_data.py
This command determines simulation properties from the input .log file and block averages
from the input .ene file, and writes the results to the output .sba file. The syntax is as
follows:

simulation_block_data.py [-n block-length] [-s simboxfile] -e enefile

-l logfile -c sbafile
The options are described in Table C.1.

Table C.1. Options for the simulation_block_data.py command.

Option Description

-e enefile Input .ene file name from a simulation. No default.

-l logfile Input .log file name from a simulation. No default
-c sbafile Output .sba file name. No default.
-n block-length Input block length in ps for calculation of block averages. Default: 10 ps.
-s simboxfile Simbox .dat file name.This is the file that Desmond writes out to track how the
shape of the simulation box evolves during the simulation.

Desmond 3.1 User Manual 143

 Appendix C: Analyzing a Simulation from the Command Line

C.2 simulation_block_test.py
This command evaluates the results in an input.sba file using tests specified in an .sbt file.
The output file is a plain text file that prints the job details block from the .sba file and then
provides information on whether block averages are within the tests specified in the .sbt file.
The syntax is as follows:

simulation_block_test.py [-d] -i sbafile -t sbtfile -o outfile

The output file has the following information for each test in the .sbt file.

• Test (Pass or Fail)

• Testing criteria from input .sbt file
• Corresponding block values from input .sba file
An example output file has the following information besides the job details block from the
.sba file.
Test for E: Fail
Testing criteria: SD = 5.000, Slope = 4.300 kcal/mol/ps , Average = -435320.200
+/- 4000.000 kcal/mol
Block data: SD = 44.941, Slope = -0.034 kcal/mol/ps, Average = -128666.563 kcal/
mol

C.3 Simulation Block Analysis (.sba) File Syntax

The .sba file contains properties obtained from the .log file and block averages from the
.ene file. It has a header block followed by the job details block and then by data blocks.

The header block contains the information on the energy and log files and has the following
format:

Version: version
Energy_File: enefile
Log_File: logfile
The job details block contains the information from the log file about the simulation, such as
the status of the simulation, number of atoms, ensemble, and so on. An example block is
shown below:

Block: Job_Details
Status = Normal
Temperature = 300.0
Job_name = rin
Degrees_of_freedom = 103139
Molecules = 3

144
 Appendix C: Analyzing a Simulation from the Command Line

Duration = 1.2
Atoms = 50274
Ensemble = MTK_NPT
End_Block
Subsequent data blocks printed in the .sba file are block averages enclosed between Block
and End_Block lines. A sample data block is shown below:

Block: E
Time(ps) E(kcal/mol)
5.0 5.0
10.0 4.9
End_Block
The first row is a heading that indicates what quantities are listed below, and it is followed by
rows of values. This block indicates that block average for E for data points up to the first 5 ps
was 5.0 kcal/mol, and for the next 5 ps (5ps to 10ps) it was 4.9 kcal/mol.

C.4 Simulation Block Test (.sbt) File Syntax

Simulation block test (.sbt) files are used to test the data in .sba files and determine the
stability of the simulation. These files contain the test parameters for various properties. A
sample test set is as follows:

E {
sd = 5.0
slope = 4.3
average = -435320.2
average_tol = 4000.0
}
This block indicates that the block values for the property called E in an input .sba file should
have following properties:

• Standard deviation < 5.0

• Slope (i.e. drift with respect to time) < 4.3
• Average should be within -435320.2 ± 4000

Desmond 3.1 User Manual 145

146 Schrödinger Suite 2012 Update 2
Desmond User Manual

References

1. Bowers, K.J.; Chow, E.; Xu, H.; Dror, R. O.; Eastwood, M. P.; Gregerson, B. A.;
Klepeis, J. L.; Kolossvary, I.; Moraes, M. A.; Sacerdoti, F. D.; Salmon, J. K.; Shan, Y.;
Shaw, D. E. Scalable Algorithms for Molecular Dynamics Simulations on Commodity
Clusters, Proceedings of the ACM/IEEE Conference on Supercomputing(SC06),
Tampa, Florida, November 11-17, 2006, http://sc06.supercomputing.org/schedule/
event_detail.php?evid=9088.
2. Shaw, D.E. A fast, scalable method for the parallel evaluation of distance-limited pair
wise particle interactions. J. Comput. Chem. 2005, 26, 1318.
3. Bowers, K.J.; Dror, R.O.; Shaw, D.E. The midpoint method for parallelization of
particle simulations. J. Chem. Phys. 2006, 124, 184109.
4. Bowers, K.J.; Dror, R.O.; Shaw, D.E. Zonal methods for the parallel execution of range-
limited N-body simulations. J. Comput. Phys. 2007, 221, 303.
5. Lippert, R.A.; Bowers, K.J.; Dror, R. O.; Eastwood, M.P.; Gregersen, B. A.; Klepeis, J.
L.; Kolossvary, I.; Shaw, D. E. A common, avoidable source of error in molecular
dynamics integrators. J. Chem. Phys. 2007, 126, 046101.
6. Arkin, I.T.; et al. Mechanism of Na+/H+ Antiporting. Science, 2007, 317, 799.
7. Humphrey, W.; Dalke, A.; Schulten, K. VMD - Visual Molecular Dynamics, J. Molec.
Graphics, 1996, 14, 33.
8. Cornell, W.D.; Cieplak P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.; Ferguson, D. M.;
Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.; Kollmann, P. A. J. Am. Chem. Soc. 1995,
117, 5179. Parameters converted from those at http://amber.scripps.edu/
amber9.ffparms.tar.gz
9. Kollman, P. A. Acc. Chem. Res. 1996, 29, 461. Parameters converted from those at http:/
/amber.scripps.edu/amber9.ffparms.tar.gz
10. Wang, J.; Cieplak, P.; Kollman, P. J. Comput. Chem. 2000, 21, 1049. Parameters
converted from those at http://amber.scripps.edu/amber9.ffparms.tar.gz
11. Hornak et al. Proteins: Structure, Function & Genetics, 2006, 3, 712.
12. Shaw, D. E.; Maragakis, P.; Lindorff-Larsen, K.; Piana, S.; Dror, R. O.; Eastwood, M.
P.; Bank, J. A.; Jumper, J. M.; Salmon, J. K.; Shan Y.; Wriggers, W. Science 2010, 330,
341.

Desmond 3.1 User Manual 147

 References

13. Duan, Y.; Wu, C. Chowdhury, S; Lee, M. C.; Xiong, G.; Zhang, W.; Yang, R.; Cieplak,
P.; Luo, R.; Lee, T.; Caldwell, J.; Wang, J.; Kollman, P. J. Comput. Chem. 2003, 24,
1999. Parameters converted from those at http://amber.scripps.edu/
amber9.ffparms.tar.gz. Bugfix from http://amber.scripps.edu/bugfixes/9.0/bugix.5
applied to correct torsional assignments.
14. Parameters generated from http://mackerell.umaryland.edu/CHARMM_ff_params_files
/toppar/toppar_c35b2_c36a2.tar.gz.
15. Beglov, D.; Roux, B. J. Chem. Phys. 1994, 100, 9050 (ions).
16. MacKerell, A. D., Jr., et al. J. Phys. Chem. B. 1998, 102, 3586 (proteins).
17. MacKerell, Jr. A. D.; Feig M.; Brooks, A. D., III J. Comput. Chem. 2004, 25, 1400
(protein CMAP). Missing CMAP term applied to protonated HIS.
18. Foloppe N.; MacKerell, A. D., Jr. J. Comp. Chem. 2000, 21, 86 (nucleic acids).
19. MacKerell, A. D., Jr. Banavali, N. K. J. Comp. Chem. 2000, 21, 105 (nucleic acids).
20. Feller, S. E.; MacKerell, A. D., Jr. J. Phys. Chem. B. 2000, 104, 7510 (lipids).
21. Feller, S. E.; Gawrisch, K.; MacKerell, A. D., Jr. J. Am. Chem. Soc. 2002, 124, 318
(lipids).
22. Klauda, J. B.; Brooks, B. R.; MacKerell, A. D., Jr. J. Phys. Chem. B, 2005, 109, 5300
(alkanes/lipids).
23. Klauda, J. B.; Venable, R. M.; Freites, J. A.; O’Connor, J. W.; Tobias, D. J.; Mondragon-
Ramirez, C.; Vorobyov, I.; MacKerell, A. D., Jr.; Pastor, R. W. 2010, 114, 7830.
24. Jorgensen, W. L.; Maxwell, D. S.; Tirado-Rives, J., J. Am Chem. Soc. 1996, 118, 11225.
25. Damm, W.; Frontera, A.; Tirado-Rives, J.; Jorgensen, W. L. J. Comput. Chem. 1997, 18,
1955.
26. Jorgensen, W. L., et al. Theochem. 1998, 424, 145.
27. McDonald, N. A., Jorgensen, W. L. J. Phys. Chem. B. 1998, 102, 8049.
28. Rizzo, R. C.; Jorgensen, W. L. J. Am. Chem. Soc. 1999, 121, 4827.
29. Watkins, E. K.; Jorgensen, W. L., J. Phys. Chem. A. 2001, 205, 4118.
30. Kaminski, G. A.; Friesner, R. A.; Tirado-Rives, J.; Jorgensen, W. L. J. Phys. Chem. B
2001, 105, 6474.
31. OPLSAA/L reparameterization, version 1 torsions used for SER, version 1 for ASP,
version 3 (combined) for LEU, VAL, Jacobson, M.P., et al. J. Phys. Chem. B. 2002, 106,

148 Schrödinger Suite 2012 Update 2

 References

11673. The charges for HISE used in this force field have not been published to our
knowledge.
32. Jensen, K. P.; Jorgensen, W. L. J. Chem. Theory Comput. 2006, 2, 1499.
33. Jacobson, M.P.; Kaminski, G. A.; Friesner, R. A.; Rapp, C. S. J. Phys. Chem. B. 2002,
106, 11673.
34. Berendsen, H. J. C. et al. in Intermolecular Forces, edited by B. Pullman (Reidel,
Dordrecht,1981), p. 331.
35. Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P. J. Phys. Chem. 1987, 91, 6269.
36. Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. J. Chem.
Phys. 1983, 79, 926. Parameters as tabulated in Mahoney, M. W.; Jorgensen, W. L. J.
Chem. Phys. 2000, 112, 8910.
37. Neria, E.; Fischer, S.; Karplus, M. J. Chem. Phys. 1996, 105, 1902.
38. Jorgensen, W. L.; Madura, J. D. Mol. Phys. 1985, 56, 1381. Parameters as tabulated in
Mahoney, M. W.; Jorgensen, W. L. J. Chem. Phys. 2000, 112, 8910.
39. Horn, H. W.; Swope, W. C.; Pitera, J. W.; Madura, J. D.; Dick, T. J.; Hura, G. L. J.
Chem. Phys. 2004, 120, 9665.
40. Mahoney, M. W.; Jorgensen, W. L. J. Chem. Phys. 2000, 112, 8910.
41. Earl D. J; Deem, M. W.; Phys. Chem. Chem. Phys., 2005, 7, 3910.
42. Guo Z; Mohanty U.; Noehre J.; Sawyer T. K.; Sherman W.; Krilov G. Probing the α-
helical structural stability of stapled p53 peptides: molecular dynamics simulations and
analysis. Chem. Biol. Drug Des. 2010, 75, 348.
43. Patriksson, A; van der Spoel, D., A temperature predictor for parallel tempering simula-
tions. Phys. Chem. Chem. Phys. 2008, 10, 2073.
44. Gervasio, F. L.; Laio, A.; Parrinello, M. Flexible docking in solution using metady-
namics. J. Am. Chem. Soc. 2005, 127, 2600.
45. Shivakumar, D.; Williams, J.; Wu, Y.; Damm, W.; Shelley, J.; Sherman, W. Prediction of
Absolute Solvation Free Energies using Molecular Dynamics Free Energy Perturbation
and the OPLS Force Field. J. Chem. Theory Comput. 2010, 6, 1509.
46. Liu, P.; Kim, B.; Friesner, R. A.; Berne, B. J., Replica exchange with solute tempering: a
method for sampling biological systems in explicit water. Proc. Natl. Acad. Sci. U.S.A.
2005, 102, 13749.

Desmond 3.1 User Manual 149

150 Schrödinger Suite 2012 Update 2
Desmond User Manual

Getting Help

Information about Schrödinger software is available in two main places:

• The docs folder (directory) of your software installation, which contains HTML and
PDF documentation. Index pages are available in this folder.
• The Schrödinger web site, http://www.schrodinger.com/, particularly the Support Center,
http://www.schrodinger.com/supportcenter, and the Knowledge Base, http://www.schro-
dinger.com/kb.

Finding Information in Maestro

Maestro provides access to nearly all the information available on Schrödinger software.

To get information:
• Pause the pointer over a GUI feature (button, menu item, menu, ...). In the main window,
information is displayed in the Auto-Help text box, which is located at the foot of the
main window, or in a tooltip. In other panels, information is displayed in a tooltip.
If the tooltip does not appear within a second, check that Show tooltips is selected under
General → Appearance in the Preferences panel, which you can open with CTRL+, (,).
Not all features have tooltips.
• Click the Help button in a panel or press F1 for information about a panel or the tab that is
displayed in a panel. The help topic is displayed in your browser.
• Choose Help → Online Help or press CTRL+H (H) to open the default help topic in your
browser.
• When help is displayed in your browser, use the navigation links or search the help in the
side bar.
• Choose Help → Manuals Index, to open a PDF file that has links to all the PDF docu-
ments. Click a link to open the document.
• Choose Help → Search Manuals to search the manuals. The search tab in Adobe Reader
opens, and you can search across all the PDF documents. You must have Adobe Reader
installed to use this feature.

Desmond 3.1 User Manual 151

 Getting Help

For information on:

• Problems and solutions: choose Help → Knowledge Base or Help → Known Issues →
product.
• Software updates: choose Maestro → Check for Updates.
• New software features: choose Help → New Features.
• Scripts available for download: choose Scripts → Update.
• Python scripting: choose Help → Python Module Overview.
• Utility programs: choose Help → About Utilities.
• Keyboard shortcuts: choose Help → Keyboard Shortcuts.
• Installation and licensing: see the Installation Guide.
• Running and managing jobs: see the Job Control Guide.
• Using Maestro: see the Maestro User Manual.
• Maestro commands: see the Maestro Command Reference Manual.

Contacting Technical Support

If you have questions that are not answered from any of the above sources, contact Schrödinger
using the information below.

E-mail: help@schrodinger.com
USPS: Schrödinger, 101 SW Main Street, Suite 1300, Portland, OR 97204
Phone: (503) 299-1150
Fax: (503) 299-4532
WWW: http://www.schrodinger.com
FTP: ftp://ftp.schrodinger.com

Generally, e-mail correspondence is best because you can send machine output, if necessary.
When sending e-mail messages, please include the following information:

• All relevant user input and machine output

• Desmond purchaser (company, research institution, or individual)
• Primary Desmond user
• Installation, licensing, and machine information as described below.

152 Schrödinger Suite 2012 Update 2

 Getting Help

Gathering Information for Technical Support

This section describes how to gather the required machine, licensing, and installation informa-
tion, and any other job-related or failure-related information, to send to technical support.

For general enquiries or problems:

1. Open the Diagnostics panel.
• Maestro: Help → Diagnostics
• Windows: Start → All Programs → Schrodinger-2012 → Diagnostics
• Mac: Applications → Schrodinger2012 → Diagnostics
• Command line: $SCHRODINGER/diagnostics
2. When the diagnostics have run, click Technical Support.
A dialog box opens, with instructions. You can highlight and copy the name of the file.
3. Attach the file specified in the dialog box to your e-mail message.

If your job failed:

1. Open the Monitor panel in Maestro.
Use Applications → Monitor Jobs or Tasks → Monitor Jobs.
2. Select the failed job in the table, and click Postmortem.
The Postmortem panel opens.
3. If your data is not sensitive and you can send it, select Include structures and deselect
Automatically obfuscate path names.

4. Click Create.
An archive file is created in your working directory, and an information dialog box with
the name of the file opens. You can highlight and copy the name of the file.
5. Attach the file specified in the dialog box to your e-mail message.
6. Copy and paste any log messages from the window used to start Maestro (or the job) into
the email message,or attach them as a file.
• Windows: Right-click in the window and choose Select All, then press ENTER to
copy the text.
• Mac: Start the Console application (Applications → Utilities), filter on the applica-
tion that you used to start the job (Maestro, BioLuminate, Elements), copy the text.

Desmond 3.1 User Manual 153

 Getting Help

If Maestro failed:
1. Open the Diagnostics panel.
• Windows: Start → All Programs → Schrodinger-2012 → Diagnostics
• Mac: Applications → Schrodinger2012 → Diagnostics
• Linux/command line: $SCHRODINGER/diagnostics
2. When the diagnostics have run, click Technical Support.
A dialog box opens, with instructions. You can highlight and copy the name of the file.
3. Attach the file specified in the dialog box to your e-mail message.
4. Attach the file maestro_error.txt to your e-mail message.
This file should be in the following location:
• Windows: %LOCALAPPDATA%\Schrodinger\appcrash
(Choose Start → Run and paste this location into the Open text box.)
• Mac: Documents/Schrodinger
• Linux: Maestro’s working directory specified in the dialog box (the location is
given in the terminal window).
5. On Windows, also attach the file maestro.EXE.dmp, which is in the same location as
maestro_error.txt.

154 Schrödinger Suite 2012 Update 2

 Index
A desmond..................................................... 68
desmond_restraints ........................... 102
absolute solvation free energy........................... 48
manipulate_trj...................................... 98
Advanced Ion Placement dialog box ................ 16
mold_gpcr_membrane ........................... 100
Amber
multisim................................................... 70
converting input files to CMS..................... 99
rebuild_cms .......................................... 103
force fields available via viparr .................. 87
solvate_pocket...................................... 94
analysis
system_builder...................................... 75
metadynamics ..................................... 64, 123
trajectory_extract_frame .............. 101
simulation events ................................ 62, 123
viparr ....................................................... 85
simulation quality ....................................... 61
communication plugin ...................................... 69
atom groups
composite model system (CMS) files....... 80, 137
defining in multisim file ........................... 113
specifying for replicas .............................. 140
specifying ................................................... 38
composite system builder (CSB) file ................ 75
thermostat ................................................... 33
example ...................................................... 76
multisim keyword..................................... 116
B config file ........................................................ 127
barostat back-end, adding keywords for ................ 138
coupling style ..................................... 34, 133 macros ...................................................... 128
methods available ....................................... 34 reading ........................................................ 19
pressure............................................... 34, 133 syntax ....................................................... 127
barrier, metadynamics distance variables . 29, 141 units .......................................................... 128
bond, attachment ............................................... 42 configuration file– see config file
bonded time step ....................................... 32, 134 conventions, document ...................................... ix
box shape ............................................ 11, 78, 116 coordinate origin, resetting ............................. 116
box size ....................................................... 11, 78 corrections
box volume, minimizing ..................... 11, 76, 116 charged molecules, solvation free energy... 52
buffer distance enthalpic, for restraints ............................. 124
in plane of membrane ................................. 78 long-range dispersion, for absolute free
solute to box boundary ....................... 11, 116 energy .................................................... 124
coupling scheme, barostat............................... 133
C cutoff radius, nonbonded interactions....... 35, 131
charged molecules, correction to solvation free
energy ............................................................ 52 D
CHARMM, force fields available via viparr..... 87 desmond command .......................................... 68
checkpoint file........................................... 37, 106 directory
config file keyword................................... 136 installation .................................................... 6
importing .................................................... 20 Maestro working........................................... 6
probing................................................ 71, 108 dispersion correction, long-range ................... 124
clashes, ligand-protein, relieving ...................... 43 domain decomposition.............................. 39, 105
collective variables............................................ 28 config file specification ............................ 135
bin size for analysis .................................. 123
config file block........................................ 141 E
maximum number....................................... 29
energy components, config file keyword ........ 138
range for analysis...................................... 123
energy recording ............................................... 22
command syntax
config file keyword................................... 136
amber_prm2cms ........................................ 99

Desmond 3.1 User Manual 155

 Index

ensemble saving trajectory frame ............................... 61

classes available.......................................... 23 input files, writing....................................... 20, 42
config file keyword................................... 133 integration algorithm, setting parameters for.... 32
replica exchange ......................................... 27 interactions
valid classes .............................................. 133 Coulomb method, config file keyword..... 132
environment variable short-range and long-range......................... 35
SCHRODINGER .............................................. 6 truncation at cutoff ............................. 35, 131
Ewald tolerance................................................. 35 interval
averaging, for simulation quality analysis.. 61
F metadynamics potential addition.............. 141
velocity randomization ............................. 137
far time step............................................... 32, 134
interval, recording
FEP calculations ............................................... 41
checkpoint file .................................... 37, 136
use of solvate_pocket................................ 121
energy ................................................. 22, 136
FEP protocol ..................................................... 50
energy group............................................. 138
adjusting ..................................................... 53
simulation box .......................................... 138
default ......................................................... 51
trajectory............................................. 22, 135
files
ions
analysis report........................................... 123
adding ........................................... 14, 79, 116
checkpoint ................................................ 106
excluding from region .......................... 15, 79
deleting unwanted stage data............ 105, 125
placing near residues ............................ 15, 79
extern stage extra ...................................... 123
types available ............................................ 15
input, writing ........................................ 20, 42
multisim (msj) .......................................... 105
output structure......................................... 111 J
relaxation protocol...................................... 24 job name............................................................ 39
structure, for FEP jobs.............................. 118 macro for, in config file ............................ 129
force fields .................................................. 14, 79 multisim job................................................ 71
available with viparr ................................... 87 jobs, restarting multisim ................................. 106
merging....................................................... 86
modifying ................................................... 85 L
specifying with viparr................................. 86
lambda schedule................................................ 54
system builder assignment........................ 116
config file keywords ................................. 130
free energy
default for absolute free energies ............. 131
average as function of time....................... 125
default for relative free energies............... 130
plot, as function of collective variables ...... 64
lambda windows ............................................... 54
relative binding ............................... 41, 46, 51
index ......................................................... 130
solvation ......................................... 41, 48, 52
structure files ............................................ 118
ligand environment, FEP simulations ............... 49
G ligand fragments ............................................... 43
Gaussian potential, metadynamics............ 28, 141 clashes with protein .................................... 43
glue.................................................................... 37 editing......................................................... 44
config file keyword................................... 135 Ligand Functional Group Mutation by FEP
panel .............................................................. 43
I ligand mutation ................................................. 42
long-range dispersion correction for absolute free
images
energy .......................................................... 124
plotter to use ..................................... 123, 124

156 Schrödinger Suite 2012 Update 2

 Index

M mutation FEP
ligand .......................................................... 42
master job.......................................................... 39
protein residue ............................................ 46
host ............................................................. 71
ring atom .................................................... 44
membrane
mutation FEP, use of Hamiltonian replica
adding ................................................... 11, 78
exchange .................................................. 51, 55
adjusting orientation ................................... 13
available models ......................................... 11
buffer distance ............................................ 78 N
embedding GPCR in................................. 100 near time step ............................................ 32, 134
OPM ........................................................... 12 neutralizing the model system .......................... 80
placing ........................................................ 12 node locking.................................................... 106
relaxation protocol................................ 24, 30
surface area................................................. 30 O
surface tension ............................................ 30
metadynamics ................................................... 28 output, config file keyword............................. 137
analysis ....................................................... 64
bin size for analysis .................................. 123 P
config file block........................................ 141 partial charges ................................................... 13
distance barrier ................................... 29, 141 particle mesh Ewald (PME) method............... 132
log file name ............................................. 141 periodic boundary
mutlsim block ........................................... 119 setting up .............................................. 11, 78
Metadynamics Analysis panel .......................... 65 wrapping molecules at.............................. 135
Metadynamics panel ......................................... 29 pressure
minimization ..................................................... 21 config file keywords ................................. 133
algorithm .................................................... 21 simulation, setting ...................................... 23
convergence threshold ........................ 22, 139 probability profiles.......................................... 123
maximum iterations ............................ 21, 139 processors
number of LBFSG vectors........................ 139 allocation .................................................... 39
setting parameters for ................................. 34 choosing number of .............................. 39, 70
steepest descent steps ............................... 139 config file keyword................................... 135
Minimization panel ........................................... 21 product installation ......................................... 152
model system properties, for simulation event analysis .......... 62
adding ions to ............................................. 79 protein side chains, mutating ............................ 46
adding salt to .............................................. 80
importing .................................................... 21 R
neutralizing ................................................. 80
preparing..................................................... 75 random seed
relaxing ....................................................... 23 Bennett method, FEP analysis.................. 124
selecting...................................................... 20 velocity ..................................................... 137
solvating ..................................................... 80 recording
Molecular Dynamics panel ............................... 22 checkpoint file .................................... 37, 136
msj file, example ............................................... 72 energy ................................................. 22, 136
multisim .................................................... 70, 105 energy group............................................. 138
file syntax ................................................. 109 simulation box .......................................... 138
job file example .......................................... 72 start time ..................................................... 37
restarting jobs ........................................... 106 trajectory............................................. 22, 135
template commands .................................. 105 update frequency ........................................ 37

Desmond 3.1 User Manual 157

 Index

relative binding free energy .............................. 41 solutes

relaxation protocol keeping together ................................. 37, 135
default ......................................................... 23 reading for model system ........................... 76
membrane ................................................... 30 solvation free energy......................................... 41
modifying ................................................... 24 correcting for charged molecules ............... 52
replica exchange................................................ 27 solvent
config file temperature keyword............... 140 models ........................................................ 10
default ensemble......................................... 27 multisim keyword............................. 116, 117
graph of exchanges ..................................... 63 specifying ............................................... 9, 77
mutation FEP use.................................. 51, 55 stages, multisim job ........................................ 110
solute tempering (REST).......................... 140 common keywords.................................... 111
temperature ........................................... 27, 69 deleting unwanted data files ..................... 125
temperature profile ..................................... 27 Desmond................................................... 111
Replica Exchange Dynamics Review panel...... 63 inserting on restart .................................... 107
Replica Exchange Graph panel......................... 64 modifying on restart ................................. 107
report file......................................................... 123 restarting from completed ........................ 107
restarting jobs skipping .................................................... 111
from a completed stage............................. 107 supported .................................................. 110
with different config file........................... 107 Start dialog box................................................. 40
with new stages......................................... 107 start time
restarting subjobs checkpoint file recording.................... 37, 136
maximum attempts ..................................... 71 energy group recording ............................ 138
restraints energy recording....................................... 136
enthalpic correction for............................. 124 metadynamics potential addition.............. 141
on atom positions........................................ 36 simulation ................................................. 134
simulation box recording.......................... 138
S trajectory................................................... 135
velocity randomization ............................. 137
salt, adding .................................................. 16, 80
structures
Schrödinger contact information .................... 152
checking validity of ...................................... 4
Set Up Membrane panel ................................... 12
preparation................................................ 4, 9
side chain
reading solute for system building ............. 77
modifying ................................................... 48
specifying for viparr ................................... 86
reorienting .................................................. 48
subjobs
simulated annealing
changing stage parameters ....................... 107
activating, config file keyword ................. 133
host to run on.............................................. 71
number of stages......................................... 25
job name ................................................... 111
temperature schedule, config file
location ......................................... 39, 71, 106
keyword ................................................. 132
maximum number .............................. 71, 107
Simulated Annealing panel ............................... 25
maximum restarts ....................................... 71
simulation event analysis .................................. 62
subdirectory names................................... 111
specification block, MSJ file .................... 123
synchronizing ........................................... 111
Simulation Event Analysis panel ...................... 63
substitution group, ligand mutation .................. 42
simulation parameters
surface area, membrane .................................... 30
extra for desmond command...................... 69
surface tension
Simulation Quality Analysis panel ................... 62
config file keyword................................... 133
simulation time ................................................. 22
simulation, setting ...................................... 23
config file keyword................................... 134
system builder ................................................... 75

158 Schrödinger Suite 2012 Update 2

 Index

System Builder panel superimposing on Workspace structure...... 60

Ions tab ....................................................... 14 viewing during simulation.......................... 61
Solvation tab ............................................... 10 viewing in Maestro ..................................... 57
viewing with VMD..................................... 81
T visible atoms............................................... 60
Trajectory panel ................................................ 58
tapering schemes....................................... 35, 131
transmembrane atoms, defining........................ 12
temperature
config file keyword................................... 132
desmond command option for replica U
exchange.................................................. 69 units, in config file .......................................... 128
randomized velocity target ....................... 137 update frequency............................................... 37
replica exchange ................................. 27, 140
simulated annealing, config file keyword. 132 V
simulation, setting ...................................... 23
velocities
solvate_pocket .......................................... 120
randomizing........................................ 38, 137
thermostat ................................................. 132
recording in trajectory ................................ 37
temperature profile, replica exchange............... 27
temperature program......................................... 24
thermodynamic properties W
plot over time.............................................. 62 water
thermostat force-field models....................................... 87
atom groups ................................................ 38 moving out of solute entry........................ 116
temperature ............................................... 132 penetration between protein and
thermostat groups.............................................. 33 membrane................................................ 30
thermostat method............................................. 33
time
free energy as function of ......................... 125
simulation start ......................................... 134
total simulation ......................................... 134
time series, variables for analysis ................... 123
time steps .......................................................... 32
config file keyword................................... 134
topology, processor ..................... 39, 70, 105, 135
Total Free Energy by FEP panel ....................... 49
trajectory
combined .................................................... 98
config file keyword................................... 135
directory................................................ 24, 37
frame image ................................................ 61
frame positioning........................................ 59
frames per file........................................... 135
path to ....................................................... 142
recording interval........................................ 22
replica exchange ......................................... 28
saving as movie .......................................... 61
saving structures from ................................ 61
smoothing during play................................ 59

Desmond 3.1 User Manual 159

160 Schrödinger Suite 2012 Update 2
120 West 45th Street 101 SW Main Street 245 First Street 8910 University Center Lane
17th Floor Suite 1300 Riverview II, 18th Floor Suite 270
New York, NY 10036 Portland, OR 97204 Cambridge, MA 02142 San Diego, CA 92122

155 Gibbs St Dynamostraße 13 Zeppelinstraße 73 Potsdamer Platz 11

Suite 430 D-68165 Mannheim D-81669 München D-10785 Berlin
Rockville, MD 20850-0353 Germany Germany Germany

Quatro House 8F Pacific Century Place No. 102, 4th Block

Frimley Road 1-11-1 Marunouchi 3rd Main Road, 3rd Stage
Camberley GU16 7ER Chiyoda-ku, Tokyo 100-6208 Sharada Colony
United Kingdom Japan Basaveshwaranagar
Bangalore 560079, India

SCHRÖDINGER ®