lOMoARcPSD|49403025

BIF501 Handouts PDF - BIF

Biotechnology (Virtual University of Pakistan)

Scan to open on Studocu

Studocu is not sponsored or endorsed by any college or university

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)
 lOMoARcPSD|49403025

MUHAMMAD IMRAN

BIF501 - Bioinformatics II
Topic - 1 Applications of Bioinformatics
Applications of Bioinformatics
1. Drug Development 10. Personalized Medicine
2. Crop Improvement 11. Preventive Medicine
3. Microbial Genome 12. Waste Cleanup
4. Gene Therapy 13. Antibiotic Resistance
5. Biotechnology 14. Alternate Energy Science
6. Comparative Study 15. Insect Resistance
7. Evolutionary Studies 16. Climate Change Studies
8. Veterinary Science 17. Nutritional Quality
9. Molecular Medicine

Drug development
✓ Drugs target only about 500 proteins
✓ Disease mechanisms and using computational tools identify and validate new
drug targets

Crop improvement
✓ Comparative genetics of the plant genomes
✓ Information obtained from the model crop systems can be used to suggest
improvements to other food crops.
✓ At present the complete genomes of Arabidopsis thaliana (water cress) and
Oryza sativa (rice) are available.

Microbial genome applications

✓ Complete genome sequences
✓ Environment, health, energy and industrial applications
✓ Isolation of the genes that give them their unique abilities to survive under

extreme conditions.

Gene Therapy
➢ Gene therapy-used to treat, cure or even prevent disease
➢ Clinical trials

Biotechnology
➢ Archaeoglobus fulgidus and Thermotoga maritima
➢ Corynebacterium glutamicum
➢ Xanthomonas campestris
➢ Lactococcus lactis

Evolutionary studies

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

➢ The sequencing of genomes from all three domains of life; eukaryota, bacteria
and archaea

Topic - 2 Applications of Bioinformatics

Veterinary Science
 Farm animals including cows and sheep have been sequenced
Molecular Medicine
➢ The human genome project
➢ 3000-4000 hereditary disease including Cystic Fibrosis and Huntingtons
disease
➢ Response to an environmental stress causes cancers, heart disease, diabetes
➢ Human Genome Project Data Base

Personalized medicine
➢ Pharmacogenomics
➢ Sequence variants in DNA
➢ Trial and error to find the best drug
➢ Patient's genetic profile
➢ With the specific details of the genetic mechanisms of diseases
being unraveled, the development of diagnostic tests to measure a
persons susceptibility to different diseases may become a distinct
reality.
➢ Preventative actions such as change of lifestyle or having treatment
at the earliest possible stages when they are more likely to be
successful, could result in huge advances in our struggle to conquer
disease.
Waste cleanup
➢ Deinococcus radiodurans
➢ Potential usefulness in cleaning up waste sites that contain radiation
and toxic chemicals
Antibiotic Resistance
➢ Enterococcus faecalis
➢ Virulence region-resistant genes
➢ The discovery of the region, known as a pathogenicity island

Alternative energy sources

➢ Chlorobium tepidum
➢ Capacity for generating energy from light

Insect resistance
➢ Bacillus thuringiensis
➢ Control serious pests of cotton, maize and potatoes
➢ Insecticides can be reduced and hence the nutritional quality of the
crops is increased
Climate change Studies
➢ Increasing levels of carbon dioxide emission-global climate change.
➢ Study the genomes of microbes that use carbon dioxide as their sole
carbon source.
Improve nutritional quality

MUHAMMAD IMRAN 2

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

➢ Genes transferred into rice to increase levels of Vitamin A, iron and

other micronutrients
➢ Reducing occurrences of blindness and anaemia

Topic -3 Nucleotide Sequence Databases

Biological Databases:
Biological databases in general store biological data and their main goals are
1. Data storage
2. Information retrieval
3. Knowledge discovery

Classification:
Biological databases can be classified as
➢ Primary databases (that stores the Primary Sequences)
➢ Secondary databases (the primary sequences are annotated and kept in Secondary
Databases)
➢ Specialized Databases (they are dedicated towards some specific organism or can
have some disease data)

Biological databases can also be classified on the bases of types of data which they contain,
such as:
➢ Nucleotide databases
➢ Protein databases
➢ RNA databases
➢ Genome databases
➢ Expression databases (Gene Expression Databases)

Issues:
The issues which are present generally in other databases are also found to be in Biological
databases that may be co-related with the relatively slow pace of quality assurance techniques
as compared to the pace with which new data is emerging, so the issues are similar and are as
follows:
Due to limited Q/A
➢ Redundancy
➢ Inconsistency
➢ Incompatibility (format, terminology, data types, etc.)

Nucleotide Sequence databases:

The Nucleotide Sequence Databases are one of the types of Biological Databases that
contains nucleotide sequences in it, which can be DNA and cDNA or EST sequences.

Here, we have a diagram where we have a genomic DNA which has different Exons(we know
that in Eukaryotes, we have exons and introns). So exonsgets transcribed into mRNA and we
can get cDNA from this mRNA through reverse transcriptionand then we can store this cDNA
into our databases whereas the ESTs are the subsets within those cDNA’s.

MUHAMMAD IMRAN 3

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Origin:
The Nucleotide Sequence Databses were first assembled into Genebank (1982) at Los Alamos
National Laboratory (LANL), New Maxico under the leadership of Walter Goad. GeneBankis
now working under the umbrella of NCBI (National Center for Biotechnology Information).
NCBI is the central repository that stores multiple types of biological data that includes
genomes, their assemblies, their sequencing data, their expression data and what not. In this
diagram, we can see the page where you can search for any kind of data; a drop-down list
which provides you with various options. The link to this page is
http://www.ncbi.nlm.nih.gov/.
Here, is the page for GeneBank, so if you want search about nucleotides and genome
sequences, this is the best resource
NCBI was established in United States.
National Center for Biotechnology Information

GeneBank

EMBL and DDBJ:

• European Molecular biology Lab (EMBL established1980) was established in Europe.

MUHAMMAD IMRAN 4

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• DNA databank of Japan (DDBJ) established in Mishima japan (1984).

INSDC:
• Genebank, DDBJ and EMBL joined together in International Nucleotide Sequence
Database Collaboration (INSDC)

You can see in this diagram, the NCBI (National Center for
Biotechnology Information), DDBJ(DNA Databank of Japan)
and EBI (EuropeanBioinformaticsInstitute) /ENA (European
Nucleotide Archive)forms an International collaboration known
as INSDC (International Nucleotide Sequence Database
Collaboration).
Where EMBLestablishedEBI, to deal with Bioinformatics kind
of stuff and within them they have established ENA to maintain
the DNA sequence datasets

Here, is the page of INSDC (International Nucleotide Sequence Database Collaboration), and
you can observe that all three collaborators’ logos are there. Similarly, if you look into the
data, we can have Next Generation reads, Capillary reads and information about samples and
annotated sequences all on this first page.
(We’ll discuss it later).
Growth of Genebank:
If we look into the growth of Gene Bank as shown in the figure below (left), we can see the
number of bases in the GeneBank which are uncountable as they are in trillions which is a
huge number starting somewhere in 1982 and if we look into these curves, blue is the growth
of GeneBank and red one is the whole genome sequences (which we are comparing) which is
starting somewhere in 2003 or 2004 after the publication of Human genome Project.
So, if you look into the number of bases, it seems like they double after every 18 month which
means the growth is huge and is exponential.
Similarly, if we look into the sequences (right figure), are also around somewhere in 1000’s in
1982 but now they are more than hundred million sequences in this GeneBank.

MUHAMMAD IMRAN 5

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

GeneBank
WGS
http://www.ncbi.nlm.nih.gov/genbank/statistics

Conclusions:
In the end, we conclude some of the followings:
➢ Biological databases store biological data.
➢ INSDC is joint venture of NCBI, EMBL and DDBJ.
➢ Growth of bases in GeneBank is exponential, doubling every 18 months.

Topic - 4 Protein Databases

Introduction:
Protein databases store protein data which may include the following:
✓ Protein sequences
✓ Motif (patterns of amino acids)
✓ Structure
✓ Structure alignments (aligned structures)

Origin:
First sequences to be collected were Proteins (before Nucleotide Sequences) using Sanger and
Tupy’smethods (1951) where Common protein families like cytochromes were sequenced (as
in that era people were focusing on the sequences made from cytochrome molecules).
Atlas of protein sequences (mainly cytochromes) was assembled by Margret Dayhoff and her
collaborators at National Biomedical Research Foundation (NBRF) in 1960s.
PIR (Protein Information Resource):
The collection (of Dayhoff and co) became PIR (Protein Information Resource) which is now
a collaboration of NBRF, Munich Center for Protein Sequences (MIPS) and Japan
International Protein Information Database (JIPID).

Protein Sequences:

MUHAMMAD IMRAN 6

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Swiss-Protis a Collaboration between the SIB (SwissInstitue of Bioinformatics) and EBI

(EuropeanBioinformaticsInstitute) and it weekly releases from about 50 servers across the

world, the main source beingExPASy in Geneva (i.e. it’smainlycontrolled by ExPASywhichis

the main server located in Geneva).
Here, is the page for ExPASy, and you can find different structural alignments, proteomic data,
genomic data.
Protein Sequences:
International partnership between PIR, EBI and SIB created UniProt, by unifying the PIR-
PSD (Protein Information Resource – Protein Structure Database i.e. they kept the PSD of PIR
into UniProt known as PIR-PSD), Swiss-Prot, and TrEMBL(is where we put the translated
sequences from the DNA-where DNA is translated into the protein and the protein sequences
are coming from different reading frames of those DNAs using all 6 reading frames- will be
discussed later) databases.

MUHAMMAD IMRAN 7

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, is the page of UniProt and you can see, we have 3 main sections i.e. Protein Ontologies
labelled as PRO then we have ProClass where we can have the sequences and ProLINK tells
us about the literature.

Here, we look into the PRO which is the Protein Ontologies- ontologies is where we can
classify those proteins on the basis of their functions and different functions have their
hierarchy so ontologies are labelled in form of different hierarchies, so there is a major
function and a trend towards moving the specific function.
Here, we can see just a PRO Hierarchy Ontology in this example.
http://pir.georgetown.edu/

MUHAMMAD IMRAN 8

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

In this figure, as we mentioned earlier, iProClass is the integration of different protein

resources, so we can have sequences from here, protein expression data, and protein
modifications. We can also integrate the genomic data with the proteomic data.
http://www.uniprot.org/

In this figure, we have iProLINK which provides literature information and most of the
research papers can be found here.

MUHAMMAD IMRAN 9

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

The link for it is: http://www.uniprot.org/.

Here, in this figure we have PDB and PDB stands for Protein Data Bank, basically it’s a
repository where we have the protein structures.
These structures are obtained by different chemistry and molecular biology techniques like X-
ray Crystallography in the labs and then those structures are submitted into the PDB where
researchers can get those structures and can compare their predicted structures with them, so
it’s a good resource if you are working on structural protein bioinformatics. The link for this
page is http://www.rcsb.org/pdb/home/home.do

In this figure, we can see SCOP which is a similar effort that utilizes different structural
elements on those proteins and it classify those proteins on the basis of their structural
elements like family, fold super family, domains and then classes.
So, Class is the biggest in this SCOP hierarchy, there
are different major group of classes.
The link is : http://scop.mrc-lmb.cam.ac.uk/scop/

We can see here for an example, we have the class in which we have all the alpha helices;
these helices are formed by special arrangement of amino-acids. Basically when the protein
sequences- just a linear sequence of amino-acids when it turns around on it selves, it forms
those secondary structures so those structures are then recognized as alpha and beta (we are not
going into the details; you can go for molecular biology course or Google about alpha or beta).
The main idea to present here is that SCOP actually classify the proteins on the basis of those
structures so for an example, alpha (is that class where we have all those proteins that has
alpha helices in them), we can also have beta (where we have all those proteins that has beta
chains in them), alpha/beta (where we have alpha helices then comes beta then comes alpha so
they are present one after the other), alpha + beta ( we can have separate regions where we can
have alpha helices stacked together and then we have beta chains stacked together). And the
link for it is - http://scop.mrc-lmb.cam.ac.uk/scop/.

MUHAMMAD IMRAN 10

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Conclusions:
We conclude that:
✓ First sequences to be collected were Protein sequences.
✓ Protein databases are classified on the basis of sequences, motifs, structures and different
structural alignments.
✓ Growth of Sequence in Databases is exponential (just like as in Nucleotide Databases the
growth of sequence is higher).

Topic - 5 Genome & Organism Specific Databases

Origin:
First attempt to sequence free living organism was launched in late 1990’s (Blattner et al.
1997) and Viruses had already been sequenced (Fleischmann et al. 1995).
Haemophilusinfluenzaewas the first genome that was published and the project was initiated
at The Institute of Genome Research (TIGR) under the leadership of Craig Ventor (the same
person who’s name we’ll see in the human genome project). At that time, a method which was
already established known as shotgun sequencing method was being tested by this project to
verify its reliability and efficiency. And by utilizing this method they sequenced the genome
which was about 1.8 million base pairs (bp), it took 9-months and the cost was around 1
million US dollars and this project Paved the way for sequencing of many other organisms.
Examples
• AceDB (AC. elegansDataBase) was the first genome database for genome sequences
was developed in 1989 and was established by Richard durbinandThierry-Miegi.

(Same Durbin whose book “Biological Sequence Analysis” we’ll consider in the latter
half of the course).

Here is the figure of AceDB webpage, you find the sea elegans; a worm and there are other
organisms.
Link for this page is: http://www.acedb.org/.
Examples:
TAIR (The Arabidopsis Information Resource) which is a database for Arabidopsis
(http://www.arabidopsis.org/) and SGB (Saccharomyces Genome Database) actually uses the
system of AceDB (http://www.yeastgenome.org/).

MUHAMMAD IMRAN 11

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Human Genome Project:

Human Genome Project started initially as a Pilot Project which begun by Department
OfEnergy (DOE) of USA in 1986. Two organizations, one is National Human Genome
Research Initiative (NHGRI), which was federally funded organization through NIH
(National Institute of Health) that started in 1988 by Francis Collin which was joined to
Commercial organization named Celera (Celera Genomics)in 1998, a commercial under the
leadership of Craig Venter.
Both of them claimed that they sequenced the full Human Genome
and the issue was raised that who completed the project first, and
who had a major role in it but later on it was resolved by President
Bill Clinton at that time and they published it together back in year
2000. In the end, they concluded that there areTotal 3.4 billion bases
which are sequenced at a cost of $1/base.

While we have those genomes available, we want to see their graphical views where we can
get the reports, get the idea about where different genes are located, so in order to do that we
needed to make something which we call it as genome browsers- are the webpages where we
can look into the different features within our genomes so UCSC is one of the example (shown
on the left) which is University of California Santa Cruz which is the biggest genome browser.
The link to this browser is http://genome.ucsc.edu/.
The figure of UCSC Genome Browser, where we can have information, so on the top we see a
chromosome and down below we see various lines which are known as different tracks (for
snips, genes, EST’s etc.) so we can look or zoom into different regions of the genome by
using those genome browsers.
The link to this webpage is: http://genome.ucsc.edu/.
Conclusion:
In the end, we conclude the following:
✓ Success of Haemophilusinfluenzaepaved the way for other genome sequencing
projects

MUHAMMAD IMRAN 12

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

✓ Human Genome Project was accomplished by NHGRI and Celera (they were
working independently from one another).
✓ Genome browsers help in exploring different regions of the genome.

Topic -6 Gene Expression Databases

Gene Expression Omnibus (GEO):
Genes are expressed into mRNA, and whenever we talk about gene expression, we generally
mean the mRNA sequences so we can normally get those mRNA from techniques like
microarray and another famous technique nowadays which is being established is known as
RNAseq. And microarray data and RNAseq can be classified into Gene Expression Data
which is stored in Gene Expression Databases.
Basically, Gene Expression Databases are the public repository for the archiving and
distribution of gene expression data submitted by the scientific community.
The gene expression data are MIAME compliant data (where MIAME is Minimum
Information About a Microarray Experiment and can be searched on this link:
http://www.mged.org/Workgroups/MIAME/miame.html). Whenever you need to submit a
microarray experiment, you need to give descriptions like how exactly the samples were
prepared, normalization methods which you have used and other counts and normalized files)
so while keeping in view these datasets have different records in it.
Gene Expression Omnibus is convenient for deposition of gene expression data, as required by
funding agencies and journals and it’s also a curated resource for gene expression data where
we can do Browsing, querying, analysis and retrieval of the data.

Here, is the webpage of GEO which is Gene Expression Omnibus running under NCBI (you
can visit NCBI where you can get to the GEO Database) which are having different datasets,
has expression profiles where we can see the change in expression of genes across different
treatments and we can also analyze this expression data. There is a tool called as GEO2R, we
can use BLAST in it. (We’ll discuss later)
http://www.ncbi.nlm.nih.gov/geo/
Gene Architecture:
GEO has four kinds of records or data files (keeping in view the MIAME rules) and are as
follows:
✓ Sample(GSM) – these files stores the sample information like how the samples are
prepared, how the treatments are given, how the experimental design was established.
✓ Platform (GPL) – The idea about platforms, they are stored in GPL files so here we can see
whether it’s a microarray data or RNAseq data (there are different protocols coming from
different agencies so we can have that information).

MUHAMMAD IMRAN 13

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

✓ Series (GSE) – Sometimes different treatments are recorded as separate files so GSE are
the files where we can have the similar treatment files and they are put together in a
shape of series (are a set of samples and which are somehow related).
✓ Datasets (GDS) – Whereas the actual data is stored as GDS files which are the sample data
collections and are assembled by GEO.

Here, is the Gene Expression Omnibus page and if we look into the different types of datasets
it have, we can have Series (on the top left side of right figure), different records for the
Platform, Samples. If you look into the types of series, you can see there are expression
profiling by array, expression profiling by high throughput sequencing (in our course we’ll be
getting some RNAseq data which is under the expression profiling by high throughout
sequencing), similarly there are other various techniques for getting the expression which are

MUHAMMAD IMRAN 14

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

listed below in the Series section as can be seen and number of datasets available are also
present in the
column called as
count.
If you want to look
into some dataset,
you can simply type
into search bar say
for example, you
write colon cancer
RNAseq data which
leads us to the sets of records it gets and when we click onto one of them the page appears
(shown below).

And we get this file, so here is

the information of the
particular dataset.
You can see on the top, that
each dataset is submitted as a
series with a unique number,
say for example here the
number is GSE57043 which is
basically an id number for this
dataset.
So, if we look into this page, we can have the idea about the experiment, the organism from
which it’s coming, type of experiment, a little bit summary of the experiment, we can also see
the contributors name, their publications and addresses.
It is a huge page which is portioned and the other side of it is shown in the figure below

This is the bottom part

of the same webpage
shown above.
Here, we can see that
we need to be associated
with the platforms so
the platform information
is can be found on this
page, and we get those
sequences which are
sequenced by machine
Illumina HiSeq 2000
(we’ll be discussed in
upcoming lectures).

MUHAMMAD IMRAN 15

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

There are total six samples in this dataset, so individual samples are put together labelled as
GSM.
We can also download the sequence expression counts or values in different formats and there
are also some
normalized counts as
shown in the figure,
these are compressed
files. There are also raw
reads data are present in
the format, which we
call as SRP or
Sequencing Read
Archive so that stores
the raw read data.
Since, funding and the publication agencies demands that your data should be submitted and
shared with the community so here is an example (in the figure shown) where we can see a
publication and they have put GSE into their publication which helps other scientists to get
access to this data using this ID number as highlighted (in the figure).If you are submitting
your paper, you need to provide this information to the publication agencies which is an
essential consideration.
Conclusion:
So we sum-up that GEO is a public repository for the archiving and distribution of gene
expression data and is the Best resource to get microarray and Next Generation Sequencing
(RNASeq) data.

Topic -7 Medical Databases

Introduction
Informatics in health care may be called as health informatics
• It deals with the resources, devices, and methods required to optimize the
acquisition, storage, retrieval, and use of information in health and
biomedicine.
(wiki)
Medical databases store and provide medical information
• The premier database for biomedical literature is the National Library of
Medicine (NLM)’s MEDLINE, accessible through PubMed
PUBMED
• Comprises of more than 24 million citations for biomedical literature from
MEDLINE, life science journals, and online books
• Citations may include links to full-text content from PubMed Central and
publisher web sites

MUHAMMAD IMRAN 16

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MEDLINE
• MEDLINE is the primary resource for biomedical journal articles
• Millions of citations to articles in biomedical journals
• MEDLINE uses the MeSH vocabulary
Other Databases
MEDLINE is the primary resource, but other databases may also be helpful
• Academic OneFile
• CINAHL (Cumulated Index of Nursing and Allied Health Literature)
• PsycINFO
• Web of Knowledge
Academic OneFile
• Academic OneFile lists articles from journals covering a broad range of
subjects
• While it does not primarily focus on medical topics, useful articles can
still be found here

PsycINFO
• PsycINFO searches the psychological literature
• While it does not primarily focus on medical topics, useful articles can still
be found here

MUHAMMAD IMRAN 17

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• http://www.apa.org/pubs/databases/psycinfo/coverage.aspx

http://www.apa.org/pubs/databases/psycinfo/coverage.aspx
Web of Science
• Major source for articles in a wide range of fields, including the sciences,
social sciences, and humanities.
• Excellent place to find articles from scientific journals that may not be
included in MEDLINE
Conclusions
Informatics in health care may be called as health informatics
• Medical databases deal with the acquisition, storage, retrieval, and use of
information in health and biomedicine.

Topic -8 Sequence Submission Introduction

Introduction
Sequences are submitted to the databases in order to share them with the scientific community
(sometimes they are also required by the Publication and funding agencies to submit them).
Generally sequences are submitted at the time of publication and are reviewed by peers.
Caution
It is important to ensure that sequence files do not contain any special characters because
sometimes the control characters can also be incorporated into or normal sequences, which can
then mess-up the down-stream analysis or data-transfer.

MUHAMMAD IMRAN 18

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

So, there is an issue of how to put the ambiguous

nucleotides or amino acids in the sequences (because at
some places you are not sure whether it is ‘A’ or ‘T’ or
‘G’ or ‘C’ and you are restricted to put a single letter).
So, there is an organization known as International
Union of Biochemistry (IUB), it has established some
standard codes to represent those ambiguous bases or
amino acids.
For example, here we see that G, A, T or C are just
Guanine, Adenine, Thymine and Cytosine respectively.
If we see R, it can be either A or G and the word is
derived from the group they are coming from i.e. the
puRines. We see Y that is the pYrimidine, it can be C or
T.
M stands for if they are having some amine group / amino group in them, K is if they have
Keto group i.e. G or T.
S is if they have strong interactions (3 hydrogen bonds) like C and G, who forms triple
bonds.
W is for weak interactions, A or T.
Since H follows G in Alphabet so it’s everything except G, it can be A, C or T and similar
procedure is followed for B,V, and D whereas N can be any base.

Similarly, for amino acids, we have

single letter codes i.e. from A to Z. And
we can see in the figure on the left that
some letters are missing.
There are four amino acids that are
starting with G, but we gave that G
letter to Glycine and for rest of them,
we might use some other letters like
Glutamic acid is represented as E.
Y stands for Tyrosine (down below) and
X can be any amino acid like N (in case
of the nucleotide sequences).

NCBI:
NCBI has two options for sequence submission
BANKIt - for simple sequences (not related with down-stream analysis) and annotations
and can be submitted through web (if the datasets are small) which does not requires any
advanced tools.
Sequin - For Complex sequences and annotations and is also good if we want to do some
off-line submissions normally where we have our datasets which are huge ones and can be
used in future with some advanced tools (for analysis) and graphical reports.

MUHAMMAD IMRAN 19

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

http://www.ncbi.nlm.nih.gov/WebSub/?tool=genbank
In the figures above, are the glances BankIt and Sequin webpages.

UniProt:

For protein sequences, just like NCBI tools, we have UniProt and the similar tool is called
asSPIN which is a web-based tool for submitting directly sequenced protein sequences
and biological annotations to the knowledgebase.
Shown in this figure, is the webpage of SPIN.
We can register here and then we can submit our data.
https://www.ebi.ac.uk/swissprot/Submissions/spin
Conclusion:
We conclude that sequences are stored in databases in specific format and when we want to
submit them into a database then we need to follow the guidelines provided by those
databases.

Topic #9 DNA Sequence Retrieval

Introduction
Databases not merely collect and organize data (i.e. not only stores it) but allow intelligent data
retrieval (we can do some down-stream analysis on those data sets). Let’s see how we can get
the DNA data from the NCBI.

MUHAMMAD IMRAN 20

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

So, here is the webpage of NCBI, for example you want to search for say p53 gene; tumour
suppressor gene. We write p53 on the search bar, then we get then results, so here we can find
many ID entries like 9000 entries are there, we are just looking into the first page in this we
choose the first two. So let’s click the first one, the p53 where the ID is 2768677, there is a
description that what sort of gene is it, and its actually coming from Drosophila melanogaster,
the location is Chromosome number 3 and we see some Aliases; the alternative names of this
gene. The link to NCBI is http://www.ncbi.nlm.nih.gov/.

When we clicked on the first gene as shown in the figure above, we now come to this webpage
which is a huge page that is portioned into different figures.
In this figure (on the left), we can see the summary of this gene.
The official symbol is p53 provided by FlyBase which is also written in the Primary source
(FlyBase is the databases that stores the genome of this fruit-fly Drosophila), then the locus
tag, gene type is protein coding, RefSeq says reviewed (sometimes the genes are submitted and
reviewed by some other scientist so it means that this gene has been REVIEWED). In the

MUHAMMAD IMRAN 21

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

organism section, we see the classification of that organism and the Aliases are written beneath

it.
In this figure, we can look into the structure of this gene and its coordinates (genomic
coordinates), where we can see the location from where it is coming from, we can also see the
orientations- the directions in which it is going (down below).

In this figure, we can see the

genomic region, the transcripts
and products tabs, we can look
into the products of this gene
(the gene when is expressed,
the DNA is converted into the
RNA). Since it’s a eukaryotic
genome where there is
alternative splicing, so we can
find different alternative splice
variants of this gene.
On the upper right side of the figure, it is written as Go to nucleotide, Graphics, FASTA and
GeneBank , so these are the different views with which we can get access to data files
associated with this gene. When we click GeneBank, we are guided to another page, shown in
the next figure.

We can see the entry in GeneBank and

how does it look.
Here, again we see the name of the
gene, locus (where it’s ID is written),
length of the gene (it is 4426 BP),
DNA, it is a linear type of DNA then
we have the submission date.
Then the definition which is describing
the organism’s name, chromosome
from which it is coming, then it has
accession (the regions of the genome
from which it is coming from), then we have the version (which is NT_033777.3, so there
should have been .1 and .2 and since this is the third review, we can see .3 version here), we
also see the reference (down below) and the authors from which this gene is coming and then
their publications (it was seen to be published in Science).

MUHAMMAD IMRAN 22

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

When we scrolled down,

we can see in the figure on
left, that there are features
of this gene so the total
length of the gene is 4426.
We can see mRNA (down
below), and since it’s a
eukaryotic gene, so mRNA
is coming from the exons
and the regions from which
it came are shown below
with the word join.

Then, within this mRNA

we find the coding
sequences (shown in the
figure on the left), where
coding sequences are the
parts of the mRNA
which are translated into
the proteins so there are
further sub-sets within
those mRNA regions.
Down below, we see the
translated version where
we can see the word
written as translation,
and here we see the
amino acid sequences
coming from this gene.

In the end, till we reach the word called as origin, and here we can see the actual nucleotide
sequences which are present starting from 1 until the last nucleotide and the sequence ends
with a double slash sign (//).
Conclusions:
So, we conclude that DNA Sequences are stored in DNA sequence databases in specified
formats and Genebank format is a standard format.

MUHAMMAD IMRAN 23

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

TOPIC no.10 Protein Sequence Retrieval

Protein Sequences:
Now we talk about the data retrieval and first we’ll talk about the protein sequence
retrievals and structures. Protein data is of the following types:

• Actual sequences (from the proteomic data or some other experimental

techniques) or translated sequences (sometimes, we go to nucleotides
databases, we get those nucleotides and then we translate them by using some
softwares, so these are kind of predicted protein sequences) .
• Structures (we can also make structures from those proteins that maybe
predicted or the real structures coming from various X-ray Crystallography
Techniques).
• Annotations (sometimes, we are interested in the functions of the proteins so
those are stored as annotations).

UniProt (It is an international partnership between PIR, EBI and SIB):

Now as far as the resources are concerned, we have multiple resources for protein
sequences but UniProtclaims to be the biggest and integrated resource whereas
for the structures PDB seems like a good resource.
As shown in his figure, is the webpage for data retrieval from UniProt, so we

want to search a protein, say

p53, where we put it into the search box and press enter which gives us the
output. And we see that there are 18,000 different records and it is showing us the
first 25 out of them.
We can have different columns for the output on this webpage so we can have
entry; it’s ID, entry name (the Suffix Human is written so it’s coming from
Human, it can be from mouse, rat and Arabidopsis), the protein name is Cellular
tumour antigen, then gene name which is TP53 (where TP stands for Tumour
Protein), the organism is obviously the human (here) and in the end we have it’s
length i.e. 393 bp (base pairs).
The link to this webpage is http://www.uniprot.org/uniprot/.

MUHAMMAD IMRAN 24

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

So, let’s check the first one and here we reach on the record for this protein

(shown in the figure on the left) which is Cellular

Tumour Antigen p53 protein, commonly known as TP53.
We can have different tabs, showing us the outputs. We can look into the
functions, its taxonomy, and lot many other characteristics so if we look into the
function so it gives us some description about what it’s doing.

After scrolling the same webpage (shown in the figure on the left), we can see the
feature key and in some site written (there are unique sites in different proteins
which are having some specific properties in them so this is just one amino-acid
present in this protein that interacts with the DNA). Similarly, there are different
metal binding sites and we can see that it’s mainly binding to the Zincmetal.The
number of amino-acids is shown here so these are the regions where it interacts
with the metal.
Down below, we can also see the DNA binding region, for example here, the amino
acids are from 102 to 292 and that is also shown in the Graphical view as well.
GO-Molecular function or GO-Gene Ontologies, so gene ontologies are the
different functional annotation term, there they define different functions, so
amongst them we have molecular functions, biological processes, and we have
cellular components. So here we just see a Molecular function, so it tells us that it
performs the functions as shown in the figure , mainly it’s a ATP binding, it’s p53
binding with various other functions like DNA binding. So all those functions
related to these proteins are present in the heading of GO-Molecular Function.

MUHAMMAD IMRAN 25

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Next, we move on to some other functions, in the Biological process category

(shown in the figure) we see that it is related to Apoptosis (which is a cell death
and it is related to cell-cycle and some other components).
In the below section, we see some enzymes and pathway databases,
andReactomeis a database in which we have a group of reactions which
are categorized so these are the list of those reactions with which it is
related.

When we move
further (as shown in the figure on the left) till we reach its Taxonomy.
On the top, we can see something written as Protein family or group databases
whichis TCDB. Basically, there is another classification in which the proteins are
classified on the basis of being as transporter proteins so it is associated with the
transportation across the membranes and there is 5-digit number, so there is a
specific classification code which is given to each protein, and this protein has the
specific code as shown in the figure.
So then we have the names and taxonomies, where there are protein names, and
thetaxonomyof the individual can be seen in the Taxonomic lineage row. Let’s
see how we reach to its sequence and is shown in the figure below:

In this
figure,
we can
see the
sequence
of the
protein
which is
found to
be at the
end of
the page.
Here,
it says
Isoform 1, so different proteins have different isoforms, different alternative
splice variants so this is Isoform 1 as exhibited by its name which is P04637-1,
and is the kind of first isoform. We can see the sequence of the protein and starts
with a methionine (always a first amino acid in those proteins) and ending at
390TH amino acid. So, it’s a 393 aa long protein and the sequence is right here.
You can click on the FASTA button on the top and then you can get this output in
FASTA format (we’ll discuss it later).
NCBI:
We can also get the same protein from NCBI (as shown in the figure on the left)

MUHAMMAD IMRAN 26

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

In NCBI,
obviously
the
sequence is
pretty
similar and
the
arrangement is slightly
different so it is
ORIGIN, where the sequence starts and sequence ends at those two slashes
(//). So, we can get the protein sequence from NCBI as well and the link to this
website is http://www.ncbi.nlm.nih.gov/.
PDB:

PDB gives us the structures, so we can go to PDB webpage (as shown in the
figure on the left) and search for the same ID i.e. P04637 and it gives us the sections
or the regions from where it can make up some specific structures.
You can see the turns in Annotations section, the black ones are the empty lines
where no secondary structure can be formed, blue ones show those bends and the
orange ones are designated as alpha helices regions. So in PDB, we can have
structures in this format as well as the 3D-Structures as shown in the figure
below:

The link towards PDB is:

http://www.rcsb.org/pdb/explore/remediatedSequence.do?structureId=2LY4
&bionumb er=1
Conclusions:
We conclude that:
• UniProt is the integrated resource between PIR, EBI and SIB and
• PDB is a good resource to get the protein structure.

MUHAMMAD IMRAN 27

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

TOPIC # 11 Sequence Formats

Sequences are stored in different formats in databases and since software requires those
sequences in specific format so it’s good to have an idea about what major formats are, we’ll
look into few of them.
FASTA Sequence Format
FASTA is the most recognized and well distributed format to present DNA and Protein
sequences.
The sequence starts with a ‘greater than’ sign (>), whereas the actual sequence is always on
the next line. It is recommended that all lines of text should be shorter than 80 characters in
length (generally we have 60 characters).
Example:
>gi|568815581:c7687550-7668402 Homo sapiens chromosome 17, GRCh38 Primary
Assembly
GATGGGATTGGGGTTTTCCCCTCCCATGTGCTCATCTAGAGCCACCGTCCAGGGAG
CAGGTAGCTGCTGGGCTCTCCACGACGGTGACACGC--------
>gi|120407068|ref|NP_000537.3| cellular tumor antigen p53 isoform a [Homo sapiens]
MEEPQSDPSVEPPLSQETFSDLWKLLPENNVLSPLAPPVLGFLHSGTAKSVTCTYSPAL
NKMFCQLAKT--*
This sequence is of DNA in the fasta format (shown above), which starts with the ‘greater
than’ sign (>), same as in the case of the protein sequence in the fasta format (shown below)
that also starts with the same symbol.
Then we have ‘gi’ written which stands for ‘gene identification’ and the numbers shown are
the ‘ID’ in both of the sequences.
In the DNA sequence, we have ‘c’ followed by the ‘ID’, this ‘c’ basically means the sequence
is of the complementary strand and the regions from where it is coming are designated here;
the base positions in between them, this gene is located. Then we have a short description of
this gene that it belongs to ‘Homo sapiens ’, ‘chromosome 17’ and the ‘Primary Assembly’
(assembly is where we get short sequence reads or small sequences and we put them together
into a gene, known as assembly). Then finally, we have the actual sequence which is around 60
characters long in each line (as the sequence was quite long, we have used dashes to represent
further characters).
In the protein sequence, we have ‘ref’ followed by the ‘ID’, which gives us an idea that it is a
reference sequence (reference sequences are the curated sequence, there is a sub-section in
NCBI called as ref seq, so they have reference sequences ; a kind of standard sequences to
avoid any kind of redundancy. So, we can say these are the primary or the main sequences and
we might have other alternative splice variants but references are the kind of true
representative of the class). Followed by ref, we have another ID, which is the ‘protein ID’.
Then we have its brief description that it’s a ‘cellular tumor antigen’ protein ‘p53 isoform’ and
is also from ‘Homo sapiens’. Finally, we have the actual sequence of this protein and in the
end we have dashes that represents it is an incomplete sequence and steric (*) is shown
(sometimes the steric (*) is found to be seen in fastafiles but sometime it don’t, so the software
must know what does this specific steric (*) stands for).
GeneBank Sequence Format:
GeneBank sequence format is found to be in the GeneBank Database which is a kind of
standard format and other formats are pretty similar to it.
A sequence file in Gene Bank format can contain several sequences. One sequence starts with
a line containing the word LOCUS and a number of annotation lines. The start of the sequence
is marked by a line containing "ORIGIN" and the end of the sequence is marked by two
slashes ("//").

MUHAMMAD IMRAN 28

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Example:
LOCUS AAU03518 237 bpDNA PRI 04-FEB-1995
DEFINITION Aspergillusawamori internal transcribed spacer 1 (ITS1) and
18S rRNA and 5.8S rRNA genes, partial sequence.
ACCESSION U03518
BASE COUNT 41 a 77 c 67 g 52 t
ORIGIN
1 aacctgcggaaggatcattaccgagtgcgggtcctttgggcccaacctcccatccgtgtc
61 tattgtaccctgttgcttcggcgggcccgccgcttgtcggccgccgggggggcgcctctg
121 ccccccgggcccgtgcccgccggagaccccaacacgaacactgtctgaaagcgtgcagtc
181 tgagttgattgaatgcaatcagttaaaactttcaacaatggatctcttggttccggc
//

Here is the GeneBank format, which starts with the word ‘LOCUS’, then we have it’s ‘ID’, it
is 237 base pairs long, we have some short description that it is a ‘DNA’, ‘PRI’ – primary
sequence, submitted on ‘04-FEB-1995’.
Then we have a ‘DEFINITION’ line where we can have some description/explanation about
this gene. Then again we have an ‘ACESSION number’. It also provides us with the ‘BASE
COUNT’ (i.e. how many A’s (Adenines), G’s (Guanines), C’s (Cytocines), T’s (Thymines) are
there).
Then finally the word ‘ORIGIN’ tells us that the actual sequence is right here, we have these
lines (60 bases on each line) that are separated into chunks of 10 bases and is a kind of
standard practice. The sequences ends with the those slashes (//).
EMBL Format:
This format is similar to that of GeneBank Format. An example sequence in EMBL format is:
ID AA03518 standard; DNA; FUN; 237 BP.
XX
AC U03518;
XX
DE Aspergillusawamori internal transcribed spacer 1 (ITS1) and 18S
DE rRNA and 5.8S rRNA genes, partial sequence.
XX
SQ Sequence 237 BP; 41 A; 77 C; 67 G; 52 T; 0 other;
aacctgcggaaggatcattaccgagtgcgggtcctttgggcccaacctcccatccgtgtc 60
tattgtaccctgttgcttcggcgggcccgccgcttgtcggccgccgggggggcgcctctg 120
ccccccgggcccgtgcccgccggagaccccaacacgaacactgtctgaaagcgtgcagtc 180
tgagttgattgaatgcaatcagttaaaactttcaacaatggatctcttggttccggc 237
//
Here, we have ID, accession number (AC), descriptions (DE), and the sequence actually starts
from where the word ‘SQ’ is there, and we can observe that we have pretty similar lines as
seen in the previous example. Finally, the sequence ends with doubles slashes same as in
GeneBank format.
SwissProt Format:
SwissProt protein sequence format is similar to EMBL format but there is considerably more
information about physical and biochemical properties of a protein (as you can see below there
is more description).
ID - Identification.
AC - Accession number(s).
DT - Date.
DE - Description.
GN - Gene name(s).
OS - Organism species.
OG - Organelle.
OC - Organism classification.
RN - Reference number.
RP - Reference position.

MUHAMMAD IMRAN 29

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

RC - Reference comments.
RX - Reference cross-references.
RA - Reference authors.
RL - Reference location.
CC - Comments or notes.
DR - Database cross-references.
KW - Keywords.
FT - Feature table data.
SQ - Sequence header.
// - Termination line.
XML Format:
It is a modern practice in which we try to put those sequences in kind of a machine language.
So, XML stands for Extensible Markup Language. The format is similar to HTML (language
for Web programming).
The good part is that this language is in between machine and man readable so it’s kind of easy
to code over this.
And it is becoming standard data format for transferring genome data.
Example:

<xsd:annotation>
<xsd:documentation>
XML Schema for SBOL core data model compatible with RDF/XML serialization.
<dc:date>2012-01-19</dc:date>
<dc:creator>EvrenSirin</dc:creator>
<dc:contributor>Michal Galdzicki</dc:contributor>
</xsd:documentation>
</xsd:annotation>

This format seems pretty weird but not for the people with computer science
background.NBRF Format:
>DL;seq1
seq1, 16 bases, 2688 checksum.
agctagctagctagct*

>DL;seq2 seq2,
16 bases, 25C8 checksum.
aactaactaactaact*

The format is pretty similar to fasta but in addition to that it gives us the checksum value
(checksum- we take those nucleotides and since we know that in computers every digit is
related to some ‘ascii’ value, we can take those values and add them up together and then we
can come up with this number known as checksum. So, it’s a good thing to have this number
as when somebody is downloading the sequence, he can again check on his computer and find
the checksum, if they are equivalent to one another, the sequences are correctly downloaded
otherwise there must be some issues with the downloading)
GCG FORMAT:

GCG stands for Genetics Computer Group (basically it was a group of scientists who were
helping the biological community to develop different software and training programs to help
with the biological sequence analysis problems, so they also came up with the sequence
formats). This format is kind of similar to the NBRF format (we have checksum but we don’t
have greater than (>) sign as in fasta, we have length of the sequence). There can be multiple
sequences in one file.
Example:

seq1 seq1 Length: 16 Check: 9864 .. 1 agctagctagctagct seq2 seq2 Length: 16 Check: 9672 ..
1 aactaactaactaact

MUHAMMAD IMRAN 30

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Sequence converters:
Sometimes, we need to convert between sequences so you can come up with your own script
or you can come up with your own codes and there are also some programs meant for this
purpose alone such as READSEQ is a useful sequence converter (developed by D.G.Gilbert at
Indiana University, USA) basically it recognizes DNA or Protein sequence file and
interconvert them between different formats.
Conclusions:
What we conclude in the end of this lecture is the following:
• Databases store sequences in specified formats
• Genebank, DDBJ and EMBL has similar formats
• Different software need sequences in different formats

We might convert the sequences into other formats on our own or we can also simply use one
of the programs available for converting like READSEQ

Topic -12 Data Retrieval

Data Retrieval:
Nearly all biological databases are available for download as simple text (flat) files. Sometimes
we are interested to download the database and do the analysis locally in our own machines
which might save our time as the local version of the database allows one greater freedom in
processing the data.
ENTREZ:
It is an integrated search engine that works behind NCBI, so you can do lot of researches and
can look for variety of data using it (It can be accessed from the site
www.ncbi.nlm.nih.gov/Entrez/). It integrates PubMed and 39 other scientific literatures,
nucleotide and protein databases. For example, it can be protein domain data, population
studies, expression data, pathways, genome details and taxonomic information.

Here, we can see it integrates between GEO (gene

expression sets), OMIM (Online Mendelian Inheritance in
Man), Genome Databases, taxonomy Databases, etc. And
we can see that in the middle we have Nucleotide, PubMed
and Protein. So it is an integrated system which operates
between different databases, so you can simply search for
whatever you are looking after and ENTREZ will search
it for you.

MUHAMMAD IMRAN 31

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, is the page of ENTREZ that allows you to search anything by the help of a search bar at
the top. It has different connections like we have Literature resources, we have Health
Databases, Genomes, different Genes Databases, Proteins and Chemicals.
Bulk Data Retrieval:
Sometimes, we need to obtain data in bulk amount and for this purpose normally we use Linux
but for Windows users, there are some packages or programs available and are known as FTP
clients so the best option is to use FTP (File transfer protocol). The File Transfer Protocol
(FTP) is a standard network protocol used to transfer files Via command line or application
programs like FTP clients (we’ll be using it).
Once, we get the data which is mostly not in a proper format and every other software require
different specific formats so we might want to use some programming languages to help
convert the data into the required format. The programming languages like PERL and Python
are good for processing Biological data in Bioinformatics.
Conclusions:
We have learned that :
• Data is transferred over the internet.
• Data needs to be transformed or processed before handing it over to any software.

Topic # 13 Genome Informatics

Now we talk about the main subject which we are seeking in this course that is
the Genome Informatics and let’s look into it.

Genome Informatics:
It is about the Genome sequencing that provides the sequences of all the genes of an organism.
The major application of Bioinformatics is the analysis of full genomes that have been
sequenced. Whereas the challenge is to identify those particular genes that are predicted to
have a specific biological function.

Genomics definition:
NHGRI (National Human Genome Research Institute) defines Genomics as:
“Study of all of a person's genes (the Genome), including interactions of those genes with
each other and with the person's environment.”
So, Genome Informatics can be defined as:
It is the field in which computational and statistical techniques are applied to derive biological
information from genome sequences.

MUHAMMAD IMRAN 32

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

“Genome informatics includes methods to analyze DNA sequence information and to

predict protein sequence and structure.”
(Iossifov, et al. 2014)
Genome Sequences:
So, after we have the genome sequences, we do the analysis of those sequences and it includes
the following:
• The discovery and utilization of sequence polymorphisms (different sequence vary
from one another, so we can identify those polymorphisms).
• Opportunity to explore genetic variability both between and within the organisms
(we can help identify different traits of those organisms by using those
polymorphisms).

Genome Analysis:
In Genome Analysis we mainly perform the following tasks;
Sequencing

➢ Assembly (since the sequencing is done in a way that whole genome is broken
down into short fragments and once those fragments are sequenced, we need to put
them together, this step in genome analysis is known as Assembly).
➢ Repeat identification and masking out (once we assemble that genome, we try to
find out the regions in which we have large number of repeats because assemblies
jumble up where we have those repeats so we need to find those regions and it is
one of the important task to go and look into those assemblies while keeping in
mind those regions in which we have those repeats).
➢ Gene prediction (after we have assembled a finished genome, now we can go for
the prediction of the genes where we can find the genes by using different patterns
or features of those genes).
➢ Looking for EST (Expressed Sequenced Tags) and cDNA (complementary DNA)
sequences.
(EST and cDNA are basically originated from the DNA where the genes that are
expressed are transcribed into mRNA which is then reversed transcribed back into
cDNA. So by the help of cDNA, we can look into where those expressing regions
are present in the genes that will give us the idea of the gene expression or the
regions from where the mRNAs are made).
➢ Genome annotation (in which we can find out similar functions performed by
different genes)
➢ Expression analysis (once we have the idea about the regions of the gene in which
we can have the gene expression then we can explore the quantification i.e. how
much those genes are being expressed).
➢ Metabolic pathways and regulation studies (once those genes are expressed, their
products interact with each-other and then they perform different metabolic roles in
the shape of different metabolic pathways and networks).
➢ Functional genomics (where we are actually looking into the different functions
performed by different regions of the genome that are under the control of different
genes and what exactly would be the effect of changes in those genes specifically if
we want to study about the genes related to diseases).
➢ Gene location/gene map identification (map the location of those genes on the
chromosomes).
➢ Comparative genomics (in which we can take one genome and compare it with
another genome, where we can find the comparative features; what is present in the
first genome and not in the second one and the intersections between them, etc.).

MUHAMMAD IMRAN 33

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

➢ Identify clusters of functionally related genes (those genes they might be having
similar structures, sequences and also performs similar functions, which can give us
the idea about the evolution).
➢ Evolutionary modeling (so the identification of the clusters of functionally related
gene can help us in making an evolutionary model).
➢ Self-comparison of proteome (sometimes we are interested in finding genes which
are kind of duplicated within the same organism, so in order to do that this self-
comparison of proteome is made, where proteome is the collection of the proteins
which are derived from those genomes. Therefore, the whole collection of one
organism’s proteins can be termed as proteome and we can compare it with itself
and can find about those sequences which are being duplicated in it).

• Model organisms:

Most of the times, while we are doing those genome sequencing projects, our objective is to
find the cure of some disease, or improving some variety of the crop for enhancing its
production, or looking into some drugs against different organisms so it’s a good idea to have
some model organisms that can be used for studying various processes in labs and there are is
a range of model organisms which includes:
➢ E. coli – bacteria
➢ S. cerevisiae – yeast
➢ C. elegans – worm
➢ D.melanogaster – fly
➢ Daniorerio – zebrafish
➢ Musmusculus - mouse
➢ Homo sapiens – you and me
➢ Arabidopsis - plant

Here, in this
diagram we see
auniversal tree of
life that has been
made with the
structures of small
ribosomal RNA
unit. It divides
whole living
organisms into three
groups, we have
Bacteria at the top,
we have Archaea(they are special organisms that lives under hard conditions) and then we
have Eucarya (which is obviously the biggest among all groups). We pick those model
organisms from important branches of this tree of life so for example, E-coli is shown,
yeast as an example of fungi, from animals we have worms, flies, fish, mice, and
Arabidopsis, rice, soybean are the examples from plants. So, we try to get these organisms;
best representatives from different classes from important branches on this tree of life.
Conclusions:
We conclude the following:
• Sequencing and analysis of full genomes paves the way for future discoveries

• Different model organisms are best source to explore our Genome and to interpolate
the results towards the higher organisms.

MUHAMMAD IMRAN 34

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic -14 Prokaryotic Genome

Now, we study the prokaryotic genome, prokaryotes are the organisms whose Genetic
material (DNA) is not enclosed in a nuclear membrane, so there is no nucleus in them. As
there is no nucleus in prokaryotes, there is no justification to have other membrane bound
organelles. These are relatively simple cells.

Here, in this diagram we see a prokaryotic cell which is a

bacterium (here). We have a genome (DNA) in the shape
of a big chromosome in the middle, and ribosomes (small
structures important for protein synthesis that occurs in
every other organism so ribosomes can also be seen here).
It’s relatively simple cell, having cell wall with different
layers.

Here, in this diagram we see

a comparison between a
eukaryotic cell and a
prokaryotic cell. We can
clearly see the membrane
bounded organelles in the
eukaryotic cell, like
mitochondria (involved with
the respiration process; food
is broken down into the
energy. There is a
hypothesis that mitochondria
actually evolved from
bacteria and is known as
endosymbiont hypothesis). Here we can also see the difference in the size of both cells, so
eukaryotic cells are complex and bigger than prokaryotes.
The first prokaryotic genome sequenced was that of Hemophilus influenza (we have seen in
the previous section) and this organism was sequenced in a relatively moderate cost and with
an efficient pace that paved the way for sequencing of many other organisms. So study of
those prokaryotic organisms is important.
Selection Criteria:
There are different criteria for the selections of the organisms which are then send to the
genome sequencing projects. Following are the criteria for selection:
➢ They had been subjected to a detailed biological analysis/ extensive studies and
thus were model organisms.
➢ They might be important human pathogens (so it’s important for us to study its
genome).
➢ They were of phylogenetic interest.
➢ Sequences were annotated as they were sequenced.

MUHAMMAD IMRAN 35

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, in this table,

we see different
prokaryotic
organisms;
representative
prokaryotic
organisms and
their genomes.
E-coli that was
sequenced by
Blattner et al, the
phylogenetic group
is bacteria, genome
size is 4.6 Mbp (4288 protein encoding genes), and the Novel functions or description of E-
coli is that it is a model organism.
In phylogenetic section, we just have one Archaea whereas rest of them are bacteria.
Methanococcus is an archaea with genome size of 1.66 Mbp (1682 protein encoding genes)
and it grows at high temperature and pressure and produces methane (maybe a good source to
have natural gas from it). Hemophilusis a bacterium with genome size of 1.83 Mbp (1743
protein encoding genes) and is a human pathogen. Mycoplasma is another bacterium with
genome size of 0.82 Mbp (676 protein encoding genes) and is also a human pathogen that
grown inside cells; metabolically weak. In the end, we have Synechocystiswhich is again a
bacterium with 3.57 Mbp (3168 protein encoding genes) genome size and is an ancient
organism that produces oxygen by light-harvesting; may have oxygenated atmosphere.
Conclusions:
We conclude that:
➢ Prokaryotes are simple Genomes.
➢ They are easy models to study Biochemistry, physiology and Molecular biology of
life processes.
➢ Sequencing is done on economically important organisms (i.e. first it’s
implemented on simpler genome which is then used to explore complex genomes).

Topic -15 Eukaryotic Genome

Introduction:
We have seen that prokaryotes are simple genomes in comparison to eukaryotes which are
relatively complicated. So distinct properties of eukaryotes are mentioned below:
➢ Eukaryotes have larger genomes
➢ Have tandem repeats
➢ Have introns in their protein-coding genes (introns are between the exons which are
the protein coding regions within the genes).
➢ Heterochromatin and euchromatin region (eukaryotes have complicated genome, so
the chromosome is grouped as heterochromatin; densely packed region and
euchromatin; lightly packed region).

MUHAMMAD IMRAN 36

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, in this diagram we can see a typical eukaryotic cell which is pretty stuffed as
compare to prokaryotic one. We have nucleus in the middle, channels coming out of
the nucleus known as endoplasmic reticulum (helps in transportation), ribosomes (for
protein synthesis), mitochondria (energy synthesis), we can also see the cytoskeleton
that makes the structure of this cell intact and Golgi apparatus (are concerned with the
secretion). So complicated membrane bounded organelles are present in the eukaryotes

Here, in this diagram we see the connection between the DNA and the chromosomes.
On the left-hand side, we see a DNA strand that is a 2nm wide strand. So, the DNA wraps over
the protein complex molecules (histones - labelled as 1, 2, 3...), and this structure is known as
nucleosome. Then these histones, turn around and makes a wider structure and it makes a 3nm
filament (in third section). Then these nucleosome structures supercoil on their selves to make
those further bigger fibres and until they reach the chromosome the width is 1,400 nm. So, if
we look into the chromosome, we can recognize that there are different arms in it, which are
known as sister chromatids (remember this is just one chromosome but we have two
chromatids), somewhere in the middle we see a constricted part known as centromere, whereas
the terminals are known as telomeres, (remember these nomenclature while we are discussing
the heterochromatin and euchromatin parts).
Staining with dyes:
So chromosomes if stained with the dyes, they give different coloring patterns, we can come
up with the following:
• Dense heterochromatin (dense regions obviously take more color)

• Light Euchromatin (light regions take less color)

MUHAMMAD IMRAN 37

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

If we look into the gene expression, the heterochromatic regions are packed so the enzymatic
machinery cannot reach there; hence these regions are poorly transcribed (expressed).
Whereas, the euchromatic regions are highly expressed because they are loosely packed and
the enzymatic machinery can easily reach to them.

Here, is the diagram in which we can see the relationship between heterochromatin and
euchromatin. You can look into these nucleosomes (combination of DNA and histones), which
are quite jammed packed with one another, so definetly the enzymes cannot access the DNA
which is embedded inbetween. There are different modifications on the DNA or histones that
bring about those structures (we can see on the top), so for example there are histone
methylations (in which methyl groups are added to those histones) in that case the system moves
towards downside (as shown in the diagram) so it becomes euchromatin and similarly, we see
that there are some other methylations on some other aminoacids that can move back into the
opposite direction also. So, histone methylation, histone deacetylation and there are some other
complex proteins which gets attached and give us this heterochromatin region and in the
reverse process, we get the euchromatin region. In the Euchromatin region, the histones are
quite spaced and DNA can be acessible. So, this is the reason why the euchromatin region is
expressed more as compared to the heterochromatin region.
Conclusions:
We conclude that:
• Eukaryotes are distinguished by the presence of prominent nuclei

• Eukaryotes have larger genomes, tandem repeats and introns in their protein-coding
genes (i.e. they are complicated).

Topic # 16 Epichromosomal Elements (EEs)

Introduction:
Genome is the total collection of genetic material and is made up of
➢ Chromosomes
➢ Epichromosomal Elements

Prokaryotic EEs:
1. Plasmids
2. Self-replicating
3. Additional rings
4. Bacteriophages
5. Host colonization

MUHAMMAD IMRAN 38

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

6. Transposons
7. Parasitic DNA elements

Eukaryotic EEs:
Eukaryotes have extra organelles that contain the genome (DNA) which we call it as
Organellar DNA.
Examples are:
Mitochondrial DNA (both in animals and plants), Chloroplasts DNA (in plants), these are
membrane-bound organelles and they may be present in hundreds to thousands of copies (so
there is also multiple copies of these genomes). Mitochondrion is the site for respiration
whereas chloroplast is the site for photosynthesis. Their DNA’s can be labeled as mtDNA or
cpDNA respectively.
Plasmids, yeast, Transposons, Viral genomes and retroviruses are other examples of organellar
DNA.
Endosymbiont hypothesis:
How do these organelles evolved?
So there is a hypothesis known as Endosymbiont hypothesis. According to this hypothesis,
these organelles originated as separate prokaryotic organisms that were taken inside a
primordial eukaryotic cell. Such symbiotic relationships in which two species are dependent
upon one another to varying extents served as crucial elements of the evolutionary progression
of eukaryotic cells.
This hypothesis was originally proposed in 1883 by Andreas Schimper, but extended by Lynn
Margulis in the 1980s.
So according to this theory, Mitochondria and Chloroplastare derived from endosymbiotic
bacteria (that got incorporated into the cells).
Organelle Genome:
Organelle genome (of mtDNA/cell or cpDNA/cell) features are as following:
➢ Circular
➢ Double stranded
➢ Supercoiled
➢ No histones
➢ Multiple copies

Here, in this table we can see the size of these genomes. For example, plant genome is 150kb
circular genome, plant mitochondria is 150-2000kb multipartite, human mitochondria is
17kb circular and saccharomyces mitochondria is 75kb circular.
*Mostly the genomes are circular.
As far as the expression of these organelle genomes is concerned, it has been observed that
their functions are actually dependent on nuclear genomes (they cannot make functions for
themselves).

MUHAMMAD IMRAN 39

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

They encode only a subset of genes required to elaborate a functional organelle like rRNAs,
tRNAs, ribosomal proteins, membrane-associated respiratory or photosynthetic components.
Other components which are encoded by nuclear genome are translated in the cytosol of the
cell and are imported into the organelle. It has been observed that 10% of nuclear genes are
devoted to mitochondrial function whereas 15% to plastid function.
Conclusions:
We conclude the following:
• Organelle genome is similar to prokaryotes.
• It is in high copy number.
• The mtDNA and cpDNA depends on Nuclear DNA (genome) for their function.

Topic # 17 Genome Repeats

We have seen that the major proportion of genomes in the eukaryotes is made up
of repeats (they are also present in prokaryotes). So let’s look into what these
repeats actually are.

Sequence Repeats:
These repeats skew the base composition (normally the A’s, T’s, G’s and C’s relative
proportion is similar to one another but if repeats are present and these repeats are of same
types, for example if we have runs of GC’s, then obviously they’ll change the proportion of
different bases) which can contribute to having differences in there buoyant densities (so those
fragments can then be separated on the basis of those differential densities).
The repeat containing DNA can be separated as satellite DNA on the bases of these densities.

Here, is an example in which we can see the repeats,

they can be Tandem repeats (array of repeats
together), interspersed repeats (those repeats which
are separated by normal genome followed by repeats
i.e. normal genome is between those repeats).
The link to this figure is:

http://mcb1.ims.abdn.ac.uk/djs/web/lectures/repeats1.html#anchor10305
Satellite DNA:
Satellite DNA has following features:
➢ It may be one to several thousand bp long and it can also be present as Tandem; array of
100 million bases long.
➢ They are present near centromere and telomere and
➢ They can be classified as Mini-satellite and Micro-satellite.

Mini-satellite:
Mini-satellite features are as follows:
• They are 15 bases long in array of several hundred to thousands kb.
• They are typically present in euchromatin region.
• Example is VNTR and is used to identify human individuals in forensics.

MUHAMMAD IMRAN 40

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, in this figure we can have

different repeat patterns. In the
individual #1, we have those
GC repeats (runs of GC’s
together), and obviously the
organisms are diploid i.e. they
are having two chromosomes,
from their parents. So, the same
Allele A2 has two of those
repeats, if we look into another
locus (gene position), we can
see there is Allele B2 that has
AGCT repeat (2 copies in both allele) and in the second individual we have 3 copies in 1st allele
and 2 copies in second allele. So, when we digest them with restriction enzyme (restriction
enzyme cut them on the repeat regions), we can have differential banding patterns when we run
them on gel. In this way, we can recognize those individuals. In the figure (below), we can see
the bands, on the left-hand side are the bands of the individual#1 and on the right-hand side are
the bands of the individual#2. In this way, we are getting those repeat regions and by running
them on the gel, we can detect which DNA belongs to which organism.
Microsatellite:
Micro-satellite features are as follows:
1. They are 2-6 bases long and can be in arrays of 10-100 bases.
2. They are inherited to offspring.
3. Mainly they are useful markers for genetic analysis and evolutionary studies (just like in
previous example)
4. They are found in telomeres TTAGGG (one example of six nucleotides segment).
5. SSRs and STRs (Simple Segment Repeats and Short Tandem Repeats are typical examples).

Transposable Elements (TEs):

These are also important kind of repeats and their features are as follows:
1. They are making up the larger proportion of the eukaryotic genome.
2. They thought to play an important role in the evolution of these genomes.
3. They move (Jump) from one location to another even faster than chromosome replicate
(and are known as jumping genes).
4. They have a potential to increase in number,
5. So they make up a large proportion of eukaryotic genome.
6. They are detectable but sometimes they blend into the genome due to mutations and
cannot be detected.

Here, in this diagram, we can see the

proportion of these transposable elements
(red), whereas the green ones are the other
sequences. We have different organisms like
Homo sapiens, Z. mays, Drosophila
melanogaster, Arabidopsis, C. elegans and S.
cerevisiae (yeast).
We can see that in human genome, these
Transposable elements make up 35% (huge
number), in Z. mays they are even more than
that i.e. 50%, in Drosophila melanogaster,
they are comparatively less i.e. 15% genome
(y- axis represents the genome size), and in rest of the organisms, they have less proportions of
those repeats.

MUHAMMAD IMRAN 41

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

So, mainly in humans and Z. mays, they are present as a major proportion.
Conclusions:
We conclude that:
➢ Large proportion of eukaryotic genome is composed of repeats
➢ Different repeats act as markers to detect genetic variation (of organisms) and are
also used to study evolution of those organisms

Topic . 18 Transposable Elements (TEs)

Introduction
Transposable Elements are the elements which can transpose i.e. they can move from one
place to another in the genome and then they can cause the repetitive elements (repetitive DNA
within that genome).
Insertion Sequences (IS elements)
These are the simplest transposable elements and their features are as follows:
➢ They code only for the ability to transpose and are found in prokaryotes.
➢ They are usually very small (< 1 Kb to 2 Kb)
➢ They are flanked by inverted repeat sequences (IRs).
➢ They encode at least one gene that provides their own transposition functions.
➢ They do not code for noticeable (phenotypic) traits.
➢ They can cause mutations by transposition into genes.

Here, we see a structure of

Insertion Sequences (IS
elements). We have Invert
Repeats (IR) which are anti-
parallel or facing each other
(repetitive elements; these are
the distinctions in them). We
have a transposase gene that
gives the transposition
properties to transposable
elements. Another distinction in them is that when they get into the host genomes, they cause
target site duplication- normally they create the sticky ends in the genomes of the host and
later on when the complementary nucleotides are formed, they become those duplications (in
pink).
Transposons:
Transposons are more complex transposable elements as compared to the simple IS (Insertion
Sequence) elements and they code for additional characters in addition to the gene responsible
for their transposition.

MUHAMMAD IMRAN 42

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, in this diagram, we can see that

there are two major classes termed as
Class I and Class II.
Within Class I, we have LTR (Long
Terminal Repeats) retrotransposon
(they have the ability of reverse
transcription, so they have RNA’s
which gets converted into cDNA’s; this
is how they replicate). They have LTR
at both ends (they are not invert but they
are kind of forward repeats and are in
the same direction).
In the same Class I, we have another
category known as Non-LTR (Long
Terminal Repeats) retrotransposon, which doesn’t have those LTR and are of two types
namely LINE (Long Interspersed Nuclear Elements) and are larger in size and SINE (Small
Interspersed Nuclear Elements) and are smaller in size. In Class II, we have DNA transposon.
Eukaryotic TEs:
As we have seen earlier, eukaryotic Transposable Elements has two classes in them; Class I
and Class II.
In Class I, they have reverse transcriptase in them so they use RNA-mediated mechanisms of
transcription. This Class I can be categorized intoLTR (Long Terminal Repeats)
retrotransposons, simple Retrotransposonsand Retrovirus like Elements.
As we have seen earlier, that Non-LTR (Long Terminal Repeats) retrotransposon comprises
of SINES and LINES.SINES are Short Interspersed Nuclear Elements and are 80-300 bp long
and the example is Alurepeats (found in humans). Whereas LINES (Long Interspersed
Nuclear Elements) are 6-8 kb long and are relatively longer than SINES.
The example of Class II TEs is Ac-Ds (which were studied in maize - the work of Barbara
Mcclintock led to the discovery of these transposons where he gave different characteristics to
the kernels of the maize). Another example of Class II TEs is P elements in Drosophila.
Additional elements are known as MITES (Miniature Inverted Repeats TEs) that has the
features of both Class I and II and are 400 bp long.

MUHAMMAD IMRAN 43

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, in this figure, we see

the comparison of human,
yeast, maize and E-coli.
In blue color, we see those
repetitive regions
(transposons), red are the
exons, whereas orange are
introns and brown colored
are the pseudo-genes.
We are comparing the
50kb region of the genome
where in human, we can
see the blue color prevails
i.e. many transposable
elements can be seen. In
yeast the transposons are
less in number. Whereas in
maize, they are found to be
in huge proportion and in
E-coli we can see the IS
(Insertion Sequences)
elements which are
extremely low in number.

Conclusions:
We conclude the following about Transposable elements:
➢ They make up a significant part of organisms’ genome especially in that of the
eukaryotic genome.
➢ They move within and across genomes and
➢ Causes genome expansion.

Topic # 19 Eukaryotic Gene Structure

Eukaryotic genes:
Eukaryotic genes are relatively complicated and are not simple as prokaryotic genes. They
possess exons and introns.
Exons:
Exons are protein coding regions and are interrupted with introns (i.e. in between exons are
introns). During the gene expression, both exons and introns are first transcribed into mRNA
and then the introns are removed out, so the remaining structure is known as ORF (Open
Reading Frames) which only consists of the exons.
Introns:
They contribute a very small proportion in yeast i.e. only 239 introns in its genome whereas in
human the introns makes up 95% of its genome.
The introns stay on the same location and they might also have embedded genes in them.

MUHAMMAD IMRAN 44

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

They can be distinguished by the presence of GT at the 5’ ends and AG towards the 3’ end
(GT-------------AG) and this trend is highly preserved all over the genome.

Here, in this
diagram is the
structure of a typical
eukaryotic gene.
We see the
chromosome and
gene is the region
which has specific
patterns so we can
observe the
promoter region (in
the beginning of a
gene), the blue ones are the exons and those orange ones are introns.
There is start of transcription (marked by black line) which ends at the exon3 (as shown here
and marked by black line), this whole region is then transcribed into mRNA. We can see a 5’-
UTR (Un-translated region) region, so this region is transcribed into mRNA but is not
translated i.e. no protein is formed from this region, similarly we also have a 3’-UTR region.
When we see the ORF i.e. the region from start codon (initiator) to stop codon, and in between
them we can see there are number of amino-acids, so this is the region from where translation
takes place and we get a protein.
After transcription, the transcript is known as Primary RNA transcript and we can see that it
also contains those introns. Which are later on removed through a process called as splicing
and then we get a Mature RNA transcript, so that transcript is then translated into the proteins.
This mature RNA transcript is also recognized by the presence of a poly-A tail (long runs of
A’s)
Intron origin
So, about the origin of introns, there are two theories which are as follows:
• Intron-early- According to this, they used to assemble the genes from already existing
exons (so they brought the exons together and then these structures became the genes).

• Intron-late- According to this, the exons were already present with one another, then
introns got into them (i.e. they Broke up previously continuous genes by inserting into
them).

Number of Genes:
Now we talk about the degree of compactness, so the compact genomes whose size is small
and the relative proportion of gene is higher which contributes to the variation in gene density.
In short, we can say that compact genomes have higher genome density.

MUHAMMAD IMRAN 45

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, in this table, we have

different genomes (mostly
eukaryotes).
We can see that the size of
genome in Arabidopsis is
130MB and number of genes
is approximately around
25,000.
E-coli (prokaryote; bacteria)
can be seen here, with genome
size of 4.7MB and there are
over 4,288 genes.
In humans, the genome size is 3000 MB (it’s not megabytes, it’s mega base pairs), and 45,000
to 120,000 genes (slightly around 30,000 have been identified).
So, if we look into those smaller organisms, like those having smaller genomes (E-coli), and if
we take the number of genes and divide that to the genome size, we’ll see that they have more
densities as compared to the larger genomes.
Pseudogenes:
These are non-functional genes (sometimes there are mutations in the genes and if those
mutations are present in some important regions then the genes’ functions gets knocked out –
known as psudogenes).
There is one category of them and is known as processed pseudogenes that lack introns and
promoters.

Here, is the diagram

in which we see the
normal gene and a
psuedogene.
In psudogene, we
see that the bases
are deleted from the
promoter region (promoter must have specific pattern of bases in them), so in this case, this
deletion is lethal. Similarly, in start codon (initiation code) we can see GTG instead of ATG,
and we can also see the mutations in the splice sites (its GC rather than GT and GA rather than
GT respectively). And in exon2, there are 20 bases which are also been deleted. Hence, this is
how a normal gene has now become a pseudogene.
Gene families:
Sometimes, the set of genes have similar sequences as well as similar functions. And if we try
to find out similar genes within an organism or between different organisms, performing
similar functions can be categorized as gene families. Gene families arise from gene
duplication and subsequent divergence events.

MUHAMMAD IMRAN 46

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, is the diagram showing the

example of gene duplication, we have
human globin gene.
So, first duplications give rise to two
types of globins (one is myoglobin and
other globins).
Within the second type (globins), we
have alpha globin and and beta globin.
So this process is called as gene
duplication.

Conclusions:
In the end, we conclude the following:
➢ Eukaryotic genes have exons and introns.
➢ Introns make up a significant portion of higher organisms’ genome (Human
genome).
➢ Pseudo genes are non-functional genes.
➢ The genes which are similar in function, they make up the gene families

Topic # 20 Comparative Genomics

Introduction:
Comparative genomics is where we compare the genomes of different organisms like we do
comparison of gene number, gene content and gene location in both prokaryotic and eukaryotic
groups of organisms.
The availability of genome makes it possible to have a comparison of all the proteins;
proteome (so we can have the genome and translate them into the proteins or we can get the
actual proteins and can do those protein comparisons – known as comparative proteomics).
Orthologs:
Orthologs are the genes that are present in two different organisms and are so similar that they
must have the same function and the evolutionary history (Fitch, 1970).
Paralogs:
Paralogs are the gene families that originate from gene duplication events (they maybe within
the same organism) over the evolutionary time.

MUHAMMAD IMRAN 47

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

(Unlike pseudogenes, 2nd copy of gene remains functional)

Here, in this diagram we can see
that the similar genes are called
as homologs, which is further
divided into orthologsand
paralogscategories.
The Orthologs, for example if
we start from the bottom, we
can see some globin gene (early)
and then there is a gene
duplication (as seen earlier), so
it becomes a beta chain gene
and an alpha chain gene.
Then those alpha chain genes
can be seen in three different organisms (frog, chick and mouse) and are similar so that’s why
they are known as orthologs. Similarly, those beta chain genes can also be seen in three
different organisms.
Now, if we look into those genes that are similar but gets diverted (and are in the same
organism) as in mouse, chick and frog (we have alpha globins and beta globins) so these
organisms maybe classified as paralogs.
Drosophila and yeast:
When Drosophila is compared with yeast, we see that Drosophilahas core proteome only
twice the size of that of the yeast and Drosophila proteome is comparatively more similar to
mammalian proteomes than worm or yeast.
Drosophila and C. elegans:
Now if we compare the fly Drosophila with the wormC. elegans, we can see that despite the
large differences between them, the core proteome is of the similar size in both.
And nearly 30% of the fly genes have putative orthologs in the worm.
Drosophila and Human:
When we compare fly Drosophila with Human, interestingly some human disease genes are
absent in Drosophila but we can see a number of previously unknown counterparts to human
cancer and neurological disorders genes are present in Drosophila, so it can be used as a good
model in cancer studies.
Conclusions:
We conclude the following:
➢ Comparative genomics reveals the relationship among different organisms.
➢ Fruit fly has more similarities with mammals, so we can utilize it especially as a
model for cancer studies

Topic #21 Comparative Proteomics Introduction

Introduction:
Since, genes encode proteins and these proteins perform actual life functions, we can have the
genomes and get the translated protein from it and by comparing those translated proteins with
one another, we can say it’s a Comparative Proteomics study.
The collection of the protein sequences that are encoded by the genome makes up the
proteome of that individual.

MUHAMMAD IMRAN 48

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, is the figure of

how comparative
proteomics can be
done.
So from the genome
sequence, we can get
the predicted genes
out of it, then we do
translations which
gives us the proteome
of that individual. We
can use those proteins and get some Database searches (we can find the homologous proteins,
whereby we can predict their functions and roles).
We can also do the comparison of proteome by themselves so that will help us finding the
paralogs (studied earlier) within that individual.
We can also do the comparison of proteins between different organisms and we can also find
the clusters of co-related proteins in terms of their sequences and functions.
All against all, self-comparison:
Firstly we talk about the all against all; self-comparison, where we can do the following:
• Comparison of all proteins with each other within the individuals’ proteome.

• Identify unique proteins from the ones having paralogs.

• Identify Gene families

• If we have a good match between query sequence and some other sequence, we can
suspect those two are
the paralogs (because
they are present
within the same
organism).

Here, in this figure

we have the example
of database searches.
We have some
proteins which are
aligned together.
Here, we can see some proteins which are similar to one another and as they are
sharing the similar domains (domains are specific protein structure which is made up of
specific amino-acids into particular structure).We do the sequence similarity searches
by the software named as BLAST (will be discussed later).
We have some parameters (P/E), normally low P/E ratio is taken as a good score, for
example if we have a good match we can have the range of P/E value less than 10-20
and it keeps on decreasing if we go down to other similar sequences (as shown).
Cluster Analysis:
In cluster analysis, we make the groups of proteins which are quite similar to one another. And
reasons of doing it are as follows:
• To sort out the relationships of all the related proteins.

MUHAMMAD IMRAN 49

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Clustering classify the proteins based on some objective criteria e.g

– E value cut off

– Distance in alignment (so the proteins which are more similar will be grouped
together and distant proteins will be grouped from them so in this way we can
have sub-groups or clusters in our data).

There are different clustering methods and are explained briefly in the below section.
Clustering by subgraph:
The way of clustering or grouping by the method of sub-graph is as follows:
➢ Each sequence is a vertex (vertex or vertices are the point or dots by which the edges
(links) in a graph are connected. There can also be a vertex that’s without any edge
connecting to it, known as isolated vertex).
➢ Significant alignment score is an edge (on the vertices, we put our sequences and on the
edges we put our alignment scores).
➢ Trimming by removing weak edges (if we have High P/E ratio, we will remove them).

Here, in this diagram we are

representing the clustering, we have
those vertices (shown as balls) which are
the proteomes here, we have the edges
(lines) which are the connection between
those vertices and are the alignment
scores (the thickness of edges relates
with the small P/E value).
And if they are weakly co-related like the values are greater than 10-6 , so then they are
not connected.
The dashed lines (or dotted lines) are where we have loose connections.
Clustering by Linkage :
There is another technique, where the clustering is done by linkage (almost similar to the
previous ones except some changes). The method of doing it is as follows:
➢ Each sequence is a vertex.
➢ Significant alignment score is an edge.
➢ Trimming by removing weak edges (High P/E).
➢ Or remove > e-6 (we remove the ones which are not linked).
➢ Remaining sub-graph should share 2/3rd of edges.

Single Linkage:
Linkage is done by the following method:
• A group of sequences in all-against-all comparison is subjected to MSA (group those
proteins which are co-related with multiple sequence alignment by first aligning them
and then calculating their distances).

• Create distance matrix (by using those distance calculation just made).

• Neigbour joining is then used to do clustering (by distance matrices, we create those
trees and the method used is Neighbor joining- will be discussed later).

MUHAMMAD IMRAN 50

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, in this figure we have the

single linkage cluster analysis
results, so we have those
proteomes which are present in
the end of those trees which we
call them as leaves. The
closely related ones are
connected with one another.
We can see two types of arrangements, one is circular arrangement (on left) and we have
typical binary tree arrangement (on right).
In binary tree, we can see that we have come up with two big clusters, and within those
clusters we find those further groups or sub-clusters.
Core Proteome:
Core Proteome is when we do all-against-all comparison; which tells us about the proteins
which are duplicated and it also gives us the information about those proteins which are
uniquely present in those organisms- the core proteome.

Here, in this table, we

compare the core
proteome with the
total number of genes.
Here, for example in
bacteria we have 17 hundred genes, 14 hundred are the gene families (represents the core
proteome) and we have number of duplicated genes. In worm, we can see that we have 18
thousand genes where almost half of them are in the shape of gene families and rests of them
are duplicated genes (so almost half of them are duplicated amongst them). In Drosophila, we
have 13 thousand genes, 8 thousand gene families and 5536 are the duplicated genes.
Conclusions:
We conclude the following:
➢ Genome is translated into proteome.
➢ Self-comparison of proteome yields gene families and duplications

Topic # 22 Between-Proteome comparisons

Introduction:
In between-proteome comparisons, we compare the proteomes from different organisms with
one another, so in this way we can find the genes which are similar between those organisms,
and we call them as orthologous genes.
Here we take the proteome as a query and we do a database similarity search against another
proteome or there can be a whole database where we have the set of proteomes.
If the proteome is not available, we can search the EST (Expressed Sequence Tag) Database as
well.
Significance:

MUHAMMAD IMRAN 51

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

As mentioned earlier, this helps in finding the orthologs, gene families and the domains
(between different organisms). There can be other significance of between-proteome
comparisons search and are as follows:
• Proteins that have a highly significant alignment score can be suspected as the orthologs.
• Mostly the proteins that are related to core biological functions (basic functions of life) are
likely to be orthologs.

Finding true orthologs:

How can we find true orthologs?
There is a technique which we call it as:
Method 1:
Where we have the “Reciprocal Hits”, in this method, you take one organism at a time as a
query and search against the other as a database and then you flip around and take first as
database and second one as a query. So, if you are getting similar end results, it means that the
same genes are co-related with one another (or highly similar to one another), and we can keep
them as the best hits.
We can also apply some criteria on hit, for example here we can do a E-value cut off when we
do BLAST (BLAST gives us the parameter called as E-value and a lower value is considered
good and we’ll talk about BLAST algorithm in coming up lectures but here the point is to let
you know that we do some E-value cut off) and we can retain those gene pairs in those
Reciprocal hits.
So, for example here we say that E < 0.01, we can retain them.
Similarly, while we are doing BLAST, since it is a local search tool, we can compare different
regions amongst different genes or proteins, we look for the matches between different regions
and sometimes both proteins are not greatly covered in the alignment, so we want to have at
least some coverage criteria, so here we have like for example 60% coverage.
So we keep those matched pairs normally with a very conservative or low P value like 10-10 to
10-100 .
Clusters of Orthologous Group (COG):
In this way, while we are finding the true orthologs, we can group those organisms which are
similar to one another and this method is known as Clusters of Orthologous Group (COG).
Orthologs are assumed to be derived from common ancestor, so they might also have paralogs
(within the organisms, the orthologs might also go through the duplications).
Orthologs are clustered to form COG (can be studied as COGs).

Here, in this table we

have the example of
the closely related
organisms’ sequences
i.e. Orthologs
(between yeast and
worms).
We have different P
cut-offs (like <10-10,
<10-20, <10-50, <10-
100
).
We have the ‘total number of sequence groups’ and at different cut offs if we go on a stringent
criteria, we have less common orthologous groups and if keep the criteria less strict we can
have more of these orthologous groups.

MUHAMMAD IMRAN 52

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Then we have the ‘number of groups with more than two members’ (as shown).
Lastly, we have the ‘percentage of yeast’ and ‘percentage of worm’ (i.e. how many amongst
the total, they are present) and are presented in these two groups (the yeast and the worm), say
we have 40 percent and 19 percent on <10-10 cut-off P-value, and we have 5 percent and 2
percent on <10-100 cut-off P-value (if our criteria is strict).
So, in this way we can group the similar proteins at different cut offs of P-values, and we can
have the various results.
Proteomes to EST databases:
Sometimes, we take those proteomes and we match them or align them with Expressed
Sequence Tags (EST) (which is cDNA copies of cell’s mRNA sequences). We do this
procedure for those organisms’ genomes whose sequences are not available.
ESTs are single DNA reads and are mostly 3’ biased (since we get them from mRNA and
mRNA extraction protocols relies on getting those mRNA by using their 3-prime poly A-tail
which is present on their 3-prime end, so that is why they are kind of more tiled or oriented
towards 3-prime ends as they are mainly extracted from this site).
EST may be incomplete because it is wholly dependent upon the gene expression, so if we do
not have genomes rather than we only have the ESTs, we might be biased towards only those
genes which are expressed.
The softwareor the package in BLAST which is being frequently used for this purpose is
TBLASTN.
Family and Domain Analysis:
Proteins are organized into domains that represent modules of structure or function (as
domains are specific arrangements of amino-acids). And domain comparison sometimes is co-
related with their biological functions.

Here, in this figure for an

example, we take the
domains from different
proteins and we put them
into Domain Databases,
and then in the end we
can come up with the
shared domains; the
domains which are
present in these different
groups and lastly we can co-relate this information to have an idea about their functions.
Conclusions:
• Proteome comparison helps finding orthologs, gene families and protein domains

• Domain comparison reveals their biological roles.

Topic #23 Horizontal Gene Transfer

Introduction:
Horizontal Gene Transfer is where the genes are transferred at the same levels or horizontally
(so two organisms they transfer their genes and it’s not like transferring the genes from top to
bottom like vertical transfers for example, from parent to the offspring) and is relatively slow
process in evolution.

MUHAMMAD IMRAN 53

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, in this chart we have

the length of the coding
sequences and the
proportions which are
coming from the foreign
elements (in black). (Grey
are the native ones).
So this is an indication of
the horizontal gene
transfer. We can see here
that mostly in E-coli K12,
there is the biggest portion
of foreign gene and in
Synechocystis, we can observe the huge part of foreign genes as well.

Here, is another case in which there

is the comparative map of two
genomes from E-coli 0157-h7 and
K12.
We can see that here are about 1400
genes present in H7 which are not
present in K12.
Similarly there are 500 genes found
in K12 which are not seen to be in
H7.
In this circular arrangement, we see the red which indicates the genes present in H7 but not in
K12. Whereas green are the ones which are conserved in both the organisms.
Here, is the table of another example in
which we observe the horizontal gene
transfer in eukaryotes.
We have different plants, which has
hormone synthases gene (a foreign gene)
and is suspected to be coming from the
bacteria.
The Aphids (insects) which has
carotenoids as a foreign gene and it might be coming from fungal.
Sturgeons have some 15 genes which are foreign and are thought to be coming from
Trematodes.
Sea Slug that has a psoBgene which encodes a nuclear factor and the source is some Alga.
Conclusion:
Horizontal Transfer of genes between different organisms is a relatively slow process that
leads to acquisition of new traits.

Topic #24 Gene Order (Synteny)

Gene Mapping
• Gene mapping is determining the location of and relative distances
between genes on a chromosome

MUHAMMAD IMRAN 54

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Genetic vs physical Mapping

• Genetic map distances are based on the genetic linkage information
measured in Centi-Morgans (CM)
Genetic vs physical Mapping
• Physical maps use actual physical distances usually measured in number of
base pairs
Synteny
• Arrangement of genes on the chromosomes
• Comparing gene orders provides a good measure of similarities and
differences among different organisms
Conserved Synteny
• Species diverged from common ancestor have similar chromosomes and
gene order
• Gene duplication and rearrangements changes synteny
• As a results species diverge

MUHAMMAD IMRAN 55

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 56

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Conclusions

Species diverged from a common, expected to have similar synteny

Functionally related genes stay close as clusters

Topic # 25 Genome Annotations

Genome annotation:
After the genomes are assembled together, the genome annotation is a very important task
whereby we could gain information of the genome. It has several procedures involved and
which are as follows:
➢ We try to locate some important genes which are protein coding genes and also their
products.
➢ We also try locating the RNA-encoding genes.
➢ We recognize the non-coding regions in a genome.
➢ We can also predict the function of the genes.
➢ The Biochemistry and structure of gene products can also be obtained.
➢ We can explore the Literature links.

MUHAMMAD IMRAN 57

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

➢ We can also explore the links to genetic maps where they are located on the
chromosomes.
➢ We can look into the location of the repeats.
➢ We can also look into the location of STS (sequence tag sites).
➢ We can also look into the location of sequence polymorphisms.
➢ And we can find the significant alignment to some protein sequences of known function in
databases (by comparison).

Annotations steps:
Annotations are divided into two types and are as follows:
• Structural annotation

• Functional annotation

Structural Annotation:
Structural Annotation is where we try to identify certain gene features like;
• Promoters
• Terminators
• Shine-Dalgarno sites; the ribosomal binding sites during the protein synthesis)
• DNA motifs (patterns of nucleotides within the genes)
• Co-transcription units
• Operons in microbes (in micro-organisms, lots of genes are transcribed together known as
operons)

Annotations Tools:
There are two tools which are important worth mentioning here, one is MAGPIE and the other
is GENEQUIZ and these are designed to assist with gene the genome annotations.
MAGPIE (Multipurpose Automated Genome Project Investigation Environment) - It’s an
automated genome analysis tool that is used for structural annotation.
GENEQUIZ- Focuses on deriving a predicted protein function based upon the available
evidence; including evaluation of similarity to the closest homologue in the database (i.e. it is
good tool for functional annotation).

Here, in this figure, we have an

outline and is a kind of a work
flow of how exactly MAGPIE
works.
We take some source sequence,
and we give it to some software
program known as Magpie
Daemon; it takes the sequences
(which are added to the database,
so it automatically gets them) and
sends that data to the local tools,
remote tools (over the internet) and then it explores some specific features or annotation
patterns and it puts them into the Feature Database.
Then later on those results are interpreted (Interpretation and Reconstruction) and then we get
the reports.
In this way, we have a kind of automated annotation gathering tool
Functional Annotations:

MUHAMMAD IMRAN 58

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

The attributing biological information to the genes is called as functional Annotations and we
can have it via;
➢ Biological function
➢ Biochemical function
➢ Gene expression (transcription of the gene is considered as gene expression)
➢ Regulation and interactions among different genes

8 Group Classifications:
There are different classification schemes, which are meant for functional classification, to
classify the genes and their products into one of these groups;
➢ Enzymes
➢ Transporters
➢ Regulators
➢ Membranes
➢ Structural elements
➢ Protein factors
➢ Leader peptides (control transcription and translation)
➢ Carriers (transporters)

In this way, scientists have seen that 90% of the E-coli genes fit into these categories
(so their annotations can be explained).
Enzyme Commission (EC) numbers:
It is another scheme which was put forward by the Enzyme Commission (EC) that was
working under the IUBMB (International Union for Biochemistry and Molecular Biology).

They say that the enzymes are classified on the basis of the reactions they catalyze and have a
4-digit scheme which is actually the enzyme commission number: EC a.b.c.d

‘a’ (first digit) informs that it is from one of the 6 classes of biochemical reactions (enzyme
might be coming from one of these classes).
‘b’ (second digit) informs that is from the group of substrate (the thing on which the enzyme
attacks).
‘c’ (third digit) informs us that it is anaccepter molecule.
‘d’ (fourth digit) gives thedetails of biochemical reaction
For example,
tripeptideaminopeptidases
EC 3.4.11.4
Where 3 – tells us that it is a Hydrolase (use water to break substrate).
This 3.4- tell us thatit is aHydrolase that acts on the peptide bonds.
The 3.4.11- tells us that it is a Hydrolase that cleaves the amino terminal amino acids of
polypeptide.
While putting everything together, EC 3.4.11.4- it tells us that it is a Hydrolase that cleaves the
amino terminal amino acids of a tri-peptide.
With Enzyme Commission Scheme, they classified that 70% of E-coli genes shared a and
b (first two classes), which means that they catalyzes the same biochemical reaction.
Three Groups Scheme:
This is another classification scheme known as a ‘Three Groups Scheme’, where we divide all
those functions which are related to the following:
• Energy

MUHAMMAD IMRAN 59

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Information
• Communication

It was found that plants devotes half of their genome to the energy metabolism (they make
food), whereas animals devotes half of their genome to communication (they talk a lot :D )
Conclusions:
We conclude the following:
• Finding genes and their coding regions is an important task in Genome annotations.
• Functional annotations correlate the genes to different classes of functions

Topic # 26 Genome Sequencing

Introduction
Genome sequencing involves recognition of nucleotides in a Genome and determining their
precise order of arrangement. And we have seen that the advances in sequencing technologies
have revolutionized the pace of scientific discovery.
DNA sequencing began in 1970s with the development of Maxam-Gilbert’s method and later
got the pace with Sanger’s method so most of the modern day sequencing employs the variants
of Sanger’s methods.

Here, we see the

first published
nucleotide, which
was published by
Gilbert and Maxam.
They presented the
first 24 base pairs of
the DNA (working
on the lac
Operator).

First genome sequenced:

So first genome was RNA of a virus (phage) MS2 and was sequenced by Walter Fiers and
colleagues at University of Ghent, Belgium (1972-76).We can say that RNA was first
sequenced among nucleic acids (first complete genome sequence made was that of RNA).
Among the free living organisms, Haemophilusinfluenzaewas the first ever published genome
by Fleischmann et al. in 1995. (Since we are in a discussion that viruses are living or non-
living creatures, so if we talk about the living ones, it is the bacterium which was the first
living organism whose published genome was made).
MS2:
MS2 is a virus that infects E-Coli and Enterobacteriaceae. It is a single stranded sense RNA
(in the host it gets duplicated and where anti-sense is formed and from that the RNA copies are
again formed) and it encodes MS2 coat protein.
MS2 sequencing:
The entire nucleotide sequence was established by nuclease digestion and characterization of a
fragment (it had the enzymatic digestion of the nucleotides and that’s how it was sequenced).

MUHAMMAD IMRAN 60

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MS2 genome
MS2 has 49 different codons in the genetic code that specify the sequence of the 129 amino-
acids long coat polypeptide (virus has a coat which is made up of proteins on its outer side and
it has a RNA; it’s genome).
Here, we see the virus genome,
lys which has the genes like mat (helps
mat cp rep in assembly; putting those proteins
5'- bacteriophage MS2 RNA - 3' together), cp (codes for coat
protein), rep (codes for replicase
protein) and lys(codes for lysis
protein; that breaks the host cells). When we observe cp, rep, andlys, we can see the lysgene is
embedded between these two
genes, so we can have the
genes within the genes (here).

Here, is the history of different

milestones, which were done
while we are moving towards
this DNA sequencing.
So we start way back in 1870 with the discovery of DNA, then in 1953, we see the major
breakthrough is Watson and Crick Model

Here, in 1977, we can see

that Sanger’s method
arose. And worth
mentioning here is of
Hood’s name which gave
us the automated sequencer
that also changed the pace
of technologies in this
sequencing business.
Automated sequencing:
Leroy E. Hood's laboratory at the California Institute of Technology invented the first semi-
automated DNA sequencerin 1986, which is the key technology used in Human Genome
Project.

Leroy E. Hood
Institute of system biology
Seatle Washington
Conclusions:
We conclude the following:
• Genome sequencing involves recognition and determining the precise order of
nucleotides in a Genome.
• Advances in sequencing technologies have revolutionized the pace of scientific discover

MUHAMMAD IMRAN 61

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 27 Advanced Computing Approaches

Structure of RNA

Topic # 28 DNA Transcription

MUHAMMAD IMRAN 62

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 29 Protein Translation

Initiation

MUHAMMAD IMRAN 63

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 64

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 30 31 Dynamic Programming

Block Game

0 1 2 3 4 5 6 7 8 9 10

0 W

1 W W

MUHAMMAD IMRAN 65

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

10

Topic #31 Dynamic Programming

Block Game
(i-1, j), (i-1, j-1), (i, j-1)

0 1 2 3 4 5 6 7 8 9 10

0 L W L W L W L W L W L

1 W W W W W W W W W W W

2 L W L W L W L W L W L

3 W W W W W W W W W W W

4 L W L W L W L W L W L

5 W W W W W W W W W W W

6 L W L W L W L W L W L

7 W W W W W W W W W W W

8 L W L W L W L W L W L

9 W W W W W W W W W W W

10 L W L W L W L W L W L

FASTBLOCK(n, m)
1. if n and m are both even
2. return L
3. else
4. return W

Topic # 32 Algorithm for Dynamic Programming

Block Game Algorithm

MUHAMMAD IMRAN 66

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

FASTBLOCK(n, m)
1. if n and m are both even
2. return L
3. else
4. return W

Rn,m
R2,2 = L
R4,4 = L
R4,5 = W

Topic # 33 Restriction Mapping

EcoRI (Escherichia coli)
GAATTC

MUHAMMAD IMRAN 67

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

ATGTTTGCATTACGATAGAATTCCGTCAAAGTGCTAG
TACAAACGTAATGCTATCTTAAGGCAGTTTCACGATC
GCCGTTATACGCTGGATTTAAATTGCTGTGAAATGGT
CGGCAATATGCGACCTAAATTTAACGACACTTTACCA
TACTGCCAAGACCGAATTCCTGCGAGTGCTGAAACG
ATGACGGTTCTGGCTTAAGGACGCTCACGACTTTGC
GCGATATTACGAATGTGCTTACAGCACCGAATTCATC
CGCTATAAAGCTTACACGAATGTCGTGGCTTAAGTAG

ATGTTTGCATTACGATAGAATTCCGTCAAAGTGCTAG
TACAAACGTAATGCTATCTTAAGGCAGTTTCACGATC
GCCGTTATACGCTGGATTTAAATTGCTGTGAAATGGT
CGGCAATATGCGACCTAAATTTAACGACACTTTACCA
TACTGCCAAGACCGAATTCCTGCGAGTGCTGAAACG
ATGACGGTTCTGGCTTAAGGACGCTCACGACTTTGC
GCGATATTACGAATGTGCTTACAGCACCGAATTCATC
CGCTATAAAGCTTACACGAATGTCGTGGCTTAAGTAG

MUHAMMAD IMRAN 68

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

{2, 2, 2, 3, 3, 4, 5}
If X = {x1 = 0, x2, . . . , xn}
∆X = {xj − xi : 1≤ i < j ≤n}
X={0, 2, 4, 7, 10}, then ∆X={2, 2, 3, 3, 4, 5, 6, 7, 8, 10},
Representation of ∆X

0 2 4 7 10

0 2 4 7 10

2 2 5 8

4 3 6

7 3

10

The element at (i, j) in the table is the value xj − xi for 1 ≤ i ˂ j ≤ n.

Topic # 34 Restriction Mapping

{2, 2, 2, 3, 3, 4, 5}
If X = {x1 = 0, x2, . . . , xn}
∆X = {xj − xi : 1≤ i < j ≤n}
X={0, 2, 4, 7, 10}, then ∆X={2, 2, 3, 3, 4, 5, 6, 7, 8, 10},

Representation of ∆X

MUHAMMAD IMRAN 69

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

0 2 4 7 10

0 4 7 10
2

2 2 5 8

4 3 6

7 3

10

The element at (i, j) in the table is the value xj − xi for 1 ≤ i ˂ j ≤ n.

Topic # 35 Partial Digest Problem

Partial Digest Problem:
Given all pairwise distances between points on a line, reconstruct the positions of those

points. Input: The multiset of pairwise distances L, containing (𝑛) integers.

2

Output: A set X, of n integers, such that ∆X = L

∆A is equal to ∆(A ⊕
{v}), where A⊕ {v} is
defined to be
{a + v : a ϵ A},

Also ∆A = ∆(−A),where−A = {−a : a ϵ A}

A = {0, 2, 4, 7, 10}, ∆(A ⊕ {100}) =

{100, 102, 104,

107, 110}, and −A

= {−10,−7,−4,−2,

0}

{0, 1, 3, 8, 9, 11, 12, 13, 15} and {0, 1, 3, 4, 5, 7, 12, 13, 15}

0 1 3 4 5 7 12 13 15

0 1 3 4 5 7 12 13 15

1 2 3 4 6 11 12 14

MUHAMMAD IMRAN 70

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

0 1 3 8 9 11 12 13 15
0 1 3 8 9 11 12 13 15
1 2 7 8 10 11 12 14
3 5 6 8 9 10 12
8 1 3 4 5 7
9 2 3 4 6
11 1 2 4
12 1 3
13 2
15
3 1 2 4 9 10 12
4 1 3 8 9 11

5 2 7 8 10
7 5 6 8
12 1 3

13 2
15

{14, 24, 34, 43, 52, 62, 72, 83, 92, 102, 112, 123, 13, 14, 15}
U ⊕ V = {u + v : u ϵ U, v ϵ V }

U ⊝ V = {u − v : u ϵ U, v ϵ V }

U = {6, 7, 9} and V = {−6, 2, 6} U V -6 2 6

U⊕V -6 2 6
6 12 4 0

6 0 8 12

7 13 5 1
7 1 9 13

9 15 7 3
9 3 11 15

BRUTEFORCEPDP(L, n)
1. M maximum element in L

2. for every set of n − 2 integers 0 < x2 < · · · < xn−1 < M

3. X {0, x2, . . . , xn−1, M}

4. Form ∆X from X
5. if X = L
6. return X

7. output “No Solution”

MUHAMMAD IMRAN 71

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

ANOTHERBRUTEFORCEPDP(L, n)
1. M maximum element in L
2. for every set of n−2 integers 0 < x2< · · · < xn−1< M from L
3. X {0, x2, . . . , xn−1, M}
4. Form ∆X from X
5. if X = L
6. return X
7. output “No Solution”

Topic # 36 Practical Restriction Mapping Algorithm

Brute Force Algorithm

Largest distance in “L”

Outermost points of “X”

Remaining distance “δ”

L = {2, 2, 3, 3, 4, 5, 6, 7, 8, 10}

Size of L is (𝑛2)= 𝑛 (𝑛 −1)

2 = 10 where is “n” is number of points in the solutions.
Here “n” is 5 and positions of ‘X’ as x1 = 0, x2, x3, x4 and x5.

MUHAMMAD IMRAN 72

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

X = {0, 2, 10} L = {2, 3, 3, 4,

5, 6, 7} x4 = 7 or x3 = 3
If x3 = 3 then x3- x2 = 1 must be in L, but not, so x4 must be 7, x5- x4 = 3, x4 - x2 =
5, and x4 - x1 = 7 from L
X = {0, 2, 7, 10} L = {2, 3, 4, 6,}

Topic # 37 Partial Digest Algorithm

Brute Force Algorithm
List of pairwise distances, L, and uses the function DELETE(y, L)
Value y from L, notation (y,X) to denote the multiset of distances
For example,
∆(2,{1, 3, 4, 5}) ={1, 1, 2, 3}
PARTIALDIGEST (L)
1. width Maximum element in L
2. DELETE (width, L)
3. X {0, width}
4. PLACE (L,X)
PLACE (L,X)
1. if L is empty
2. output X
3. return
4. y Maximum element in L
5. if ∆ (y,X) L
6. 6 Add y to X and remove lengths ∆ (y,X) from L
7. PLACE (L,X)
8. Remove y from X and add lengths (y,X) to L
9. if (width − y,X) L
10. Add width − y to X and remove lengths (width − y,X) from L
11. PLACE (L,X)
12. Remove width − y from X and add lengths (width − y,X) to L
13. return

Topic # 38 Partial Digest Algorithm

PARTIALDIGEST (L)

MUHAMMAD IMRAN 73

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

X = {0, 10}

MUHAMMAD IMRAN 74

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 75

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 39 Regulatory Motifs in DNA Sequences

Sequence Motifs
Biological function
Nucleases and transcription factors
Processes at RNA level

Specific sequence located upstream of genes TCGGGGATTTCC

NF-kB binding sites (nuclear factor kappa-light-chain-enhancer of activated B
cells)
Regulatory motifs, transcription factors
Set of upstream regions in genes in the genome, each region containing one NF-kB
sites Suppose we do not know either location or sequence of NF-kB sites

“The Gold Bug” by Edgar Allan provided some clue of finding DNA motifs, one
of the character find parchment written below

53++!305))6*;4826)4+.)4+);806*;48!8‘60))85;]8*:+*8!83(88)5*!;46(;88*96*?;8)
*+(;485);5 *!2:*+(;4956*2(5*-)8‘8*;4069285) ;)6!8)
4++;1(+9;48081;8:8+1;48!85;4)485!
528806*81(+9;48;(88;4(+?34;48)4+;161;:188;+?;

“;48” codes for “THE”

53++!305))6*THE26)H+.)H+)TE06*THE!E‘60))E5T]E*:+*E!E3(EE)5*!TH6(T
EE*96*?TE
)* +(THE5) T5*!2:*+(TH956*2(5*-H)E‘E*T
H0692E5)T)6!E)H++T1(+9THE0E1TE:E+1THE!E5TH)HE5!52EE06*E1(+9TH
ET(EETH( +?3HTHE) H+T161T:1EET+?T

“;” for “T”

“4” for “H”

“8” for “E”

Topic # 40 Profiles 1

MUHAMMAD IMRAN 76

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Conserved Pattern
32 nucleotide
7 sequences
1. CGGGGCTGGGTCGTCACATTCCCCTTTCGATA
2. TTTGAGGGTGCCCAATAACCAAAGCGGACAAA
3. GGGATGCCGTTTGACGACCTAAATCAACGGCC
4. AAGGCCAGGAGCGCCTTTGCTGGTTCTACCTG
5. AATTTTCTAAAAAGATTATAATGTCGGTCCTC
6. CTGCTGTACAACTGAGATCATGCTGCTTCAAC
7. TACATGATCTTTTGTGGATGAGGGAATGATGC

Figure 1
P = ATGCAACT
l=8
1. CGGGGCTATGCAACTGGGTCGTCACATTCCCCTTTCGATA
2. TTTGAGGGTGCCCAATAAATGCAACTCCAAAGCGGACAAA
3. GGATGCAACTGATGCCGTTTGACGACCTAAATCAACGGCC
4. AAGGATGCAACTCCAGGAGCGCCTTTGCTGGTTCTACCTG
5. AATTTTCTAAAAAGATTATAATGTCGGTCCATGCAACTTC
6. CTGCTGTACAACTGAGATCATGCTGCATGCAACTTTCAAC
7. TACATGATCTTTTGATGCAACTTGGATGAGGGAATGATGC

Figure 2
P = ATGCAACT
l=8
1. CGGGGCTATGCAACTGGGTCGTCACATTCCCCTTTCGATA
2. TTTGAGGGTGCCCAATAAATGCAACTCCAAAGCGGACAAA
3. GGATGCAACTGATGCCGTTTGACGACCTAAATCAACGGCC
4. AAGGATGCAACTCCAGGAGCGCCTTTGCTGGTTCTACCTG
5. AATTTTCTAAAAAGATTATAATGTCGGTCCATGCAACTTC
6. CTGCTGTACAACTGAGATCATGCTGCATGCAACTTTCAAC
7. TACATGATCTTTTGATGCAACTTGGATGAGGGAATGATGC

Figure 3
P = ATGCAACT
l=8
7 x (32 + 8) = 280 nucleotides
Probability = 280/48 = 0.004
1. CGGGGCTATcCAgCTGGGTCGTCACATTCCCCTTTCGATA
2. TTTGAGGGTGCCCAATAAggGCAACTCCAAAGCGGACAAA
3. GGATGgAtCTGATGCCGTTTGACGACCTAAATCAACGGCC

4. AAGGAaGCAACcCCAGGAGCGCCTTTGCTGGTTCTACCTG
5. AATTTTCTAAAAAGATTATAATGTCGGTCCtTGgAACTTC
6. CTGCTGTACAACTGAGATCATGCTGCATGCcAtTTTCAAC
7. TACATGATCTTTTGATGgcACTTGGATGAGGGAATGATGC

Figure 4

MUHAMMAD IMRAN 77

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 41 Profiles 2
Conserved Pattern

1- position 8 - Sequence 1
2- position 19 - Sequence 2
3- position 3- Sequence 3
4- position 5- Sequence 4
5- position 31- Sequence 5
6- position 27- Sequence 6
7- position 15- Sequence 7

1-
CGGGGCTATcCAgCTGGGTCGTCACATTCCCCTT
2-
TTTGAGGGTGCCCAATAAggGCAACTCCAAAGCGGACAAA
3-
GGATGgAtCTGATGCCGTTTGACGACCTA
4-
AAGGAaGCAACcCCAGGAGCGCCTTTGCTGG
5- AATTTTCTAAAAAGATTATAATGTCGGTCCtTGgAAC
TTC
6- CTGCTGTACAACTGAGATCATGCTGCATGCcAtTTTC
AAC

MUHAMMAD IMRAN 78

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

7-
TACATGATCTTTTGATGgcACTTGGATGAGGGAATGATGC
Figure 6

Topic # 42 Profiles 3
Conserved Pattern

If s = (8, 19, 3, 5, 31, 27, 15) then

Score(s) = 5 + 5 + 6 + 4 + 5 + 5 + 6 + 6
= 42 Set of t DNA sequences

“n” nucleotides one position in each of

these “t” sequences s = (s1, s2, . . . , st),
with 1 ≤ si ≤ n − l + 1 l-mers can be
compiled into t × l alignment matrix
whose (i, j)th element is the nucleotide in the si + j − 1th element in
the ith sequence of figure 7

Topic # 43 Motif Finding Problem

Profile Matrix
If P(s) denotes profile matrix, starting from s
MP(s)(j)-largest count in column j of P(s)

MUHAMMAD IMRAN 79

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

For the starting positions in above figure, Score(s,DNA) =

5 + 5 + 6 + 4 + 5 + 5 + 6 + 6 = 42. Score(s,DNA) can be used to
measure
The strength of a profile corresponding to the starting positions s.
consensus score of l · t corresponds to the best possible alignment
Score(s,DNA) consensus score is defined to be Score(s,DNA) =
∑𝑙𝑗=1𝑀P(s) (j) Score(s,DNA) = 5 + 5 + 6 + 4 + 5 + 5 + 6 + 6 = 42
consensus score of l · t corresponds to the best possible alignment
A consensus score of l.t/4 , is worst possible alignment

Motif Finding Problem:

Given a set of DNA sequences, find a set of l-mers, one from each
sequence, that maximizes the consensus score.

Input: A t×n matrix of DNA, and, the length of the pattern to

find. Output: An array of t starting positions s =l (s1, s2, . . . ,
st) maximizing Score(s,DNA).

Topic # 44 Motif Finding Problem 1

Median String
Median string
Given two l-mers v and w, can compute the Hamming distance
between them, dH(v,w), as the number of positions that differ in the
two strings
ATTGTC
: x : x : :
ACTCTC
s = (s1, s1, . . . , st)
v is some l-mer dH(v, s) to denote the total Hamming distance
between v and the l-mers starting at positions s: dH
where dH(v, si) is the Hamming distance
between v and the l-mer that starts at si in the ith DNA sequence
TotalDistance(v,DNA) = mins(dH(v, s)
Finding Total Distance(v,DNA) is a simple problem:
find the best match for v in the first DNA sequence (i.e., a position
minimizing dH(v, s1) for 1 ≤ s1 ≤ n-l+1), then the best match in the
second sequence and so on

MUHAMMAD IMRAN 80

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Median string for DNA as the string v that minimizes

TotalDistance(v,DNA); this minimization is performed over all 4l
strings v of length l.
Median String Problem:
Given a set of DNA sequences, find a median
string. Input: A t × n matrix DNA, and l,
the length of the pattern to find
Output: A string v of l nucleotides that
minimizes TotalDistance(v,DNA)
over all strings of that length
Double minimization: finding a string v that minimizes TotalDistance(v,DNA),
which is in turn the smallest distance among all choices of starting points s in the
DNA sequences.

Min min
dH(v, s) all choices of
all choices l-mers v
starting positions s
The Median String problem- Minimization problem
The Motif Finding problem- Maximization problem
Computationally equal
Let s be a set of starting positions with consensus score Score(s,DNA), and let w be
the consensus string of the corresponding profile. Then dH(w, s) = lt - Score(s,DNA)

Hamming distance between the consensus

string w and each of the seven implanted patterns is 2, and dH(w, s) =2 x 7 = 14

= 7 x 8 − 42 = 14

Topic # 45 Motif Finding Problem 2

Median String
Median string
Given two l-mers v and w, can compute the Hamming distance between them,
dH(v,w), as the number of positions that differ in the two strings

MUHAMMAD IMRAN 81

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

ATTGTC
: x : x : :
ACTCTC
s = (s1, s1, . . . , st) v is some l-mer dH(v, s) to denote the total Hamming distance
between v and the l-mers starting at positions s: dH where dH(v,
si) is the Hamming distance between v and the l-mer that starts at si in the ith DNA
sequence
TotalDistance(v,DNA) = mins(dH(v, s)
Finding Total Distance(v,DNA) is a simple problem:
find the best match for v in the first DNA sequence (i.e., a position minimizing dH(v,
s1) for 1
≤ s1 ≤ n-l+1), then the best match in the second sequence and so on
Median string for DNA as the string v that minimizes TotalDistance(v,DNA); this
minimization is performed over all 4l strings v of length l.
Median String Problem:
Given a set of DNA sequences, find a median
string. Input: A t × n matrix DNA, and l,
the length of the pattern to find
Output: A string v of l nucleotides that
minimizes TotalDistance(v,DNA)
over all strings of that length
Double minimization: finding a string v that minimizes TotalDistance(v,DNA),
which is in turn the smallest distance among all choices of starting points s in the
DNA sequences.
Min min dH(v, s)
all choices of all
choices l-mers v
starting positions s
The Median String problem- Minimization problem
The Motif Finding problem- Maximization problem
Computationally equal
Let s be a set of starting positions with consensus score Score(s,DNA), and let w be
the consensus string of the corresponding profile. Then dH(w, s) = lt - Score(s,DNA)

Hamming distance between the consensus

MUHAMMAD IMRAN 82

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

string w and each of the seven implanted patterns is 2, and dH(w, s) =2 x 7 = 14

= 7 x 8 − 42 = 14

Topic # 45 Search Trees-Introduction

Introduction

Topic # 46 Search Trees Best Alternative

Median String and
Motif Finding Problems
Number of alternatives
To find best one
(n − l + 1)t

Figure 1

MUHAMMAD IMRAN 83

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

AA· · · AA
AA· · · AT
AA· · · AG
AA· · · AC
AA· · · TA
AA· · · TT
AA· · · TG
AA· · · TC
...
CC· · · GG
CC· · · GC
CC· · · CA
CC· · · CT
CC· · · CG
CC· · · CC
All 4l

Figure 2

(1, 1, . . . , 1, 1)
(1, 1, . . . , 1, 2)
(1, 1, . . . , 1, 3)
(1, 1, . . . , 1, 4)
(1, 1, . . . , 2, 1)
(1, 1, . . . , 2, 2)
(1, 1, . . . , 2, 3)

MUHAMMAD IMRAN 84

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

(1, 1, .
. . , 2,
4) ...
(4, 4, . . . , 3, 3)
(4, 4, . . . , 3, 4)
(4, 4, . . . , 4, 1)
(4, 4, . . . , 4, 2)
(4, 4, . . . , 4, 3)
(4, 4, . . . , 4, 4)

1 for A
2 for T
3 for G
4 for C

Consider all kL L-mers in a k-letter alphabet

For Motif Finding problem, k =
n−l+1, For Median String
problem, k = 4.
All 24 4-mers in the two-letter alphabet of 1 and 2.

Topic # 47 Algorithm for Search Trees 1

Search tree Algorithms
Next leave
All leaves
Preorder

L-mer a = (a1 a2…aL)

MUHAMMAD IMRAN 85

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 86

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 48 Algorithm for Search Trees 2

PREORDER(v)
1. output v
2. if v has children
3. PREORDER( left child of v )
4. PREORDER( right child of v )

L
=
4
k
=
2

MUHAMMAD IMRAN 87

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

1. (-,-,-,-)
2. (1,-,-,-)
3. (1,1,-,-)
4. (1,1,1,-)
5. (1,1,1,1)
6. (1,1,1,2)
7. (1,1,2,-)
8. (1,1,2,1)
9. (1,1,2,2)
10. (1,2,-,-)
11. (1,2,1,-)
12. (1,2,1,1)
13. (1,2,1,2)
14. (1,2,2,-)
15. (1,2,2,1)
16. (1,2,2,2)
17. (2,-,-,-)
18. (2,1,-,-)
19. (2,1,1,-)
20. (2,1,1,1)
21. (2,1,1,2)
22. (2,1,2,-)
23. (2,1,2,1)
24. (2,1,2,2)
25. (2,2,-,-)
26. (2,2,1,-)
27. (2,2,1,1)
28. (2,2,1,2)
29. (2,2,2,-)
30. (2,2,2,1)
31. (2,2,2,2)

Topic # 49 Next Vertex Algorithm

Traversing vertices

MUHAMMAD IMRAN 88

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Input: a = (a1, . . . , aL) at

level i Output: Next vertex in
the tree values a1, . . . , ai and
ignores ai+1, . . . , aL
NEXTVERTEX takes inputs that are similar to NEXTLEAF, with the exception
that the
“current leaf” is now the “current vertex,” and uses

parameter i for vertices

When i < L, NEXTVERTEX (a, i, L, k) moves down to the next lower level and
explores that subtree of a. If i = L, NEXTVERTEX either moves along the lowest
level as long as aL < k or jumps back up in the tree.

Topic # 50 Advanced Computing Approaches

Algorithm
➢ Some entity needs to carry out the steps specified by the algorithm
➢ Humans are generally slow

MUHAMMAD IMRAN 89

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

➢ A computer is less intelligent but can perform simple steps quickly and
reliably
➢ Algorithm must be rephrased in programming language
➢ Pseudocode: language often used to describe algorithm
➢ Complex operations are grouped together into mini-algorithms called
subgroups
➢ Variable is written as x or total
➢ An array of n elements is an ordered collection of n variables a1,
a2,……..an
➢ An algorithm is a pseudocode is denoted by a name, followed by the list
of arguments

Topic #51 ByPass Algorithm

Branch and Bound Algorithm
NEXTVERTEX Algorithm
BYPASS Algorithm
Skip the subtree rooted at
vertex (a, i) Increment ai
(unless ai = k

A tree that has uninteresting subtrees. The numbers next to a leaf represent the
“score” for that L-mer. Scores at internal vertices represent the maximum score in
the subtree rooted at that vertex. To improve the brute force algorithm, we can
“prune” subtrees that do not contain highscoring leaves. For example, since the
score of the very first leaf is 24, it does not make sense to analyze the 4th, 5th, or
6th leaves whose scores are 20, 4, and 5, respectively. Therefore, the subtree
containing these vertices can be ignored.

MUHAMMAD IMRAN 90

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 52 Finding Motifs

Brute force approach
BRUTEFORCEMOTIFSEARCH(DNA, t, n, l)
1. bestScore 0
2. for each (s1, . . . , st) from (1, . . . , 1) to (n − l + 1, . . . , n − l + 1)
3. if Score(s,DNA) > bestScore
4. bestScore Score(s,DNA)
5 bestMotif (s1, s2, . . . , st)
6. return bestMotif n−l+1 choices
for the first index (s1,), then for s2, s3)
number of positions is (n − l + 1)t
Score(s,DNA), which requires O(l) operations-O(lnt).
BRUTEFORCEMOTIFSEARCHAGAIN(DNA, t, n, l)
1. s (1, 1, . . . , 1)
2. bestScore Score(s,DNA)
3. while forever
4. s NEXTLEAF(s, t, n − l + 1)
5. if Score(s,DNA) > bestScore
6. bestScore Score(s,DNA)
7. bestMotif (s1, s2, . . . , st)
8. if s = (1, 1, . . . , 1)
9. return bestMotif

Topic # 53 Simple Motif Search Algorithm

Simple Motif Search Algorithm
SIMPLEMOTIFSEARCH(DNA, t, n, l)

MUHAMMAD IMRAN 91

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

1. s (1, . . . , 1)
2. bestScore 0
3. i 1
4. while i > 0
5. if i < t
6. (s, i) NEXTVERTEX(s, i, t, n − l + 1)
7. else
8. if Score(s,DNA) > bestScore
9. bestScore Score(s,DNA)
10. bestMotif (s1, s2, . . . , st)
11. (s, i) NEXTVERTEX(s, i, t, n − l + 1)
12. return bestMotif
Simple Motif Search Algorithm
Some sets of starting positions can be ruled out
If the first i of t starting positions [i.e., (s1, s2,... , si)]
Sequences i+1, i+2, . . . , t,
s = (s1, s2,... , st), define the partial consensus score, Score(s, i, DNA)- i×l alignment
matrix
Partial consensus score for s1, . . . , si, remaining t−i rows can only improve the
consensus score by (t − i) · l
First i starting positions (s1, . . . , s1) could be at most Score(s, i,DNA)+(t−i) · l

Score(s, i,DNA)+ (t− i) · l

is less than the currently best score,
bestScore t − i sequences in the
sample
Score(s, i,DNA) + (t − i) · l
(n − l + 1)t−i

Topic # 54 Branch and Bound Algorithm

Branch and Bound
Motif Search
BRANCHANDBOUNDMOTIFSEARCH(DNA, t, n, l)
1. s (1, . . . , 1)
2. bestScore 0
3. i 1
4. while i > 0
5. if i < t
6. optimisticScore Score (s, i, DNA)+

MUHAMMAD IMRAN 92

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

(t − i)·l
7. if optimisticScore < bestScore
8. (s, i) BYPASS(s, i, t, n − l + 1)
9. else
10. (s, i) NEXTVERTEX(s, i, t, n − l + 1)
11. else
12. if Score(s,DNA) > bestScore
13. bestScore Score(s)
14. bestMotif (s1, s2, . . . , st)
15. (s, i) NEXTVERTEX(s, i, t, n − l + 1)
16. return bestMotif

Topic # 55 Brute Force Algorithm

Brute Force
Median Search
BRUTEFORCEMEDIANSEARCH(DNA, t, n, l)
1. bestWord AAA· · · AA
2. bestDistance
3. for each l-mer word from AAA...A to TTT...T
4. if TOTALDISTANCE(word,DNA) < bestDistance
5. bestDistance TOTALDISTANCE(word,DNA)
6. bestWord word
7. return bestWord

A search tree for the Median String problem. Each branching point can give
rise to only four children, as opposed to the n−l+1 children in the Motif
Finding problem.
SIMPLEMEDIANSEARCH(DNA, t, n, l)
1. s (1, 1, . . . , 1)
2. bestDistance

MUHAMMAD IMRAN 93

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

3. i 1
4. while i > 0
5. if i < l
6. (s, i) NEXTVERTEX(s, i, l, 4)
7. else
8 word nucleotide string corresponding to (s1, s2, . . . sl)
9. if TOTALDISTANCE(word,DNA) < bestDistance
10. bestDistance TOTALDISTANCE(word,DNA)
11. bestWord word
12. (s, i) NEXTVERTEX(s, i, l, 4)
13. return bestWord
BRANCHANDBOUNDMEDIANSEARCH(DNA, t, n, l)
1. s (1, 1, . . . , 1)
2. bestDistance
3. i 1
4. while i > 0
5. if i < l
6. prefix nucleotide string corresponding to (s1,
s2 , . . . , si )
7. optimisticDistance TOTALDISTANCE(prefix,DNA)
8. if optimisticDistance > bestDistance
9. (s, i) BYPASS(s, i, l, 4)
10. else
11. (s, i) NEXTVERTEX(s, i, l, 4)
12. else
13 word nucleotide string corresponding to
(s1, s2, . . . sl)
14. if TOTALDISTANCE(word,DNA) < bestDistance
15. bestDistance TOTALDISTANCE(word,DNA)
16. bestWord word
17. (s, i) NEXTVERTEX(s, i, l, 4)
18. return bestWord

Topic # 56 Genomic Rearrangements

Waardenburg’s syndrome
➢ Hearing loss
➢ Two Different colored eyes
➢ Gene present on chromosome 2

MUHAMMAD IMRAN 94

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

➢ Splotch gene in mice

➢ Human genome-mouse genome
➢ Cut into 300 genomic fragments-Synteny blocks
➢ Chromosome 2 in humans-mouse chromosomes 1, 2, 3, 5, 6, 7, 10, 11, 12,
14, and 17
➢ Genome rearrangement results in a change of gene ordering
➢ Analysis of human and mouse genomes-250 genomic rearrangements

Mouse X chromosome

Human X chromosome
Transformation of the mouse gene order into the human gene order on the X
chromosome

Topic # 57 Greedy Approach for Motif Search

Approximation algorithm
Brute force algorithm to solve the Motif Finding
problem running time of O(l · nt)
Cannot run it on
biological samples
Faster greedy
technique-not correct,
good performance
Approximation
algorithm
CONSENSUS- as good
or better

GREEDYMOTIFSEARCH scans each DNA sequence only once. Once we have

scanned a particular sequence i, we decide which of its l-mer has the
best contribution to the partial alignment score Score(s, i, DNA) for the first i
sequences and immediately claim that this l-mer is part of the alignment.
GREEDYMOTIFSEARCH (DNA, t, n, l)
1. bestMotif (1, 1, . . . , 1)
2. s (1, 1, . . . , 1)
3 for s1 1 to n − l + 1
4 for s2 1 to n − l + 1
5. if Score(s, 2,DNA) > Score(bestMotif ,
2,DNA)

MUHAMMAD IMRAN 95

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

6. BestMotif1 s1
7. BestMotif2 s2
8. s1 BestMotif1
9. s2 BestMotif2
10. for i 3 to t
11 for si 1 to n − l + 1
12. if Score(s, i, DNA) > Score (bestMotif , i, DNA)
13. BestMotifi si
14. si bestMotifi
15. return bestMotif
Approximation algorithm
Two
closest l-
mers 2 × l
seed
matrix l(n
− l + 1)2
operations

t − 2 iterations-by scanning the ith sequence (for 3≤ i ≤ t)

t−2 sequences and selecting the one l-mer that has
themaximumScore(s, i) l · (n − l + 1) operations
Approximation algorithm
Running time of this algorithm is O(ln2 +lnt), which is vastly better than the O(lnt)
of
IMPLEMOTIFSEARCH or even the O(4lnt) of BRUTEFORCEMEDIAN-
STRING

Topic # 58 The Power of DNA Sequence Comparison

Introduction
Discovery of new gene-no idea of functions
Find similarities with genes of known function
Newly discovered cancer-causing -sis oncogene matched a normal gene involved
in growth and development called platelet-derived growth factor
Oncogene v-sis is the simian sarcoma virus
Scientists became suspicious that cancer might be caused by a normal growth gene
Discovery of cystic fibrosis gene

MUHAMMAD IMRAN 96

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Abnormal secretions, and is diagnosed in

children produce abnormally thick mucus that
clogs the lungs
10 million Americans are carriers of the cystic fibrosis gene
There is a 25% chance that the child will have cystic fibrosis
In 1989, 1 million nucleotides on the chromosome 7
The area around the cystic fibrosis gene was sequenced
Database of all known genes
Similarities between a gene that had already been discovered, code for adenosine
triphosphate (ATP) binding proteins
These proteins constitute the ion transport channel
The disease involves sweat secretions with abnormally high sodium content
Link between cancer-causing genes and normal growth genes and elucidating the
nature of cystic fibrosis

Topic # 59 Brute Force vs Greedy Algorithm

Dynamic Programming
Manhattan Tourist Problem:
Find a longest path in a
weighted grid Input: A
weighted grid G with
two distinguished
vertices: a source and
a sink
Output: A longest
path in G from source
to sink

Tourists only move south and east, any grid

positions west or north of the source are
unusable
Any grid positions south or east of the sink are unusable, the source vertex is at
(0, 0) Sink vertex at (n, m) defines the southeastern most corner of the grid

MUHAMMAD IMRAN 97

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

The brute force approach search for the longest path

Not an option for medium to large grid
Use a greedy strategy, choose between two possible directions (south or east) by
comparing them and selecting one with large increase for one step (local maximum)
Good at beginning
May lead to area with few attractions
No known greedy strategy for the Manhattan Tourist problem provides an optimal
solution to the problem

Instead of solving the Manhattan Tourist problem directly, that is, finding the
longest path from source (0, 0) to sink (n,m), we solve a more general problem:
find the longest path from source to an arbitrary vertex (i, j) with
0 ≤ i ≤ n, 0 ≤ j ≤ m. We will denote the length of such a best path as si,j , noticing
that sn,m is the weight of the path that represents the solution to the Manhattan
Tourist problem

MUHAMMAD IMRAN 98

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic #60 Sequence Similarity

Meaning of “sequence similarity or “distance” between DNA sequences. Hamming
distance is not typically used to compare DNA or protein sequences. Calculation
rigidly assumes that the ith symbol of one sequence is already aligned against the
ith symbol of the other

It is common case that the ith symbol in one sequence corresponds to a symbol at
different position in other. Mutation in DNA-evolutionary process: DNA
replication- substitutions, insertions, and deletions of nucleotides, leads to “edited”
DNA texts. Whether the ith symbol in one DNA sequence corresponds to the ith
symbol in the other

Topic # 61 Edit Distance

Edit distance between two strings as the minimum number of editing operations
needed to transform one string into another, where the edit operations are insertion,
deletion, and substitution of one symbol for another
It is often the case that the ith symbol in one sequence corresponds to a symbol at
different position in other. Mutation in DNA-evolutionary process: DNA
replication- substitutions, insertions, and deletions of nucleotides, leads to “edited”
DNA texts. Whether the ith symbol in one DNA sequence corresponds to the ith
symbol in the other

MUHAMMAD IMRAN 99

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 62 Alignment
The alignment of the strings v (of n characters) and w (of m characters, with m not
necessarily the same as n) is a two-row matrix such that the first row contains the
characters of v in order while the second row contains the characters of w in
order, where spaces may be interspersed throughout the strings in different places
As a result, the characters in each string appear in order, though not necessarily
adjacently.
No column of the alignment matrix contains spaces in both rows, so that the
alignment may have at most n + m columns.

MUHAMMAD IMRAN 100

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

A T -- G T T A T --

A T C G T -- A -- C

Columns that contain the same letter in both rows are called matches, while
columns containing different letters are called mismatches. The columns of the
alignment containing one space are called indels, with the columns containing a
space in the top row called insertions and the columns with a space in the bottom
row deletions. Five matches, zero mismatches, and four indels. The number of
matches plus the number of mismatches plus the number of indels is equal to the
length of the alignment matrix and must be smaller than n + m

A T -- G T T A T --

A T C G T -- A -- C

Each of the two rows in the alignment matrix is represented as a string interspersed
by space symbols “−”; for example AT--GTTAT-- is a representation of the row
corresponding to v = ATGTTAT, while ATCGT--A--C is a representation of the
row corresponding to w = ATCGTAC

Topic # 63 Edit Graph 1

The grid that is achieved after alignment is similar to the Manhattan grid where
each entry in the grid looks like a city block
The graph called Edit Graph
The main difference between the Manhattan and Edit Graph is that we can move
along the diagonals in the Edit Graph

MUHAMMAD IMRAN 101

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 64 Edit Graph 2

Analyzing the merit of an alignment-corresponding path in the edit graph. For any
two strings- different alignment matrices and corresponding paths.
Surplus of mismatches and indels and a small number of matches, while others have
many matches and few indels and mismatches.

MUHAMMAD IMRAN 102

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Relative merits of one alignment over another-scoring function- input an alignment

matrix and produces a score that determines the “goodness” of the alignment.
Variety of scoring functions-higher scores to alignments with more matches.
Column as a positive number if both letters are the same, and as a negative number
if the two letters are different. The score for the whole alignment is the sum of the
individual column scores.

Topic # 65 Longest Common Subsequences

The simplest form of a sequence similarity analysis is the Longest Common
Subsequence (LCS) problem, where we eliminate the operation of substitution and
allow only insertions and deletions. A subsequence of a string v is simply an
(ordered) sequence of characters (not necessarily consecutive) from v.
v = ATTGCTA AGCA and ATTA are subsequences
TGTT and TCG are not subsequences
A common subsequence of two strings is a subsequence of both of them. Common
subsequence of strings
v = v1 . . . vn and w =
w1 . . .wm as a
sequence of positions
in v,

1 ≤ i1 < i2 < · · · < ik

≤ n and a sequence of
positions in w,
1 ≤ j1 < j2 < · · · < jk ≤ m such that the symbols at the
corresponding positions in v and w coincide:

vit = wjt for 1 ≤ t ≤ k

TCTA is a common to both ATCTGAT and TGCATA

Typically many common subsequences between

two strings v and w-how to find the longest one
s(v,w)-be the length of the longest common subsequence of v and w
edit distance between v and w—under the assumption that only insertions and
deletions are allowed—is d(v,w) = n + m − 2s(v,w) and corresponds to the
minimum number of insertions and deletions needed to transform v into w

MUHAMMAD IMRAN 103

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Longest Common Subsequence Problem:

Find the longest subsequence common to two strings
Input: Two strings, v and w
Output: The longest common subsequence of v and w

Topic # 66 Recurrence for LCS problem

Every common subsequence corresponds to an alignment with no mismatches.
This can be obtained simply by removing all diagonal edges from the edit graph
whose characters do not match, thus transforming it into a graph like that shown in
the figure

MUHAMMAD IMRAN 104

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

The relationship between

the Manhattan Tourist problem and the LCS Problem is further illustrate by
showing that these two problems lead to very similar recurrences
Define si,j to be the length of an LCS between v1 . . . vi, the i-prefix of v and w1 . .
.wj , the jprefix of w. Clearly, si,0 = s0,j = 0 for all 1 ≤ i ≤ n and 1 ≤ j ≤ m si,j satisfies
the following recurrence

si−1,j

si,j = max si,j−1

si−1,j−1 + 1, if vi = wj
The first term- when vi is not present in the LCS
of the i-prefix of v and j-prefix of w (deletion of vi); the second term- when wj is
not present in this LCS (an insertion of wi ); and the third term-when both vi and
wj are present in the LCS (vi matches wj).
These recurrences can be rewritten by adding some zeros here and there as

si−1, j + 0

si,j = max si, j−1 + 0

si−1,j−1 + 1, if vi = wj

‘s’ is used to represent dynamic programming table, the data structure

MUHAMMAD IMRAN 105

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

The length of an LCS between v and w can be read from the element (n,m) of the
dynamic programming table, but to reconstruct the LCS from the dynamic
programming table, one must keep some additional information about which of the
three quantities, si−1,j , si,j−1, or si−1,j−1 + 1, corresponds to the maximum in the
recurrence for si,j .

Topic #67 Algorithms for LCS

The length of an LCS
Some additional information about which of the three quantities , si−1,j , si,j−1, or
si−1,j−1 + 1
The following algorithm achieves this goal by introducing backtracking pointers
that take one of the three values , or .

MUHAMMAD IMRAN 106

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Dynamic programming table-computation of the similarity score s(v,w) between v

and w, while the table on the right-computation of the edit distance between v and
w-insertions and deletions are the only allowed operations. The edit distance d(v,w)
is computed according to the initial conditions di,0 = i, d0,j = j for all 1 ≤ i ≤ n and 1
j m and the following recurrence:
The edit distance d(v,w) is computed according to the initial conditions di,0 = i, d0,j
= j for all 1 ≤ i ≤ n and 1 j m and the following recurrence:

di−1,j + 1
di,j = min di,j−1 + 1

di−1,j−1, if vi = wj

Topic #68 Scoring Alignments 1

Scoring matrices for DNA sequence- μ and σ
Scoring matrices for protein sequences are complicated – pointed accepted
mutation (PAM) and block substitution (BLOSUM), frequency amino acid ‘x’
replaces amino acid ‘y’ in evolutionary related sequences.
Random mutagenesis-change amino acid sequence
Some mutations- do not alter but other do
Some amino acid substitutions are commonly found through the process of
molecular evolution-

MUHAMMAD IMRAN 107

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Types of changes- most and least common- amino acid scoring matrix, sequence
alignment
Amino acids sequences- very few matches-scoring matrix δ(i, j) – how often a. a
‘i’ substitutes a. a ‘j’

Topic # 69 Scoring Alignments 2

Large set of alignments of related sequences
Computing δ(i,j)

Amounts to counting how many times the amino acid ‘i’ is aligned with amino acid
‘j’
Needs to know scoring matrix
Met Ala Phe Ser Gly Asp Glu Ser. . . . . . .
Met Ala Phe Ser -- Asp Glu Ser. . . . . . .
If proteins are 90% identical, premium +1 for matches and -1 for mismatches and
indels will do the job. Then “obvious” alignments are constructed that are used to
compute scoring matrix δ.
The simplified description hides subtle details are important in the construction of
scoring matrix
Ser Phe Try Phe
(related proteins in mouse and rat) LESS (related proteins in mouse and
human)
15 million years 80 million years
The best scoring matrix to compare two proteins depends on similarity of these
organisms

Topic #70 PAM matrix

Problem is rectified- analyzing similar proteins e.g, one mutation/100 a.a. Proteins-
human and chimpanzee, Such sequences, one PAM unit diverged. PAM unit as the
amount of time in which an “average” protein mutates 1% of it’s a.a. PAM1 scoring
matrix- many alignments of extremely similar proteins For a given set of base
alignments
define f(i,j) as the total number of times amino acids i and j are aligned against each
other, divided by the total number of aligned positions.

Also define g(i,j) as where f(i) is the frequency of the amino acid i in
all proteins from data set.
g(i,j) defines the probability that an amino acid i mutates into amino acids j within
1 PAM unit. The (i,j) entry of the PAM 1 matrix is defined as δ(i,j) =

log = log (f(i). f(j), the frequency of aligning amino acid

i against amino acid j that one expects simply by chance). The PAM n matrix can
be defined as the result of applying the PAM 1 matrix n times

MUHAMMAD IMRAN 108

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

If g is the 20 x 20 matrix of frequencies g(i,j), then gn (multiplying this matrix by

itself n times) gives the probability that amino acid i mutates into amino acid j
during n PAM units. The (i,j)

entry of the PAM n matrix is defined as log

For large n, the resulting PAM matrices often allow one to find related proteins
even when the practically no matches in the alignment. In this case the underlying
nucleotide sequences are so diverged that their comparison usually fails to find any
statistically significant similarities.
Similarity between the
Cancer –causing v-sis oncogene
Growth factor PDGF (Platelet derived growth factor)
Remained undetected
Russell Doolittle and colleagues has not transformed the nucleotide sequences into
amino acids sequences prior to performing the comparison.

Topic #71 Local Sequence Alignment

Global Alignment- similarity 2 strings- extends over entire length, eg, same
protein family- very conserved, same length – fruit flies to humans Score of
alignment- 2 substrings ‘v’, ‘w’- larger than score of alignment between entire
length of ‘v’ and ‘w’.
Homeobox genes, regulate embryonic development- variety of species. Different in
different species- one region- homeodomain- highly conserved. How to find
conserved area- ignore areas-little similarity. Temple Smith and Micheal Waterman
proposed modification of the global sequence alignment dynamic programming
algorithm-LAP

MUHAMMAD IMRAN 109

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

The score of this path is

The score of this path is

This path contains so many indels that it is unlikely to be the highest scoring
alignment.
Biologically irrelevant diagonal paths – likely have score- mismatches are

Topic # 72

same TOPIC 71

Topic #73 Local Alignment Problem

Biological significant in certain parts of DNA fragments, maximize the alignment
score s (vi
…viˊ ,wj…wjˊ) over all substrings vi….viˊ of v and wj…..wjˊ of w. Local Alignment
Problem- not extend over entire length as in Global Alignment Problem

Local Alignment Problem:

Find the best local alignment between two strings
Input: Strings v and w and a scoring matrix δ
Output: Substring of v and w whose global alignment, as defined by δ, is maximal
amoung all global alignment of all substrings of v and w.

MUHAMMAD IMRAN 110

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Global Sequence Alignment- finding the longest path between vertices (0,0) and
(n,m) in the edit graph
Local Alignment-finding the longest path among the paths between arbitary
vertices (i, j) and (iˊ, jˊ) in the edit graph.

Find the longest path between every pair of vertices (i, j) and (iˊ, jˊ)- then select
longest of these computed paths. Instead of finding the longest path from (i, j) to
(iˊ, jˊ), LAP- finding the longest path from the source (0,0) to every other vertex by
adding edges to weight 0

Topic #74 Progressive Multiple Alignment

Another approach-strong pairwise alignment Greedy progressive
multiple alignment heuristic-pair of strings with greatest similarity-new string-
“once a gap, always a gap.” Multiple alignment of k sequences is reduced to the
multiple alignment of k−1 sequences.

MUHAMMAD IMRAN 111

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

The motivation for the choice of the closest strings at the early steps of the algorithm
is that close strings often provide the most reliable information about a real
alignment
Many popular iterative multiple alignment algorithms including the tool
CLUSTAL, use similar strategies
ATGTCATATTCGGAC
ATGTCATATTCGGAC
ATGTCATATTCGGAC
ATG- CATA
ATGTCATA
Progressive multiple alignment algorithms- problem with CLUSTAL-may be
misled by some spuriously strong pairwise alignment effect, a bad seed. The error
in initial pairwise alignment will propagate all the way through to the whole
multiple alignment. Many algorithms have been proposed,-even with systematic
deficiencies are quite useful in computational biology
Multiple alignment for k sequences
Generalization of the Pairwise Alignment problem Existence of a k-dimensional
scoring matrix k-dimensional scoring matrices are not very practical Describe two
other scoring approaches that are
more biologically relevant. The choice of the scoring function can drastically affect
the quality of the resulting alignment, and no single scoring approach is perfect in
all circumstances.
Multiple alignment of k sequences-a path of edges in a k-dimensional-Manhattan
gridlike edit graph.
The weights of the edges-scoring function
Intuitively, assign higher scores to the columns with a low variation in letters-high
scoreshighly conserved sequences
Multiple Longest Common Subsequence problem, the score of a column is set to 1
if all the characters in the column are the same, and 0 if even one character disagrees

Topic # 75 Gene Prediction 1

Sydney Brenner and Francis Crick
Every triplet codes for one amino acid
Introduce deletions in DNA-dramatically alters its protein product. Deleting three
consecutive nucleotides results in minor changes in the protein
The phrase
THE SLY FOX AND THE SHY DOG
(written in triplets)
Turns into nonsense after deleting one letter Y from SLY
THE SYF OXA NDT HES HYD OG
or two letters LY from SLY

MUHAMMAD IMRAN 112

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

THE SFO XAN DTH ESH YDO G

but makes some sense after deleting three nucleotides SLY
THE SOX AND THE SHY DOG
Charles Yanofsky proved that a gene and its protein product are collinear.
Yanofsky’s experiment was so influential that nobody even questioned about
codons and for almost 2 decades biologists believed that a protein was encoded by
a long string of adjacent triplets.
Discovery of split human genes-collection of substrings
Raised the computational problem-prediction of genes
Human genome is larger and complex than bacterial genome
Salamander genome is ten times larger than the human genome Large
amounts of so-called junk DNA
Human genes - exons that are separated by this junk DNA. The difference in the
sizes of the salamander and human genomes thus presumably reflects larger
amounts of junk DNA and repeats in the salamander genome.
Split genes are analogous to a magazine article- page 1, 13, 43, 51, 74, 80, and 91,
with pages of advertising appearing in between. Junk DNA represents “advertising”
that separates exons.

Topic # 76 Gene Prediction 2

The jump is inconsistent from species to species.
A gene in insect genome- different in worm
Number of exons
The information in one part in human-broken up into two in the mouse.
While the genes themselves are related, they may be quite different in terms of the
parts’ structure
Split genes-1977
Phillip Sharp and Richard Roberts- Adenovirus

“An Amazing Sequence Arrangement at the 5’ End of Adenovirus 2Messenger

RNA.” Sharp’s group- hexon.

Hexon mRNA, mRNA was hybridized to adenovirus DNA- the hybrid molecules-
electron microscopy.
mRNA-DNA hybrids-three loop structures-continuous duplex segment- classic
continuous gene model

MUHAMMAD IMRAN 113

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

An electron microscopy experiment led to the discovery of split genes

Hexon mRNA is built from four separate fragments of the adenovirus genome
These four continuous segments, exons, in the adenovirus genome are separated by
three “junk” fragments called introns.

77. One approach of Gene Prediction

Human genes-3% of the human genome- no in silico gene recognition algorithm
The intronexon model-in eukaryotic organisms Prokaryotic organisms Gene
prediction algorithms for prokaryotes tend to be somewhat simpler than those for
eukaryotes
Two categories-predicting gene location. The statistical approach to gene
prediction Splicing signals- exon-intron junctions
Dinucleotides AG and GT on the left- and right-hand sides of an exon are highly
conserved

Less conserved positions on both sides of the exons

The simplest way to represent such binding sites is by a profile describing the
propensities of different nucleotides to occur at different positions
Using profiles-detect splice sites
Profiles are quite weak-match frequently in the genome at non splice sites.
Attempts to improve the accuracy of gene prediction- second category-based on
similarity.

Topic #78 Second Approach for Gene Prediction

The similarity-based approach to gene prediction relies on the observation that a
newly sequenced gene has a good chance of being related to one that is already
known. For example, 99% of mouse genes have human analogs.
Simply look for a similar sequence-based on the genes known in another-exon
sequence and the exon structure are different
The commonality-produce similar proteins
Similarity-based methods attempt to solve a combinatorial puzzle-putative exons
in a genomic sequence – mouse-human

MUHAMMAD IMRAN 114

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Know a human protein, and we want to discover the exon structure of the related
gene in the mouse genome. The more sequence data we collect, the more accurate
and reliable similarity based
methods become. Consequently, the trend in gene prediction has recently shifted
from statistically motivated approaches to similarity-based algorithms

Topic # 79 Statistical Approach to Gene Prediction 1

Statistical approaches to finding genes-statistical variations between coding (exons)

and noncoding regions
Open reading frames (ORFs) Fenome of length n as a sequence of n/3 codons
The three “stop” codons, (TAA, TAG, and TGA)
The subsegments of these that start from a start codon, ATG, are ORFs.
ORFs within a single genomic sequence may overlap since there are six possible
“reading frames”: three on one strand starting at positions 1, 2, and 3, and three on
the reverse strand

The six reading frames for the DNA sequence

One would expect to find frequent stop codons in noncoding DNA, since the
average number of codons between two consecutive stop codons in “random”
DNA should be 64/3 ~ 21. This is much smaller than the number of codons in an
average protein, which is roughly 300. Therefore, ORFs longer than some threshold
length indicate potential genes. However, gene prediction algorithms based on
selecting significantly long ORFs may fail to detect short genes or genes with short
exons
Many statistical gene prediction algorithms rely on statistical features in protein-
coding regions, such as biases in codon usage. We can enter the frequency of
occurrence of each codon within a given sequence into a 64-element codon usage
array

MUHAMMAD IMRAN 115

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 80 Statistical Approach to Gene Prediction 2

Some more facts about genetic code and codon usage in humans

in-frame hexamer count Mark Borodovsky

Gene prediction in bacterial genomes-several conserved sequence motifs often
found in the regions around the start of transcription
Such sequence motifs are more elusive in eukaryotes
The approaches-prokaryotes-eukaryoyes
Exons-130 nucleotides-reliable peaks in the likelihood ratio plot while analyzing
ORFs-do not differ enough from random fluctuations to be detectable. Moreover,

MUHAMMAD IMRAN 116

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

codon usage and other statistical parameters probably nothing in common in

splicing machinery recognizes exons

Topic # 81 Statistical Approach to Gene Prediction 3

Biologically oriented
Approach
Recognize the locations of splicing signals at exon-intron junctions
There exists a weakly conserved sequence of eight
nucleotides at the boundary of an exon and an intron (donor splice site) and a
sequence of four nucleotides at the boundary of an intron and exon (acceptor splice
site)

Profiles for splice sites are weak-limited success Hidden Markov Model (HMM)
approaches that capture statistical dependencies between sites GENSCAN
developed by Chris Burge and Samuel Karlin. GENSCAN combines coding region
and splicing signal predictions into a single framework.
Splice site prediction-coding region appear on one side of the site
Such statistics are used in the HMM framework of GENSCAN that merges splicing
site statistics, coding region statistics, and motifs near the start of the gene
The accuracy of GENSCAN decreases for genes with many short exons or with
unusual codon usage

Topic # 82 Similarity Based Approached to Gene Prediction 1

A similarity-based approach-Previously sequenced genes Unknown genes in newly

sequenced DNA fragments Combinatorial puzzle-find a set of substrings (candidate
exons) whose splicing best fits the target Brute force approach-find all local
similarities
Each substring from the genomic sequence that exhibits sufficient similarity to the
target protein could be considered a putative exon Exon-flanking
dinucleotides AG and GT

Overlapping
Model a putative exon with a weighted interval in the genomic sequence,
parameters (l, r, w)
l is the left-hand position, r is the right-hand position, and w is the weight of the
putative exon “w”- local alignment score
Likelihood that this interval is an exon
A chain is any set of non overlapping weighted intervals.
Total weight of a chain
A maximum chain

MUHAMMAD IMRAN 117

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Weights of all intervals are positive (w > 0)

Topic # 83 Similarity Based Approached to Gene Prediction 2

Model a putative exon with a weighted interval in the genomic sequence

Five weighted intervals, (2, 3, 3), (4, 8, 6), (9, 10, 1), (11, 15, 7), and (16, 18, 4),
shown by bold edges, form an optimal solution to the Exon Chaining problem. The
array at the bottom shows the values s1, s2, . . . , s2n generated by the
EXONCHANING algorithm

Topic # 84 ORF Prediction

ORF (Open Reading Frame)
Gene finding, especially in prokaryotes starts form searching for open reading frames (ORF)
An ORF is a sequence of DNA that starts with start codon “ATG” (not always) and ends with
any of the three termination codons (TAA, TAG, TGA)

Here, in this figure we see

a comparison between
prokaryotic and
eukaryoticcells.
Prokaryotes maybe
bacteria or some related
organisms whereas rest of
the organisms is classified
in eukaryotes.
If we look into the
prokaryotic situation, we
have different genes which are separated by Intergenic regions and these ORF can be spanning
these two genes whereas in case of the eukaryotes, we have only one gene, we have exon1 and
exon2 whereas the ORF spans cross these different exons.
So on the top (in the figure), we have the longer ORF whereas in the bottom we have a shorter
one as compared to the genome size those shorter ORF which are actually spanning different
exons.
ORF and gene finding:
─ ORF provide important evidence in gene finding.
─ Generally longer ORFs are preferred.
─ However presence of ORF not necessarily means the region is translated to a functional
product

MUHAMMAD IMRAN 118

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Reading Frames:
Depending on the start point, we can define different ORFs, so if we want to go for those
triplet codons, we can start with any nucleotide, in this way we have three different
possibilities for one of the strands and since we have two strands in the DNA, so in total we
can have six ORFs (so 3 of them are from 3’ to 5’ direction whereas the other 3 are from 5’ to
3’ direction).
*Three on forward
strand and three on
complementary strand

This is how those 6

ORFs looks like.
So, there are the six
frame translations (in
this picture).
We observe that we
have a top strand which
is a forward strand and
starts at 5’ end and ends
at 3’ end.

A complementary runs in an anti-parallel fashion which starts with 3’ end and ends at 5’ end.

So, how can we get the reading frames out of it?

We can start with the position number 1 on the first strand and we label it as +1 (as shown), in
this way we can start with A, and have triplet codons like ATG, GTT, TGG and so on.
In the second strand or second possibility (denoted by +2), we can start from the position
number 2, so this frame starts from the second nucleotide like T and makes a triplet codon as
TGG, TTT, GGG and so on.
So this how the third strand (3rd possible ORF; denoted by +3), where it starts from the
position number 3 can be made. We don’t start with the position number 4 because it will be
same as the position number 1.

Similarly, we can do like this for the opposite direction strands with the possibility of position
number 1, 2 and 3.

So, this how the 6 ORF are made.

Conclusions:
An ORF is a sequence of DNA that starts with start codon “ATG” (not always) and ends with
any of the three termination codons (TAA, TAG, TGA) and there are 6 reading registers as far
as the ORFs are considered from the sequence.
Reference:
Biological Sequence Analysis
R Durbin, S Eddy, A Krogh and G Mitchison
Cambridge University Press, 1998.
Bioinformatics The machine learning approach
P Baldi and S Brunak
The MIT Press, 1998
Post-Genome Informatics
M Kanehisa

MUHAMMAD IMRAN 119

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Oxford University Press, 2000

Topic # 85 ORF Finders

ORF Finding:
─ Long ORF may be a gene.
─ Expected 64/3 ~ 21 codons before we see a stop codon.
─ Genes are longer than this.
─ We might scan for ORF longer than a threshold

Codon usage and likelihood ratio:

─ An ORF is more ‘reliable’ if it has ‘likely’ codons
─ We can do sliding window calculations (focus on some nucleotides like for example if
the segment is 1000 long, we can make a window of say size 50 and we can look into
the frequencies of different codons in that window) to ORF having ‘likely’ codon
usage.
─ An ORF is more ‘reliable’ if it has ‘likely’ codons
─ However average vertebrate exon length (130 nucleotides) is too small for reliable
peaks

An improvement may be;

In-Frame hexamer count
i.e. frequencies of pairs of consecutive codons.
ORF Finders:
Tools are mainly based on pattern finding algorithms
1. NCBI’s ORF Finder
2. ORF Investigator
3. OrfPredictor

If you go to NCBI webpage, the ORF

Finder is there.
You can place you sequence into the
text box and upload it while simply
clicking onto the ‘OrfFind’ button.
You can even grasp the sequence
automatically by writing the accession
number in the space give after the
word ‘ACESSION’.

MUHAMMAD IMRAN 120

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

When you click ‘OrfFind’, it

will take you to this page
where there are different
registers or ORFs (explained
in the previous lecture).

The steric (*) here means that

the protein synthesis stops
here.

Conclusions:
─ ORF providesimportant evidence in gene finding.
─ Generally longer ORFs are preferred.
─ However presence of ORF not necessarily means the region is translated to a functional
product

TOPIC 86 Translation Start Site (TSS)

Translation Start Site (TSS)
Translation starts with ATG that codes for methionine in a polypeptide

Here, in this example shown in the figure, we have those

TSS in the middle (it’s actually a file where sequences are
aligned with one another), we look deeply, we can see that
their neighbors are C and G which are most frequently
seen (if not always). We might use these neighbors to
identify the presence of these TSS.
Assumption:
─ Certain nucleotides prefer to be around TSS than others.
─ The “biased” nucleotide distribution is information is a basis for translation start
prediction

Coding Potential:
Hexamer frequencies in coding versus non-coding regions may provide important insights
Frequency of X(A,G,C,T) at position i is
Fi (X)= log(Ci (X)/Ni(X))(frequency of any nucleotide can be found by taking the sum of
thelog of ratio of the counts of theparticular nucleotide in that particular position divided by
the total)

*where c is the counts.

MUHAMMAD IMRAN 121

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Based upon
the
frequency
equation,
we can
come up
with a
frequency
table as
shown in
the figure.
The
frequencies
shown here
indicates
that you have the presence of true TSS here or you can expect that there are some
Transcription Start Sites (TSS) here.
You can observe that on different positions like -4, -3, -2 and -1, similarly at +3, +4,+5, and +6
you have which nucleotides and what are their percentages (in the table).
Example
Which one is more probable to be a Translation Start?

Solution
We can use frequency table and the scoring function as under;
Si =  log (Fi (X)/0.25)
-frequency from the frequency table divided by the expected frequency and then we convert it
into log scale because we want to play with big numbers.
We can call this equation as theInformation Content (IC)

Here, is our
frequency table,
we pick the
frequencies
from here.

MUHAMMAD IMRAN 122

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

And we add those

frequencies here in the
equation.
So, in this way we get a
positive number in the end,
so we have 13.69 on the left
side whereas we have 9.44 on
the right side.

So, which sequence has the strong evidence to have a translation site. In our case, we will
prefer the one with higher value so its probably the green sequence.
Algorithm:
➢ Build a mathematical model, based on collected translation start sequence
➢ For each candidate translation start sequence, apply the model and get a score
➢ If the score is larger than zero, predict it is a “translation start”; the higher score, the
higher the probability the prediction is true

Conclusions:
• TSS prediction can be an important step in gene prediction

• TSS can be predicted while using the frequency of neighboring nucleotides

References:
Biological Sequence Analysis
R Durbin, S Eddy, A Krogh and G Mitchison
Cambridge University Press, 1998.
Bioinformatics The machine learning approach
P Baldi and S Brunak
The MIT Press, 1998
Post-Genome Informatics
M Kanehisa
Oxford University Press, 2000

TOPIC 87 Prediction of splice junctions

Splice Junctions:
Donor site
• Coding region | GT (introns starts)

Acceptor
• (introns ends)YAG | coding region
-Y can be any pyrimidine.

• Canonical form

• GT-AG: 99.24%

There are also some non-canonical forms.

MUHAMMAD IMRAN 123

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Like TSS, the flanks of splice junctions show “biased” distributions of nucleotides in certain
positions
• These biased distributions of nucleotides are the basis for prediction of splice junctions

Here, is the example where we can see we have the exon

then we have the donor site where we see a big GT (this
representation is known as sequence logos), then we have
the intron region which is followed by the acceptor region
which ends up with AG and then again a next exon starts.

Sequence LOGOS:
─ A visual representation of a position-specific distribution
─ Easy for nucleotides, but we need colour to depict up to 20 amino acid proportions.

Overall height at position is proportional to the information content

• Proportions of each nucleotide/amino acid are in relation to their observed frequency,
with most frequent on top, next most frequent below

Non Canonical Splice Junctions:

In addition to canonical GT-AG (99.24%);
GC-AG: 0.69%
AT-AC: 0.05%
Others: 0.02%
Information Content (IC):
Si =  log (Fi (X)/0.25)
─ If every nucleotide has 0.25 frequency in a position, then the position’s information
content is ZERO.
─ Use “information content as a criterion for determining the length of flanks

MUHAMMAD IMRAN 124

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Here, is the acceptor site

distributions and we can
observe that at position
number -2 and -1, there
are the presence of the
acceptor sites.
100% frequencies of A
and G, so these are
predicted but if we look
into the neighbors like at
position -3, there are
mostly Cs.
We can look into the
donor sites; we have
100% frequency for Gs and Ts,
we label them as 1 and 2, and
different positions which are
neighboring to them, we can
include 10 or 15 neighbors.

Algorithm:
Mathematical model: Fi (X): frequency of X (A, C, G, T) in position I
Score a segment as a candidate donor/acceptor site by
log (Fi (X)/0.25)
For each candidate sequence, apply the model and get a score
If the score if larger than zero, predict it is “donor/acceptor”; the higher score, the higher the
probability the prediction is true
Conclusions:
Like TSS, the flanks of splice junctions show “biased” distributions of nucleotides in certain
positions
• These biased distributions of nucleotides can be used for prediction of splice junctions

References:
Biological Sequence Analysis
R Durbin, S Eddy, A Krogh and G Mitchison
Cambridge University Press, 1998.
Bioinformatics The machine learning approach
P Baldi and S Brunak
The MIT Press, 1998
Post-Genome Informatics
M Kanehisa
Oxford University Press, 2000

MUHAMMAD IMRAN 125

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

TOPIC 88 Prediction of Exons

Introduction:
Exons can be predicted within ORF from the information gathered about
• splice junctions

Approach:
For each segment [acceptor, donor], we get three scores (coding potential, donor score,
acceptor score)
Various possibilities
─ all three scores are high – probably true exon.
─ all three scores are low – probably not a real exon.
─ all in the middle -- ?.
─ some scores are high and some are low -- ??

So here, we can get the evidence by the help of that information which we gathered from those
splice sites.
Prediction:
➢ Collect a set of exons and non-exons
➢ Score them using our scoring schemes
➢ Plot them as follows
➢ “draw” a separating line between exons and non-exons

Linear Discriminate Analysis:

─ linear discriminate analysis (LDA) finds an optimal plane surface that best separates
points that belong to two classes.
─ For example, if there are ten true exons and ten introns, and two feature.
─ These samples could be represented by 20 points in a two-dimensional space.
─ LDA would compute a straight line through the space that can best separate the two
classes with the minimal classification error

For example, if we
look into this picture,
there is the coding
region and the non-
coding region and
there is a line in the
middle that tries to
separate these two
which helps in
discrimination.

We can draw a
central line or linear
regression line, we
can see the equation
over there and by drawing this line we can have the prediction, this line fits in such a way that
it fits into most of the data points.

MUHAMMAD IMRAN 126

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

If we don’t know about data point, we can predict it while using this prediction line (angular
line in the figure).

Conclusion:
➢ Collect a set of exons and non-exons
➢ Score them using our scoring schemes
➢ Plot them as follows
➢ “draw” a separating line between exons and non-exons

References:
Biological Sequence Analysis
R Durbin, S Eddy, A Krogh and G Mitchison
Cambridge University Press, 1998.
Bioinformatics The machine learning approach
P Baldi and S Brunak
The MIT Press, 1998
Post-Genome Informatics
M Kanehisa
Oxford University Press, 2000

TOPIC 89 Annotation of Assembled Genome

A genome sequence is useless without annotation
Three steps in genome annotation:
• Find features not associated with protein-coding genes (e.g. tRNA, rRNA, snRNA,
SINE/LINE, miRNA precursors)

A genome sequence is useless without annotation

Three steps in genome annotation:
• Build models for protein-coding genes, including exons, coding regions, regulatory
regions

A genome sequence is useless without annotation

Three steps in genome annotation:
• Associate biologically relevant information with the genome features and genes

Ab initio methods:
Based on sequence alone
• Gene prediction algorithms (e.g. AUGUSTUS, Glimmer, GeneMark)

• RepeatMasker(repeat families)

Evidence-based Methods:
• Require transcriptome data for the target organism (the more the better)

• Align cDNA sequences to assembled genome and generate gene models:

TopHat/Cufflinks, Scripture

MUHAMMAD IMRAN 127

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Biological Annotations:
BLAST of gene models against protein databases
• Sequence similarity to known proteins

• InterProScan of predicted proteins against databases of protein domains

– Pfam, Prosite, HAMAP, PANTHER, …

• Mapping against Gene Ontology (function of those genes) terms

– GO terms

– BLAST2GO

Pattern Finding:
─ Much of the data processing in bioinformatics involves searching and recognizing
certain patterns within DNA, RNA or protein sequences.
─ In Biology it means finding motifs in DNA or proteins while in computational means it
is finding a pattern in a string

Conclusions:
After a genome is assembled, genome annotations are performed to identify gene and other
features in a genome

TOPIC 90 Pattern Finding in a Genome

Vocabulary:
─ A pattern (keyword) is an ordered sequence of symbols .
─ Symbols of the pattern and the searched text are chosen from a predetermined finite set,
called an alphabet (Σ)

Four Cases of Pattern Finding:

─ Look for a perfect match

─ Allow errors due to substitutions

─ Allow errors due to insertions-deletions

(InDels).

MUHAMMAD IMRAN 128

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Rank possible matches according to a weight function and keep matches above a certain
threshold

For every pair, aligned

together, we come up with the
weight function and in the end
we combine all those weight
functions

Generalized Algorithm:
Goal: Finding all occurrences of a pattern in a text
Input:
Pattern p = [p1…pn] of length n
Text t = [t1…tm] of length m
Output:
An indication that pattern P exist in T
or it does not exist in text T

Here we are using arrays of

data structure, so we have text
in those arrays and the indexes
are the positions of those
letters.

The total length is 14 (starting

from 0 and ending at 13), and the pattern length is 6.

So, the pattern matches on the seventh index, so what we will get in the end is that P is a
substring of T, starting from 7th index to 12th.
Conclusion:
Pattern searching algorithms search specific sequences in strands of DNA, RNA and proteins
having important biological meaning

TOPIC 91 Pattern Finding Algorithms

Methods Devised for Pattern Finding:
➢ Exact searching methods
➢ Approximate searching methods
➢ Position weight matrices
➢ Suffix trees

Exact Pattern Matching:

─ Given a pattern p of length m
─ and a string or text T of length n (m ≤ n)

MUHAMMAD IMRAN 129

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

─ Find all the occurrences of p in T

The matching needs to be exact, which means that the exact word or pattern is found
Exact Pattern Matching Algorithms:
➢ Naïve Brute Force algorithm
➢ Boyer-Moore algorithm
➢ Knuth Morris Pratt algorithm

Approximate Pattern Matching:

Also referred as approximate string matching or matches with k mismatches or differences
Also referred as approximate string matching or matches with k mismatches or differences
Given: a pattern p of length m
and
a string or text T of length n (m ≤ n)
Find: all the occurrences of substring X in T that are similar to p, allowing a limited number,
say k different characters in similar matches
Approximate Pattern Matching Algorithms
─ Dynamic programming approach
─ Automata approach
─ Filtering and automation algorithms

Position Weight Matrices:

Also known as position specific scoring matrices (PSSM)
• A matrix representing the frequencies of residues observed for a position in multiple
alignment

Suffix Trees:
It is a compressed tree containing all the suffixes and allows many problems on strings to be
solved quickly
Conclusions:
➢ Exact searching or pattern matching methods
➢ Approximate searching or pattern matching methods
➢ Position weight matrices.
➢ Suffix trees

TOPIC 92 Methods Devised for Pattern Finding

Introduction:
Also known as exhaustive search algorithm
➢ All the possibilities are explored and the best one is chosen
➢ For the task with many possibilities brute force will take too much time

Working:
➢ Searches patterns by going through the whole sequence nucleotide per nucleotide.
➢ Always shifts the window by exactly one position to right
➢ Requires 2n expected text character comparisons

MUHAMMAD IMRAN 130

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

When a mismatch the comparison stops and starts again by moving the pattern one position
forward
Algorithm:
Brute_Force(T,P)
n length[T]
m length[P]
For s 0 to n-m
Do if P[1..m]= T[s+1…s+m]
print “pattern occurs at position” s+1
Here, Brute_Force is the function which has two arguments (T= text and P= pattern) where n
records the length of T and m records the length of P and we start with a for loop which goes
from 0 to n-m, say for example we have ‘T’ which is of length 10 and ‘P’ which is of length 5
so it starts from 0 and will go till 5.

Do if P[1..m]= T[s+1…s+m]
This line is where we are saying that nucleotide number 1 of the pattern and we go up to its
whole length. In case of the text, we started with 0, so we add one over here, so 0+1 i.e. the 1st
nucleotide is compared till the last nucleotide.
So, if we find the occurrence of the patterns we will put that in the print statement and s+1 will
give its position.

Drawback:
The repetitive use of residues in comparison leads to runtime of O(mn), which makes it very
slow

MUHAMMAD IMRAN 131

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Conclusion:
Brute force is an exhaustive search method that takes long time as it does nucleotide by
nucleotide comparison

TOPIC 93 Knuth-Morris-Pratt Algorithm

Introduction:
A linear time algorithm for string matching
• Does not involve backtracking on string s i.e., repetitive comparison of nucleotide
residues

Components:
• Π
The Prefix Function

• The KMP Matcher

The Prefix Function Π:

Encapsulates knowledge about how the pattern matches against shifts of itself
• This information can be used to avoid useless shifts of the pattern ‘p’

• This enables avoiding backtracking on the string ‘S’

The KMP Matcher:

Given: string ‘S’, pattern ‘p’ and prefix function ‘Π’
Find: the occurrence of ‘p’ in ‘S’
Return: the number of shifts of ‘p’ after which occurrence is found

MUHAMMAD IMRAN 132

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Conclusions:
A linear time algorithm for string matching
• avoid useless shifts of the pattern ‘p’

• Does not involve backtracking on string S

Topic 94 Knuth-Morris-Pratt Algorith

MUHAMMAD IMRAN 133

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 134

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 135

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 136

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic # 95 Knuth-Morris-Pratt Algorithm

The KMP Matcher (S, p)

1 n  length[S]
2 m  length[p]
3 Π  Compute-Prefix-Function(p)
4 q  0 //number of characters matched
5 for i  1 to n //scan S from left to right
6 do while q > 0 and p[q+1] != S[i] //if next character does not match

7 do q  Π[q]

8 if p[q+1] = S[i] //next character matches

9 then q  q + 1

10 if q = m //is all of p matched?

11 then print “Pattern occurs at position” i – m+1

11 q  Π[ q] // look for the next match

Let us execute the KMP algorithm to find whether ‘p’ occurs in ‘S’.

For ‘p’ the prefix function, Π was computed previously and is as follows:

q 1 2 3 4 5 6 7
p a t a t a c a
Π 0 0 1 2 3 1 0

Initially: n = size of S = 15;

m = size of p = 7
Step 1: i = 1, q = 0 comparing p[1] with S[1]

MUHAMMAD IMRAN 137

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 138

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 139

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Complexity
➢ O(m) - It is to compute the prefix function values.
➢ O(n) - It is to compare the pattern to the text.
➢ Total of O(m + n) run time.

Advantages
➢ The running time of the KMP algorithm is optimal (O(m + n)), which is very fast
➢ The algorithm never needs to move backwards that makes the algorithm good for
processing very large files

Drawback

➢ Doesn’t work so well as the size of the alphabets increases Conclusion

➢ A fast linear time algorithm for string matching

MUHAMMAD IMRAN 140

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

96.Scoring Scheme
Scoring System: Introduction:
• Total score assigned to an alignment is sum of terms for each aligned pair of residues,
plus terms for each gap

∑ S(xi,yj) + d
d = linear gap penalty
Simple Alignment Scores:
• A simple way (but not the best) to score an alignment is to count 1 for each match and 0
for each mismatch

Scoring or Substitution Matrices:

• Used for scoring amino acid substitutions in pairwise alignments

• They reflect substitution rates that are originated by evolutionary events

Some of the substitution matrices to compute sequence alignments are:

• PAM: Point Accepted mutations

• BLOSUM: BLOCK Substitution Matrix

Substitution Matrices:
For a set of well known proteins:
• Align the sequences

• Count the mutations at each position

• For each substitution set the score to the log-odds ratio

For each substitution set the score to the log-odds ratio

Positive Score:
The amino Acids are similar, mutations from one into the other occur more often than expected
by chance during evolution
Negative Score:

MUHAMMAD IMRAN 141

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

The amino Acids are dissimilar, mutations from one into the other occur less often then expected
by chance during evolution

Conclusion
• Substitution matrices are the log-odds matrices used for scoring amino acid substitutions
in pairwise alignments

97.Substitution Matrices
Introduction:
• Substitution scores can be derived from probabilistic model

Notations:
➢ Let a pair of sequences x and y of length n and m
➢ Xi be the ith symbol in x yj be the jth symbol in y
➢ Symbols are from alphabet A
o A={A,T,G,C}
➢ A={twenty amino acids}
➢ Symbols from alphabet be a,betc

Objective:
Given a pair of aligned sequences, we want to assign a score to the alignment that gives a
measure of the relative likelihood that the sequences are related as opposed to being unrelated
Unrelated or random Model R:
Letter a occurs independently with frequency qaand the probability of two sequences is the
product of probability of each amino acid

MUHAMMAD IMRAN 142

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Match Model M:
─ Aligned pairs occur with a joint probability pab.
─ pabcan be thought of as the probability that the residues a and b have been independently
derived from some unknown original residue c in their common ancestor .
─ The probability of the alignment is

─ The ratio of two likelihoods can be calculated as;

P(x,y|M)
P(x,y|R)

Odds ratio:
─ The ratio of two likelihoods can be calculated as;

─ Logarithm can be used to have an additive score

S = ∑ S(xi,yj)
where S(a,b) = log(pab/qaqb)
log likelihood of (a,b) as aligned vs unaligned pair
Dayhoff PAM matrices:
Dayhoff, Schwartz and Orcutt (1978) presented their famous PAM (Point accepted mutations)
using substitution data from similar proteins then extrapolating this information to longer
evolutionary distances
S(a,b) = log(pab/qaqb)
incorporating the evolutionary time
S(a,b|t) = log P(b|a,t)/qb
Since pab/qa= P(b|a)
Values are rounded to near integer for computational convenience
• PAM250 is scaled by 3/log2 to give scores in third-bits

BLOSUM matrices:
─ Dayhoff matrices do not capture true difference between short time substitutions and long
term ones.
─ PAM matrices do not perform well in case of distantly related proteins.
─ BLOSUM matrices are derived from set of aligned, ungapped regions from protein
families called BLOCKS database (Henikoff&Henikoff 1992).
─ Sequences from each block was clustered together with score >L%.
─ Matrices with L= 62 and L= 50 known as BLOSUM62 and BLOSUM50 respectively.

MUHAMMAD IMRAN 143

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

─ BLOSUM62 is good for ungapped alignments and BLOSUM50 is good for gapped
alignments

Conclusions
• Substitution scores can be derived from probabilistic models

• PAM and BLOSUM are famous substitution matrices

98.Optimal Algos
Introduction:
Finding the path whose total score is maximal will give the best sequence alignment
• Two methods

– Local alignment

– Global alignment

Global Alignment:
It is an alignment that essentially spans the full extents of input sequences
• Hence it covers the entire length of sequences involved

• The Needleman-Wunsch algorithm finds best global alignment between two sequences

Local Alignment:
• It only covers parts of the sequences to be aligned

• Smith-Waterman algorithm finds the best local alignment between two sequences

Dynamic Programming:
• Dynamic programming is used to find an optimal alignment of two sequences and its
scores

• It is a method by which a larger problem may be solved by first solving smaller, partial
versions of the problem

• Three steps in dynamic programming:

• Initialization

• Matrix filling (scoring)

Traceback (alignment
─ Initialization: Create matrix with M+1 columns and N+1 rows where M and N
correspond to the size of sequences to be aligned

─ Matrix filling:Fill the matrix with highest possible score

MUHAMMAD IMRAN 144

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

─ Trace back:Move from the last corner and follow the arrow

Conclusion:
• Dynamic programming is used to find an optimal alignment of two sequences and its
scores

99.Needleman_wunch Algos
Introduction:
It performs global alignment on two sequences
• The algorithm was developed by Saul B. Needleman and Christian D. Wunsch and
published in 1970

Basic idea is
• to build up the best alignment by using optimal alignments of smaller subsequences

• It was the first application of dynamic programming to compare biological sequences

Steps:
Three steps
1. Initialization
2. Matrix filling
3. Traceback

Creating the Matrix:

Initial matrix is created with M+1 columns and N+1 rows
• Where M and N correspond to the length of sequences

Initialization:
• The cell of first row and first column of the matrix is initially filled with zero

• Add gap penalty for each shift to the right

Matrix Fill:
• Move through the cells row by row, calculating the score for each cell

• Compute three scores:

– A match score

– Vertical gap score

– Horizontal gap score

• The match score is the sum of the diagonal cell score and the score for a match

MUHAMMAD IMRAN 145

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• The horizontal gap score is the sum of the cell to the left and the gap score

• The vertical gap score is computed analogously

Traceback:
➢ The final step in the algorithm is the trace back for the best alignment
➢ Start at the bottom-right corner
➢ Follow where maximum value comes from
➢ F(M, N) is the optimal score, and from Ptr(M, N) can trace back optimal
alignment

Scoring Scheme:
• Scoring scheme introduced can be user defined

• It contains specific scores for match and mismatch residues as well as gap

Conclusions:
• Needleman and Wunsh Algorithm performs global alignment on two sequences using a
dynamic programming approach

100.Smith_waterman Algo
Introduction:
• Finds the best local alignment between two subsequences

Steps:
➢ Initialization
➢ Matrix filling
➢ Traceback or alignment

Trace-back:

MUHAMMAD IMRAN 146

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

F(M, N) is the optimal score, and from Ptr(M, N) can trace back optimal alignment
Creating the Matrix:
• Initial matrix is created with M+1 columns and N+1 rows

– Where M and N correspond to the length of sequences

Initialization:
• First row and first column of the matrix is filled with zero

Matrix Filling:

• Add the match or mismatch scores diagonally

• Add gap penalties vertically and horizontally

• Replace the negative values by zero

Traceback:
• Traceback starts with maximum value in the matrix and then go backwards

MUHAMMAD IMRAN 147

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Alignment:

Conclusion:
• Finds the best local alignment between two subsequences

Topic 101-102 DNA Sequencing

Magazine cut into millions of pieces. Problem of fragment assembly in DNA sequencing Short 500 to
700 nucleotide sequences per experiment, Assembling entire genome reassembling the magazine Both
problems are complicated by unavoidable experimental error. The data are frequently incomplete. DNA
sequencing methods-Fred Sanger and Walter Gilbert. Cells make copies of DNA DNA fragments of
different lengths- if one base is missing.

Sequencing of a 5386- nucleotide virus, DNA sequence data, Human Genome Project , 3
billionnucleotide sequence, DNA sequencing technology, Sequencing reads, Continuous genome,
DNA reading process, DNA sequencing machines, Single DNA fragment.

Shotgun sequencing

Sonicated, Inserts, Vector, Bacterial host, Cloning proce, DNA sequencing- inserts and
computational

Topic-103 DNA Array

Human Genome Project, Sequencing by Hybridization, DNA array (DNA chip) , Probes,
Hybridization, Weak chemical bond. DNA probes, Biochemical problem and the combinatorial
problem Science, SteveFodor and colleagues, Light-directed polyme synthesis –similarities to
computer chip manufacturing.

In Fig. An array with all 4l probes- 4 · l separate reactions Affymetrix-64-

kb DNA array in 1994 1-Mb or larger Probes- unknown target DNA-l-mer
composition Universal DNA array contains all 4l probes of length l .

MUHAMMAD IMRAN 148

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic-104 Sequencing by Hybridization

Unknown DNA sequence- length l , String s of length n, the l-mer composition is the multiset of n −
l + 1 l-mers in s and is written Spectrum(s, l).
If l = 3 and s = TATGGTGC, then
Spectrum(s, l) = {TAT, ATG, TGG, GGT, GTG, TGC}
Lexicographic order, like ATG,GGT,GTG, TAT, TGC, TGG

Sequencing by Hybridization (SBH) Problem:

Reconstruct a string from its l-mer composition
Input: A set, S, representing all l-mers from an (unknown) string s
Output: String s such that Spectrum(s, l) = S

Topic-105- Fragment Assembly in DNA Sequencing

500- to 700-bp DNA reads Reconstruct the entire genomic DNA sequence Fragment assembly
Error rate in DNA reads Sequencing errors, Assignments to read Major problem-repeats in
DNA-

20% of human genome Alu sequence-million times Repeats occur at many scales Human T-cell
receptor locus-Trypsinogen gene (4 kb).1 million Alu repeats ( 300 bp) and 200,000 LINE repeats (
1000 bp) 3 billion-letter sequence 500-letter reads, large number of repeats.

106. Strategy for Sequencing

Strategy for Sequencing
Clone into BAC, Sequenced each one-minigenome, Simplifies the computational assembly problem,
Human Genome project, Whole-genome assembly paradigm (James Weber and Gene Myers)Celera
Genomics Mate-pair reads sequencing technique.

Inserts of length L- both ends are sequenced, Mates at distance L- length is larger than most repeats

Most fragment assembly algorithms consist of the following three steps

Overlap: Finding potentially overlapping reads
Layout: Finding the order of reads along DNA
Consensus: Deriving the DNA sequence from layout Best match between the suffix of one read and
the prefix of another. Sequencing errors-Dynamic programming algorithm. constructing the layout is
the hardest step in fragment assembly. Whole-genome shotgun sequencing. Large number of reads to
ensure that experimental errors are reduced to minor noise.

MUHAMMAD IMRAN 149

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Topic-107- Protein Sequencing and Identification

Routinely sequenced proteins-Frederick Sanger-Nobel prize, Computational problem Edman
degradation reaction to chop off one terminal a.a, Sanger digested insulin with proteases
DNA sequencing “break-read the fragments-assemble”
Edman degradation reaction, 1960s protein sequencing machines were on the market DNA
sequencing technology Protein-Obtaining reads-problem Assembly- easy

108. Computational Protein Sequencing

Computational Protein Sequencing
Proteins produced in cell Sequence of previously unknown proteins. Proteins -biological system
and Range-Brain cells-Liver cells
De novo protein sequencing Protein identification Gedanken experiment Proteins constitute the
DNA polymerase
Protein sequencing and identification , Spliceosome , Matthias Mann and colleagues purified the
spliceosome complex

109. Mass Spectrophotometry

Programmed cell death, Survival factors, Developing nematode DNA sequence data Nervous
system- mutation in several genes, Mass spectrometry.

Mass spectrum of a peptide is a collection of masses of these fragments- Derive the sequence of a
peptide given its mass spectrum. For an ideal fragmentation process the peptide sequencing problem is
simple. The fragmentation process is not ideal, and mass spectrometers measure mass with some
imprecision.

110.111 Peptide Sequencing Problem

Peptide fragmentation is characterized by a set of numbers Δ = {δ1, . . . , δk} types of ions

The set of ion types. A δ-ion of an N-terminal partial peptide Pi is a modification of Pi that has mass
mi – δ, The most frequent N-terminal ions are called b-ions (ion bi corresponds to Pi with δ= −1) and
the most frequent C-terminal ions are called y-ions (ion yi corresponds to Pi- with δ =19)

MUHAMMAD IMRAN 150

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

112.Protein identification via Database Search

De novo protein sequencing algorithms- invaluable for known, unknown proteins. Useful for complete
spectra Spectra far from complete, De novo peptide sequencing algorithms. If we had access to a
database of all proteins from a genome, then we would no longer need to consider all 20l peptide
sequences to interpret an MS/MS spectrum, but could instead limit our search to peptides present in
this database . Database search “the back of a book”, Experimental spectrum , Sequence of the
experimental peptide, SEQUEST algorithm – John, Yates and colleagues.
Protein Identification Problem:
Find a protein from a database that best matches the experimental spectrum.
Input: A database of proteins, an experimental spectrum S, a set of ion types , and a parent mass m.
Output: A protein of mass m from the database with the best match to spectrum S

Topic-113 Modified Protein Identification Problem

SEQUEST Algorithm

Exhaustive search approach

Virtual database
Potential modifications
Combinatorial problem

Modified Protein Identification Problem:

Find a peptide from the database that best matches the experimental spectrum with up to k
modifications.
Input: A database of proteins, an experimental spectrum S, a set of ion types Δ, a parent
mass m, and a parameter k capping the number of modifications

Output: A protein of mass m with the best match to spectrum S that is at most k
modifications away from an entry in the database

Modified Protein Identification problem P1 and P2- S1 and S2, Notion of spectral similarity
Shared peaks count, Limitations in detecting similarities by database search.

Topic 114-Protein Structures

Background
Complex protein structures enable proteins to perform complex functions. We know over a million
protein sequences but only about 100,000 protein structures.
Why only 100,000 proteins for over million protein sequences
Estimating exact protein structures is very difficult. Its difficult to crystallize proteins. Even if we
manage to get protein’s X-Ray, to reconstruct the structure is extremely complex
Introduction
What if we could somehow predict protein structures?

MUHAMMAD IMRAN 151

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Since we know so many sequences, they can be used for predicting protein structures. This
indeed is possible and helpful.

The Basic Idea

1. Amino acids determine the protein structure

2. We have a large protein sequence dataset (uniprot)

Hence, we can fold protein sequences and predict their structures!

Why predict and why not exact solutions?
A deterministic solution of protein folding is a major unsolved problem in molecular biology!
Proteins fold spontaneously or with the help of enzymes or chaperones.
To predict we must first learn!

To computationally predict protein structures, we need to copy or mimic the natural folding! What
are the steps in protein folding and structure formation?
To fold we must learn the steps

Step 1: "Collapse"- leading to burial of hydrophobic AA’s

Step 2: Fluid globule - helices & sheets form, but are unorganized

Step 3: Compaction, and rearrangement of 2‘ structures

Conclusion
• Protein structure prediction involves learning how the amino acids in primary
sequence fold.

• Using this information, upon getting a protein sequence, we can try to

predict how it folds!

115. Predicting Secondary Structures

Background
Since the first step in protein folding is the formation of secondary structures, we must evaluate
which amino acids in the primary sequence prefer which secondary structures?
• By looking at the structures in PDB, we know that Alanine mostly found in Alpha
Helices. So if we have several Alanines in the sequence, then we can anticipate that a
helix may be formed by them

Introduction
What if we survey the entire PDB and check the presence of each amino in each type of secondary
structure
• If we know which amino acid is found in which specific secondary structure,
then we can use it for prediction!

Conclusion

MUHAMMAD IMRAN 152

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Several algorithms have been designed to predict 2’ given an amino acid sequence

• The first such algorithm was the Chou-Fasman Algorithm!

• We will see it in the upcoming modules!

116. 2’ Structures in Chou Fasman Algorithm

Background

• For a primary sequence, and a tentative 2’ structure,propensity table can help us

compute the overall propensity

• Product of propensity values is computed for overall propensity for each 2’ structure

Introduction

• An important point to note here is that 2’ structures are formed due to hydrogen
bonding between amino acids

• So, we need to consider the neighboring amino acids as well!

Conclusion

• You only need to compute propensities for a small number 2’ structures

• The highest net propensity will be the most probably secondary structure that will be
formed!

117.121 Chou Fasman Algorithm -

Only a small number of combinations of secondary structures are possible due to their individual
properties. Such as 4 amino acids are needed to start an Alpha Helix and 5 amino acids for Beta
Sheet Note that besides the alpha helix and beta sheets, LOOPS are an other secondary
structure.

How can loops be integrated into predicting 2’ structures?

• Loops are small ~ 3-4 amino acids

MUHAMMAD IMRAN 153

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Conclusion
• For Alpha Helices, 4 contiguous amino acids are required

• Their Alpha-Helix propensity should be more than 1.0

• Once this propensity falls below 1.0, Alpha-Helix stops

Chou Fasman Algorithm - II

Background
Alpha Helices are formed from 4 contiguous amino acids having an Alpha-Helix
propensity over 1.0. The Alpha-Helix stops if this propensity falls below 1.0! Introduction
• Once Alpha Helices are constructed, and concluded, the remaining amino acids can
be evaluated for Beta sheets and turns etc

• Let’s see how Beta sheets are evaluated using Chou Fasman Algorithm

Conclusion
• Alpha Helices can be finalized if their propensity is higher than the propensity for
Beta Sheets in regions of 5 amino acids

• For those regions where that is not the case, further evaluation is required

Chou Fasman Algorithm - III

Alpha Helices were finalized if their propensity was higher than the propensity for Beta
Sheets in regions of 5 amino acids. For those regions where that are not the case, what should
be done?
We can evaluate such regions for Beta Sheets.
Let us see step by stop how to find a beta sheet and how to differentiate them from
alpha helices

MUHAMMAD IMRAN 154

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Conclusion
• Using the strategy of higher propensity, alpha helices and beta sheets can be
completely resolved

• Assignments for each beta sheet and alpha helix can be finalized

• But what about the loops?

Chou Fasman Algorithm - IV

Background
• After computing the propensity of alpha helices and beta sheets, we need to
settle for loops

• Let’s see how can we find out the loops using Chou Fasman Algorithm

Conclusion
• Chou Fasman Algorithm helps predict Alpha Helices, Beta Sheets and Turns

• The algorithm is based on statistical occurrence of Amino Acids in known structures

122.125. Chou Fasman Algorithm – Flowchart I, II & III.

• Chou Fasman Algorithm helps predict secondary structures such as Alpha Helices,
Beta Sheets and Turns. Step by step flowchart of the entire algorithm. Beta sheets
can be predicted from primary amino acid sequences

MUHAMMAD IMRAN 155

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 156

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Chou Fasman Algorithm helps predict secondary structures from amino acid
sequences. Step by step flowchart of the algorithm for extracting Alpha Helices

• Alpha helices, beta sheets and turns can be predicted using Chou Fasman Algorithm.
This algorithm is based on statistical analysis of amino acid occurrences in proteins.

Chou Fasman Algorithm – Improvements

• Secondary structure propensity values of alpha helix, beta sheet and turns should be
recalculated with the latest protein data sets.

MUHAMMAD IMRAN 157

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Special consideration for:

➢ Nucleation regions
➢ Membrane proteins
➢ Hydrophobic domains
➢ Consider variable coil and loop sizes besides the from tetra peptide turns
➢ Consider local protein folding environments
➢ Solvent accessibility of residues
➢ Protein structural class
➢ Protein’s organism

Conclusion
Chou Fasman can be improved to better predict secondary structures by incorporating
biochemical factors and updated statistics!

126. Summary of Visualization, Classification & Prediction

A. Why do we need to visualize proteins?
B. Which atoms are used to reconstruct proteins?
C. Where are the positions of these atoms stored?

Structure Classification

A. What is the relationship between protein structure and function?

B. What is the need to classify proteins
C. Hierarchy of classification

Structure Prediction
A. Why structure of proteins are important?
B. Why are so few structures reported till date?
C. Benefits of predicting structures

Conclusion
Structure visualization, classification and prediction equip us to perform functional evaluation of
proteins! This is important for understanding disease and designing drugs for treating them

Topic-127- Introduction to Homology Modelling

Proteins are 3D molecules with their own unique structures. Protein structure is reflective of
the protein function. Protein structure includes 1’, 2’, 3’ and 4’ structures. 1’ structure of
proteins is the sequence of proteins and can be obtained by mass spectrometry. 2’ structures
formed by proteins are the helices, beta sheets, loops and coils. 3’ structure of proteins is the
combination of 2’ structures such that the overall protein structure is formed. 4’ protein
structure is formed when two or more proteins complex together. X-Ray Crystallography and
NMR Spectroscopy are used to find the structures of proteins. However, these methods are
difficult and expensive. Solution: Prediction of structures. Protein sequence gives rise to its
structure. If another protein which has a similar sequence also has its structure known, the
structure of an unknown protein can be predicted based on that similar protein . So, it is then
possible to identify unknown protein structures by just examining the homologous protein
sequences.

Conclusions

MUHAMMAD IMRAN 158

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Sequence Identity

• Alignment Length

Which combination of identity and alignment length is suitable for best for structure
prediction?

128.Homology, Paralogy and Orthology

In homology modelling, proteins with similar 1’ sequences are considered. Given that one of
them has its 3’ structure known, then the 3’ structure of other protein can be predicted

Conclusions
• Good sequence alignment and identity ensures that homology modelling will give
accurate results

129. Workflow of Structural Modelling

Homology modelling is used to predict structures of proteins having high sequence similarity
with other proteins with known structures! Overall, there are three different strategies for
structure prediction
1. Homology Modelling

2. Threading/Fold Recognition

3. Ab Initio Modelling

130.135 Seven Steps to Homology Modelling – I

Protein structure can be predicted by 3 methods:

1. Homology Modelling
2. Fold Recognition / Threading
3. Ab Initio Modelling

Let’s start by looking at Homology Modelling. There are seven salient steps in any Homology
Modelling pipeline. Definition of Template (known) & Target (unknown). Homology modeling of the
target structure can be done as follows:

MUHAMMAD IMRAN 159

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

1. Template recognition and initial alignment

2. Alignment correction
3. Backbone generation
4. Loop modeling
5. Side-chain modeling
6. Model optimization

7. Model validation

MUHAMMAD IMRAN 160

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

MUHAMMAD IMRAN 161

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

So, Homology modelling works in seven steps. It is a repetitive process

136.MODELLER for Homology Modelling

• Template recognition and initial alignment
• Alignment correction
• Backbone generation
• Loop modeling
• Side-chain modeling
• Model optimization
• Model validation

Modeller is a software for homology modelling

salilab.org/modeller
Inputs:
Python script file, Sequence alignment & Template (PDB)

.log : log output from the run.

.B* : model generated in the PDB format.
.D* : progress of optimisation.
.V* : violation profile.
.ini : initial model that is generated.

MUHAMMAD IMRAN 162

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

.rsr : restraints in user format.

.sch : schedule file for the optimisation process.

Homology modelling helps predict protein structures by using prior structural

informationSeveral tools are available to perform homology modelling in a programmatic
or automated way!

137.139.Fold Recognition – Threading

When should we use Fold Recognition?

Introduction

• A protein fold is defined by the way the secondary structure elements of the structure are
arranged relative to each other in space.
• Common folds include 4-helix bundle and the TIM barrel.
• 5,000 stable folds in nature
• Fold recognition: Finding the best fit of a sequence to a set of candidate folds
Fold recognition is also called Threading. Technique for predicting protein structures.
Employed when homology modelling cannot predict quality structures. A protein fold is
defined by the way the secondary structure elements of the structure are arranged relative to
each other in space. Common folds include 4-helix bundle and the TIM barrel. 5,000 stable
folds in nature. Fold recognition: Finding the best fit of a sequence to a set of candidate folds.

MUHAMMAD IMRAN 163

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Fold recognition or Threading is a technique for predicting protein structures. It is useful in

cases where homology modelling fails to predict quality structures

Threading involves “passing” the amino acid sequence through each fold in the database. The best
match is computed using a scoring function. Combinations of secondary structures come together to
form the best prediction. Scoring typically involves using a Z-Score function based on energy of a
molecule.

140 Online Tools for Threading – iTasser

iTASSER helps thread amino acid sequences on fold and secondary structure databases. It
also helps predict function of structures output.

MUHAMMAD IMRAN 164

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

141.Advantages and Disadvantages of Threading

Fold recognition or Threading is a technique for predicting protein structures. It is useful in cases
where homology modelling fails to predict quality structures.

Advantages
Threading helps predict secondary structures of proteins towards tertiary structure prediction. For
the “Twilight Zone” with low alignment quality and identity, threading is use.

Disadvantages
Novel proteins cannot be predicted using threading. Fewer than 30% of the predicted first hits are
true remote homologues. Validation of each result is necessary.

142. 3D-1D Bowie Algorithm

Homology employed high alignment scores. Threading worked by creating combinations of primary
sequences and corresponding secondary structures. Proposed by Bowie et al in 1991. It converts 3D
structure into a 1-D string profile for each structure in the fold library. Align the target sequence to
these profiles. 3D-1D methods convert structure and environment information into “profiles”. Score
for each amino acid is computed for each profile. Inputs and outputs of 3D-1D
• Identify amino acids based on: protein core, side chain positioning, solubility etc. (6
in all)

• Part of secondary structure including -helix, -sheet etc (3 in all)

• Total of 3 x 6 = 18 distinct states

• Pa: j= prob. of finding amino acid (a) in environment (j)

• Pa=probability of finding (a) anywhere

• Maximize sum of scores for the fold:

143. Introduction to Ab Initio Modelling

Ab initio methods have Anfinsen’s thermodynamic hypothesis at the center . These methods attempt
to identify the structure with minimum free energy. Ab initio methods rely on computing the energies
of folded proteins. The protein structures with the lowest energy are deemed as plausible predictions.

Need for Ab Initio Modelling

• Applicable to any sequence

• Not very accurate biologically

• Accuracy and applicability are limited by our understanding of the protein folding
problem

Limitation

• Computationally expensive

MUHAMMAD IMRAN 165

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Suitable for proteins with less than 100 residues

144.Rationale of Ab Initio Modelling

Ab initio methods rely on computing the energies of folded proteins. The protein
structures with the lowest energy are declared as plausible predictions
Rationale
Sometimes it so happens that even slightly homologous proteins may not be available. This renders
homology modelling and threading/fold recognition as futile . Also, newer protein structures continue
to be discovered every day. These could not have been identified by methods which only rely on
matching with available structures. Lastly, homology / fold recognition predict protein structures
without computing fundamental physical/chemical properties of the mechanisms and driving forces in
structure formation. Ab initio methods, in contrast, base their predictions on physical models for these
mechanisms. Energy released during the folding process is computed for predicting structure.

145.Strategies for Ab Initio Modelling

Ab initio methods base their predictions on physical models of folding mechanisms.
Stabilization is measured by energy released during the folding process. Start with an
energy function. Fold structures in order to obtain the most stable structure. This structure
will have the minimum energy.
Energy Optimization in Ab Initio Modelling
1. Start with a rough initial model.

2. Define an energy function mapping structures to energy values. We have to minimize this
later!!

3. Solve the computational problem of finding the global minimum.

Simulation of the Folding Process

1. Build an accurate initial model (including energy and forces).

2. Accurately simulate the dynamics of the protein folding process.

3. The native structure will steadily emerge.

146.Energy States of Folded Proteins

Ab initio methods predict protein structures by folding proteins based on each constituent
atom’s volume, charge, mass etc. The protein structure reporting lowest energy is
selected to be the optimal structure.
Energies of Bonded Atoms vs. Nonbonded Atoms

MUHAMMAD IMRAN 166

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

147.Local versus Global Minima

The protein structure reporting lowest energy is selected to be the optimal structure
Best Case Energy Function
• Clear energy minimum in the native structure

• Viable path towards this minimum

• Global optimization finds the most stable structure

Optimal Energy Function

• Easier to design and compute

• Native structure not always at the global minimum

• No clear way of choosing among alternative structures that are generated

148. Pros and Cons of Ab Initio Modelling

Native structure not always at the global minimum. No clear way of choosing among
alternative structures that are generated Advantages
• Ab Initio methods can fold any target sequence using only physical atomic properties

• Predictions are mostly accurate and correctly describe the natural folding process

Disadvantages
• Ab initio methods are the very difficult to design (energy function)

• These methods are slow due to the huge possibilities

149.151 Summary of Structural Modelling

Strategies for Structural Modelling
• Homology Modelling

• Fold Recognition

• Ab Initio Modelling

Homology modeling of the target structure can be done as follows:

1. Template recognition and initial alignment

2. Alignment correction

3. Backbone generation

4. Loop modeling

5. Side-chain modeling

6. Model optimization

MUHAMMAD IMRAN 167

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

7. Model validation

Energy Optimization in Ab Initio Modelling

1. Start with a rough initial model.

2. Define an energy function mapping structures to energy values. We have to minimize this
later!!

3. Solve the computational problem of finding the global minimum.

Simulation of the Folding Process

1. Build an accurate initial model (including energy and forces).

2. Accurately simulate the dynamics of the protein folding process.

3. The native structure will steadily emerge.

Conclusion
• Homology modelling is performed in cases of high identity and alignment score

• For the “Twilight zone”, other strategies are employed

• For low identity and alignment scores, a “Twilight zone” for structure prediction
exists

• Fold recognition / threading is useful in such cases

• For cases where even the fold libraries do not give any high scoring matches, Ab
Initio strategies can help model the structure

• However, this is a complex and computationally expensive process

152.Review of Sequence Analysis

Important Concepts
How do we sequence:
Genomes, Proteomes
How do we compare sequences:
Pair-wise Sequence Alignment, Multiple Sequence Alignment
Types of Alignments:
Global Alignment , (Needle Wunsch), Local Alignment, (Smith Waterman)
• Advanced Tools:
Fast Alignment (FASTA), Basic Local Alignment Search Tool,
(BLAST) Databases: GenBank, UniProt Online Portals:
Ensemble, Expasy, UniProtKB

MUHAMMAD IMRAN 168

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

153.Review of Phylogenetics
• Important Concepts

Molecular Evolution
Insertions, Deletions, Substitutions
Phylogenetic Trees
Scaled Trees, Unscaled Trees
Phylogenetic Trees
Rooted Trees, Unrooted Trees

UPGMA: Unweighted Pair – Group Method using arithmetic Averages

Two sequences with with the shortest evolutionary distance between them are considered
These sequences will be the last to diverge, and represented by the most recent internal node.
Clustering Vs. Non-clustering Methods:
UPGMA is a clustering method
Maximum Parsimony etc are non-clustering methods (not included in this course).

154.Review of Protein Sequencing

Important Concepts
Techniques of protein sequencing, Edman Degradation, Mass Spectrometry, Protein Ionization, Mass
Analysis, Protein Fragmentation, MS1, MS2, Estimating and scoring whole protein mass, Extracting
& Scoring Peptide Sequence Tags, Searching Post-translational Modifications
Composite Scoring Schemes
Online tools:
Mascot, Sequest, Prosight PC

155 Review of RNA Structure Prediction

156.Review of Protein Structures

Important Concepts

Protein Structures are generally of four types:

Primary, Secondary, Tertiary, Quaternary

Techniques for determining protein structures

X-Ray Crystallography, NMR Spectroscopy
Types of Protein Secondary Structures
Helices, Beta Sheets, Coils, Loops

• Foundation of structure prediction algorithms

Propensities of certain amino acids to form specific secondary structures

MUHAMMAD IMRAN 169

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Algorithm for predicting protein structures

Chou Fasman Algorithm

• Protein Structure Database – PDB, Online tools for predicting structures by using
proteins sequences

157.Review of Homology Modelling

Four Strata of Protein Structures
Primary, Secondary, Tertiary, Quaternary
• Justification for homology modelling

Number of known protein sequences is much larger as compared to known proteins

structures
Three Strategies for Structure Prediction
Homology Modelling, Fold Recognition, Ab Initio Modelling, Protein Structure Database – PDB,
Online tools for predicting structures such as MODELLER and iTASSER

158.Conclusions from this Course

Definition of Bioinformatics, Need for Bioinformatics, Areas within Bioinformatics,
Bioinformatics as an interdisciplinary area, Need to store, process and analyze biological data,
Requirement of newer faster algorithms Specific areas focused were:
Comparing sequences, Comparing structures, Predicting structures
We looked at:
Algorithms, Databases, Online Tools for each topic.

• We studied the basic algorithms for each topic, With evolution and growth of
Bioinformatics, newer and better algorithms are now also available!

159.Advanced Follow-up Courses

We looked into the foundations of Bioinformatics, However, each topic that was studied
has a undergone a lot of development.

• For advanced study in Genomics, you may take “Computational Genomics” course
Topics:
Genome Assembly, Gene Finding, Annotation, GWAS etc

• For advanced study in Proteomics, you may take “Computational Proteomics”

course. Topics: Protein Sequencing, PTM search, Structure Modelling and PPI
studies

• For advanced study in Integrative Biology, you may take “Systems Biology”
course. Topics: Metabolomics, Transcriptomics, Network Biology etc

Also, now there are cutting edge courses on:

Nano-Bio-IT, Computational Drug Design, Personalized Medicine

MUHAMMAD IMRAN 170

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

160.Careers in Bioinformatics
Pakistan as an infrastructure-limited country. The onset of digital revolution. Emergence of data as
the most precious commodity, globally. Specifically, health data as a key commodity of the future.
Health and disease as the primordial challenge of mankind
• Unique opportunity for us in Pakistan

Bioinformatics requires two things

1. Smart mind
2. Internet connected computer

One man company

You can take public databases and design drugs. One man vs. Roche?
BIGDATA
You can make a startup company which manages and process health BIGDATA. All it needs is basic
software development skills coupled with Bioinformatics
The next disruption
The next Google, Facebook and Uber is going to emerge from Health and Bioinformatics.
Pharmaceutical companies are investing into bioinformatics human resource development
Jobs Market
Pharmaceutical Giants, Research Centers & Universities, Hospital & Diagnostic IT departments ,
Your own startup company

Topic-161 RNA Structure

Outline

• RNA folding

• Dynamic programming for RNA secondary structure prediction

• Covariance model for RNA structure prediction

Base Pairing

• RNA bases A,C,G,U

• Bases can only pair with one other base

• Canonical Base Pairs

• A-U

• G-C

• “wobble” pairing

• G-U

• I-U

MUHAMMAD IMRAN 171

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• I-A & I-C

Canonical Base Pairs

A-U, G-C

• “wobble” pairing

G-U, I-U, I-A & I-C

RNA Types messenger RNA

(mRNA) Non-coding RNA

transfer RNA (tRNA), ribosomal RNA (rRNA), small interfering RNA (siRNA), micro RNA (miRNA), small
nucleolar RNA (snoRNA)

MUHAMMAD IMRAN 172

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

162.RNA Secondary Structure

Some form of RNA can form secondary structures by “pairing up” with itself. This can change its
properties dramatically.
Base Pairing
Pairing of bases helps in determining the secondary structure

Aligning bases, based on pairing with each other gives an algorithmic approach to
determining the optimal structure

RNA Folding
RNA is produced as a single stranded molecule (unlike DNA)
• Strand folds upon itself to form base pairs & secondary structures

• RNA sequence analysis is different from DNA sequence

RNA Structure
Structures are more conserved than sequences •
Covariation

Secondary Structure representation

2D, Circle plot, Dot plot, Mountain, Parentheses, Tree model

MUHAMMAD IMRAN 173

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Pseudoknots

Pseudoknots cause a breakdown in the Dynamic Programming Algorithm. In order to form a

pseudoknot, checks must be made to ensure base is not already paired – this breaks down the
recurrence relations
Conclusions
Some forms of RNA can form secondary structures by “pairing up” with itself. This can change its
properties dramatically.

163.RNA Secondary Structure Prediction

There are different approaches to predict secondary structure of RNA
• Energy minimization

Comparative sequence analysis

• Folding and alignment

• Base-Pair Maximization

1. Energy minimization

Dynamic programming approach

Does not require prior sequence alignment. Require estimation of energy terms contributing to
secondary structure
Assumptions;
Energetically most stable structure is more likely structure. Energy associated with any position is
only influenced by local sequence and structure. Neglect pseudoknots

Approach

Energy minimization algorithm predicts secondary structure by minimizing the free energy ( G). G
calculated as sum of individual contributions of:
• Loops

• stacking

Energy minimization
• Thermodynamic Stability

• Estimated using experimental techniques

• Theory : Most Stable is the Most likely

• No Pseudknots due to algorithm limitations

• Uses Dynamic Programming alignment technique

• Attempts to maximize the score taking into account thermodynamics

• MFOLD and ViennaRNA

Drawbacks

MUHAMMAD IMRAN 174

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Compute only one optimal structure. Usual drawbacks of purely mathematical approaches

MUHAMMAD IMRAN 175

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

2. Comparative sequence Analysis

Need a multiple sequence alignment as input.
Requires sequences be similar enough (so that they can be initially aligned), Sequences should be
dissimilar enough for covarying substitutions to be detected, comparative analysis produces accurate
structure predictions.

164.RNASeq
Calculating transcript abundance and prevalence by Ultra high throughput cDNA sequencing
(Mortazwi et al, 2008)
The sequence reads are individually mapped to the source genome and counted to obtain the number and
density of reads corresponding to RNA from each
• known exon

• splice event

• new candidate gene

Procedures
Isolation of all mRNA, Convert to cDNA using reverse transcriptase, Sequence the cDNA, Map
sequences to the genome.
The more times a given sequence is detected, the more abundantly transcribed it is.
If enough sequences are generated (> 40 Million), a comprehensive and quantitative view of the entire
transcriptome of an organism or tissue can be obtained (Mortazvi et al, 2008)
Data analysis
Mapping reads, Visualization Genome browser , De novo assembly , Quantification, Differential
Gene Expression, Functional Analysis, Gene Networks
In RNASeq, transcript abundance and prevalence is calculated using Ultra high throughput cDNA
sequencing

165.RNASeq Normalization
Sequencing reactions may vary across different sequencing plate-forms as well as within different
lanes of the same sequencer. Transcript lengths also vary
Raw read counts may vary
RNASeq challenges
Uniformity of sequence coverage, Quantity of sequence required to reliably detect RNAs of lower
abundance classes, Quantification and conversion of relative quantification to absolute RNA
concentrations, Transcriptomes of organisms with large genomes, containing genes with more
complicated structure, present some special challenges
Mapping Biases
Read counts will be higher if sequencer produces more reads. Longer genes will have the probability of
mapping more reads than smaller ones
RPKM (reads per kilobase of transcript (or exon model) per million mapped reads)
Method for quantification of transcript levels. RPKM measure of read density reflects the molar
concentration of a transcript in the starting sample by normalizing for RNA length and for the total read

MUHAMMAD IMRAN 176

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

number in the measurement. This facilitates transparent comparison of transcript levels both within and
between samples
RPKM
Number of reads mapped per gene length in KB per total reads in that sample in millions
C = Count of Mapped Reads
L= Length of transcript
M = Mapped reads of sample

RPKM

How many reads are required to map at 1 RPKM with a transcript of 2Kb length from a total of 40
Million Mapped reads?

FPKM (Fragments per kilobase of transcript per million mapped reads)

• Paired end RNASeq experiments produce two reads per fragment

• FPKM counts fragments not the reads

• Both reads might not map

• Counting reads might doubble count some fragmnets

(Trapnell et al 2010)

RPM

While comparing the same genes expression across different samples (treatments), normalizing for

gene may not be necessary

Can compare relatively bigger numbers and get more DEG

RPKM reflects the molar concentration of transcript in starting sample normalized for
• Length of RNA

MUHAMMAD IMRAN 177

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Total reads in the sample

It facilitates transcript comparison within and across samples Topic

166.Neural Network
The human brain can be described as a biological neural network an interconnected web of neurons
transmitting elaborate patterns of electrical signals A neural network is a “connectionist”
computational system. Information is processed collectively in parallel throughout a network of nodes.
Complex adaptive system.
• Learning processes in biological systems.

• Learning as an optimization process.

• Learning by modification of synaptic strength.

167.Association Rule Mining

It is an important data mining model studied extensively by the database and data mining community.
Assume all data are categorical. No good algorithm for numeric data. An association rule has two
parts
• an antecedent (if)

• a consequent (then)

Antecedent is an item found in the data. A consequent is an item that is found in

combination with the antecedent Market basket transactions: t1: {bread, cheese,
milk} t2: {apple, eggs, salt, yogurt}

… …
tn: {biscuit, eggs, milk} Concepts:
An item: an item/article in a basket I:
The set of all items sold in the store A
transaction:
Items purchased in a basket; it may have TID (transaction ID) A
transactional dataset:
A set of transactions

168.Clustering
Clustering is “a process of organizing objects into groups whose members are similar in some
way” A cluster is therefore a collection of objects which are “similar” between them and are
“dissimilar” to the objects belonging to other clusters.
• Simplifications
• Pattern detection
• Useful in data concept construction
• Unsupervised learning process
Hierarchical agglomerative general algorithm
• Find the 2 closest objects and merge them into a cluster

MUHAMMAD IMRAN 178

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Find and merge the next two closest points, where a point is either an individual object
or a cluster of objects

• If more than one cluster remains, return to step 2

Applications
• For administrative purposes

• Hospital activity and performance

• Used for researchers • Data used by physician

• For laboratory use etc.

169.Machine Learning
Programming computers to optimize performance criterion using example data or past experience
When to learn? Calculate payroll , Solution needs to be adapted to particular cases (user biometrics)
When To Learn
Human expertise does not exist (navigating on Mars), Humans are unable to explain their expertise
(speech recognition), Solution changes in time (routing on a computer network)
Model
Build a model that is a good and useful approximation to the Data
KDD is the non-trivial process of identifying valid, novel, potentially useful, & ultimately
understandable patterns in data
Applications
Retail , Finance , Manufacturing , Medicine , Telecommunications , Bioinformatics, Web mining
Retail: Market basket analysis, Customer relationship management (CRM)
Finance: Credit scoring, fraud detection
Manufacturing: Optimization, troubleshooting
Medicine: Medical diagnosis
Telecommunications: Quality of service optimization
Bioinformatics: Motifs, alignment
Web mining: Search engines
Machine Learning

Study of algorithms that Improve performance at some task with exp. Role of Statistics, Role of CS

Applications of ML

Speech recognition, NLP, Computer vision, Medical outcomes analysis/Computational biology, Robot
control

170.ML Concepts
❖ Association Analysis
❖ Supervised Learning
• Classification
• Regression/Prediction
❖ Unsupervised Learning

MUHAMMAD IMRAN 179

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

❖ Reinforcement Learning

Learning Associations
Basket analysis:

P (Y | X ) probability that somebody who buys X also buys Y where X and Y are
products/services.

Example: P ( chips | Beer) = 0.7

TID Items
1 Bread, Milk
2 Bread, Diaper, Beer, Eggs
3 Milk, Diaper, Beer, Coke
4 Bread, Milk, Diaper, Beer
5 Bread, Milk, Diaper, Coke

Classification Apps

FR: Pose,lighting,occlusion(glasses,beard), make-up, hair style

Character recognition:

Speech recognition:

Medical diagnosis: From symptoms to illnesses

Web Advertizing

Retail: Market basket analysis, Customer relationship management (CRM)

Finance: Credit scoring, fraud detection

Manufacturing: Optimization, troubleshooting

Medicine: Medical diagnosis

Telecommunications: Quality of service optimization

Bioinformatics: Motifs, alignment

MUHAMMAD IMRAN 180

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Web mining: Search engines

171.ML Applications
Supervised Learning

• Prediction of future cases

• Knowledge extraction
• Compression
• Outlier detection

Un-supervised Learning

Learning “what normally happens”

No output

Clustering: Grouping

similar instances

Applications

• Customer segmentation in CRM

• Image compression
• Bioinformatics

Reinforcement Learning

Policies: what actions should an agent take in a particular situation

Utility estimation: how good is a state (→used by policy)

No supervised output

Delayed reward

Credit assignment problem (what was responsible for the outcome)

Applications:

Game playing Robot in a maze Multiple agents, partial observability, ...

MUHAMMAD IMRAN 181

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

172.Forensic Science

173.Advantages of bioinformatics in forensic

DNA Finger Printing / DNA Profiling:
It is a form of forensic identification that is used primarily to identify people.
Advantages
Even though all humans share 99.9% of their DNA sequences, the remaining 0.01% of sequences
isunique enough to differentiate people.Because DNA is in every cell of a person’s body is
present and same , any part of a human’s body including dead skin skills, hair, saliva, and more
contain DNA sequences.

174. .Methods used in DNA profiling.

DNA finger printing
A forensic scientist will need two pieces of DNA to be compared.For example, DNA discovered at
crime scene should be compared to a DNA sample taken from a suspect.
Steps
• Comparison of repeating DNA
sequences
• Match b/w two samples
DNA extraction from cell
1. Collecting cell from sample:
Two meters of DNA in cellCollect cells from the sample with buccal swab.Place the swab into
Eppendorf tube.
2. Burst cells open to release DNA:
Add the lysis solution to the tube to separate the cells.LS breaks Cell membrane& nuclear envelope
causing cells to burst open& release DNA . It also removes histones proteins from DNA.
3. Separate DNA from proteins and Debris:

MUHAMMAD IMRAN 182

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Cells have stayed in warm to for such a time that DNA is freed from cells.Salt causes proteins & other
cellular debris to clump together.Place tube into micro centrifuge.Inside the centrifuge tube spin
around and debris &heavy proteins sink in bottom of tube and DNA strands remains distributed
throughout liquid.
4. Isolate concentrated DNA:

Add the liquid containing DNA in separate tube Now add isopropyl alcohol to tube.DNA is not soluble
in this alcohol so it comes out and it can be seen with naked eye.DNA is collected at bottom tube after
placing in centrifuge.

175. DNA Profile

• Encrypted sets of numbers that reflect a person's DNA makeup,
• Variable number tandem repeats(VNTRS).
• Short term tandem repeats (STR) in making the DNA profile of a person.
• These DNA profiles are the basis of a national DNA databases.
DNA profiling processes

• RFLP analysis.
• PCR analysis.
• STR analysis.
• Amp FLP.
• Y-chromosome analysis.
• Restriction Fragment length Polymorphism (RFLP)

RFLP
• It analyzes the lengthof strands of DNA that include repeating base pairs (VNTRs).
• Repeated sequence of human genome can be same but the number of times it is repeated is
unique to everyone.
RFLP analysis requires investigators to dissolve DNA in an enzyme that breaks the strand at specific
points.The number of repeats affects the length of each resulting strand of DNA.Investigators compare
samples by comparing the lengths of the strands.
Example: CAT is repeated continuously 13 times in a row. In somebody else, it might be 12 times or 14
or whatever.
Limitation: RFLP analysis requires a fairly large sample of DNA that hasn't been contaminated with dirt.

176. DNA Profile Methods

MUHAMMAD IMRAN 183

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Polymerase Chain

Reaction (PCR)

Replicate a small amount of DNA to create a larger sample for analysis. First, a heat-stable DNA polymerase
-- a special enzyme that binds to the DNA and allows it to replicate -- is added. Next, the DNA sample is
heated it to 200 degrees F (93 degrees C) to separate the threads. Then the sample is cooled and reheated.
Reheating doubles the number of copies. Process is repeated about 30 times, there is enough DNA for
further analysis.

Analyzing STRs

PCR is the first step in analyzing STRs (Short Tandem Repeats), which are very small, specific
alleles in a variable number tandem repeat (VNTR).

Short Tandem repeats

Analyzing STRs is more accurate than the RFLP technique because their small size makes them easier to
separate. If you want to create a fingerprint, you might look at 20 different STRs at different places in
order to create a profile.

STRs
• It is impossible for two persons to have same number of STR repeated in a given sequence.

Y-chromosome Analysis

STRs in Y-chromosome Useful if the sample has mixed DNA. Gender analysis cases. It is processed just
like simple STR analysis.

AmpFLP

Amplified fragment length polymorphism, is another technique that uses PCR to replicate DNA. Like RFLP,
it first uses a restriction enzyme. Then, the fragments are amplified using PCR and sorted using gel
electrophoresis. Can be automated , Doesn't cost very much., DNA sample must be high quality otherwise
errors may result, which is the case with most DNA analysis techniques.

Topic 177- Introduction to Drug Discovery

Drug Discovery
Primary objective－
design & discovery of new compounds that are suitable for use as drugs
A team of workers－
chemistry, biology, biochemistry, pharmacology, mathematics, medicine & computing …
requirement
(i) Synthesis of the drug
(ii) Administration method
(iii) Development of tests
(iv) Procedures to establish how it operates in the body
(v) safety assessment
(vi) research into the biological and chemical nature of diseased state.
Drugs: Definition
Chemical substances that are used to prevent or cure diseases in humans,
animals and plants
Activity: Pharmaceutical/pharmacological effect on the subject, e.g. Analgesic or β-blocker
Potency: quantitative nature of the effect
Drugs Properties: ADMET

MUHAMMAD IMRAN 184

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Drug: agent used for the psychotic effect by the media or general public. Even the drugs abused have
their activity. No drug is completely safe. Suitable quantity to cure or excess to be poisonous! E.g.
aspirin, paracetamol can be toxic if excesses.

178. Drug Discovery Applications:-

Areas Influencing DD

• Molecular Biology on Drug Discovery

• High-Throughput Screening
• Combinatorial Chemistry

Molecular Biology Influence

Genetic information

Biochemical and chemical terms.Cloning and expressing genes that encode

therapeuticallyuseful protein

High Throughput Screening

Widely used in the pharmaceutical industry. Automation to quickly assaythe
biological or biochemical activity of a large number of drug-like compounds.

Combinatorial Chemistry
Laboratory technique in which millions of molecular constructions can be
synthesized and tested for biological activity.

MUHAMMAD IMRAN 185

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

179.Pharmacogenetics
It is the branch of pharmacology concerned with the effect of genetic factors on reactions to drugs.

How people respond to medicines, Correlating heritable genetic variation to drug response.

Defination:-

Biotechnological science combines techniques of medicine, pharmacology & genomics which

developing drug therapies to compensate for genetic differences in patients which cause varied
responses to a single therapeutic regimen.

MUHAMMAD IMRAN 186

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Applications:-
1. Detection of genetic variability of drug effects on the genome level
2. Agent selection
3. Analysis of drug
4. reactions and drug toxicity on gene expression
5. Development of new indications for already approved drugs
6. Discovery of new drug targets
7. Identification of (non) responders in clinical trials of phase I-IV
8. Identification of genotype dependent adverse drug reactions
9. Identification of individuals at risk for severe adverse drug effects

180.Pharmacogenomic applications
How genes affect persons response to drugs. Pharmacology (science of drugs). Genomics (the study of
genes and their functions). Develop effective, safe medications & doses tailored to a person’s genetic
makeup.

Applications:-
Improve drug safety, Reduce ADRs, Tailor treatments to meet patients unique genetic
predisposition, Optimal dosing, Improve drug discovery and Improve proof of principle for
efficacy trials.

Future
Blessing in research. As a simple example, for nearly a decade the ability to store more information on
a hard drive has enabled us to investigate a human genome sequence cheaper.

181.182. Drug Discovery – Pipeline

Target Identification, Target Validation, Lead Identification, Lead Optimization, Pre-Clinical

• Pharmacology &Toxicology

MUHAMMAD IMRAN 187

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

183.Drug Discovery Methods

Past:

(i) Identification of active ingredient from traditional remedies

(2) serendipitous discovery.

Current:

Diseases are controlled at molecular & physiological level. Information of Human Genome

MUHAMMAD IMRAN 188

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Methods for DD

✓ Random Screening
✓ Molecular Manipulation
✓ Molecular Designing
✓ Drug Metabolites • Serendipity

Random Screening

Higher/crude plants, opium, senna, reserpine, etc. Penicillin microorganism. Antibacterials with
improved therapeutic profiles.

Molecular Manipulation

Drug Metabolism

Xenobiotic metabolism Biochemical modification of pharmaceutical substances or xenobiotics by

living organisms, usually throughspecialized enzymatic systemsLipophilic chemical compounds into
more

readily excreted hydrophilic products.Rate of metabolism determines duration & intensity of a drug's
pharmacological action
Serendipity

Prototype psychotropic

drugs
Development of psychiatry
Finding of one thing while looking for something else

MUHAMMAD IMRAN 189

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

184.Biomedical annotated corpora

Biomedical researchers are interested

1. Understanding data
2. context information
3. background knowledge
4. curated databases &
5. Literature extensive

What is Annotated Corpora?

Dataset for extraction of disease/treatment entities relations. Corpora are usually constructed for
training or evaluation purposes during the development of particular system

Annotation Consumers
The linguistic community typically uses annotation as training data or for specific tasks. An abundance of
tools that can produce annotations in the specific format of those resources. Biomedical annotation
typically used for gene set enrichment analysis

Information deluge

Bio-databases, controlled vocabularies and bio-ontologies encode small fraction of information

Linking text to dbs and ontologies

Curators struggling to process scientific literature. Discovery of facts & events crucial for gaining insights
in biosciences: need for text mining

185.Steps for Creation of Biomedical Corpora

✓ Nature of data
✓ Standard datasets
✓ Formalism
✓ Users of dataset
✓ Evaluation measures

Data Mining

MUHAMMAD IMRAN 190

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

✓ Discovers unsuspected associations

✓ Combines & links facts
✓ and events
✓ Discovers new knowledge, finds new associations

IR: yields all relevant . Corpora; gathers, selects, filters documents that may prove useful , Finds
what is known
IE: extracts facts & events of interest to user, Finds relevant concepts, facts about concepts. Finds
only what we are looking for

– IE systems can be developed by referencing annotated corpus.

– The performance of IE systems can be evaluated by being compared to the

annotated corpus.

Text Mining Pipelines

✓ Text documents
✓ Retrieval/storage Indexaccess relevant storage
✓ Text Processing: word Filters, Pattern filters,Lexicon matching,Ontology, NLP
parsingetc, …

Feature Extractions:

Statistical: Word counts, pattern extraction & counts, etc

Domain-specific: Gene Name counts, etc

NLP-specific: Phrase counts, etc

Data Mining: Classification, Clustering, Association, Statistical Analysis,Visual Analysis,

etc …

186.Target Discovery Strategy

Target discovery to clinical application saga

• Physiology-based approach

• Target-based approach

1. Physiology-based approach

Is a disease-centric approach in which target is not identified, multiple targets are involved. In vivo
screening is done by using drugs, siRNAor antisense oligonucleotides.

MUHAMMAD IMRAN 191

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

This process relies on disease, not on the target.

2. Target-centric approach

Target based discovery starts with the identification of genes and their protein products. Aim to
develop drugs affecting one gene or a molecular mechanism. The identification of diseaserelevant
genes in vitro cellular models has been possible due to several tools. Gene-suppression tool used to
linked the genes with disease.

Target has two types

1. Genetic
2. Mechanistic

Genetic targets are represented by genes and genes products. Mechanistic targets include mechanism

based targets such as receptors, enzymes or genes, identified on the basis of the disease state.

187.Strategies To Identify Possible Drug Targets

First step of drug discovery is to identification of disease-associated targets. Genome sequencing and
screening have enhanced opportunities for target identification and lead optimization.
Structure-based Target Discovery
It helped in defining the contours of the cognate surfaces of ligands and their protein targets, permitting
optimizationof their potency and selectivity. Some drugs that are originated from structure-based
approach:

• Dorzolamide

• Captopril

• Imatinib

• Zanamivir

The protein structure contributes in the following fields:

• Target identification from sequence structure homolog recognition.

• Structural genomics and drug targets.

• Identification of ligand binding region

• Identification of hits and leads

• Structure-guided design and screening

Kinase drug discovery & Kinome

Target Discovery through Cell-based Genetics

MUHAMMAD IMRAN 192

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• New drugs are based on target identification.

• It wants the thorough knowledge of the disease processes and characterization of genes.

• Combination of three target discovery will provide the desired result.

Target Discovery Strategies

Target discovery strategies based on

• Expression profiling Proteomic approach to identify disease related genes based on differential
EP, homology and post translational modification

• Biochemical and cell biological assays - To identify genes and proteins linked with disease
pathways

• Cell-based genetics - Leads to the discovery of targets by disturbing gene function in whole
organisms, corrleation with phenotypes.

Cell-based Genetics

Cell-based assays may lead to the identification of genes involve in cellular transformation, activation,
migration and a host of biological processes relevant to a human disease.
Genetic-based Target Identification It
has some methods:

• Positional Cloning- Laboratory technique used to locate the position of a disease associated
gene along the chromosomes.

• Candidate gene approach

To identify complex disease-linked genes through SNP markers. 10 million in HGP and 3 million
identified.
Target class genetic approach

• Is applied to drug target gene families such as proteases, ion channels and GPCRs.

• 24000 protein coding genes & 2400 DTs

• Best candidate are selected from gene family for genetic analysis.

188. Target Validation

• A potential target is identified in the context of a specific disease.

• A validated target is the one that can be manipulated with drugs to produce positive clinical
effects in humans.

Requirements for TV

 Genetic Approach
Gene to disease correlation in animal model

• Forward Genetics
• Reverse Genetics

Target Validation Tools

Helps in assessing their genetic association with disease. The following tools are:

• Antisense agents
• Ribozymes

MUHAMMAD IMRAN 193

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Peptide nucleic acids

• Transcription factors
• Gene knockouts

Antisense Technology

• Drugs as molecules action cellular proteins (enzymes, receptors).

• Downstream: Block events at nuclear/ribosomal levels.
• Prevent expression of aberrant proteins at earlier stage.
• Oligonucleotide reagents

Ribozymes

• Small RNA mol. cleave other RNA mol. at sequence specific sites.
• Hairpin Ribozyme: GUC

HammerHead: NUH, H is AUC

• Mode of Delivery: Exogenously , Endogenously

• Peptide NAs: block protein translation
• Transcription Factors: Zinc finder proteins – high Affinity to bind to correct region of DNA.
Gene Knockouts:

189.Lead Identification
• Compound for biological activity on target.
• Potency threshold
• Libraries of molecules

Lead Identification – Technologies

Virtual Screening:
Protein structure , docking Chemical similarity search, Knowledge of compounds against receptor,
receptor structure & receptor ligand interactions

Visual Screening
MLCC: Multilevel chemical Compatibility scoring

• Top selling drugs

• Compounds under biological scrutiny
• Anticancer drugs
• Compounds with poor drug like characters

MUHAMMAD IMRAN 194

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

Pharmacophore Mapping
Identify lead compounds against a desired target
Definition: 3-D arrange… Usage:
interaction of receptor & legend
DB concept

• QSAR

Quantitative structure activity relationship

• SAR:

synthesizing & testing a series of structurally related compounds

Least squares & KNNs
High Throughput Docking

• Ligand & protein

• Docking algos
• Force fields, knowledge based & empirical

NMR based screening

• Nuclear Magnetic
• Resonance
• 3-D potential DC & tertiary structure of Proteins
• Need of prior information
• SHAPES
• WaterLogsy

Chemical Genetics

• Gene-product function in cellular or organismal context using exogenous ligands

MUHAMMAD IMRAN 195

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)

 lOMoARcPSD|49403025

• Knockouts
• Cell cycle - arresting agents

190. Lead Optimization

• Optimize the desirable traits of the lead
• Lead should be amenable for chemistry optimization
• Methods from Lead Identification

Lead Optimization – Technologies

• Medicinal chemists conduct extensive SARs to improve potency and selectivity.

• Improve physicochemical and drug-like properties
• Best molecules are advanced to animal Models & preliminary toxicology

LO Methods
De novo drug desing
Charge distribution, liphophilicity or pka of side chains and H-bonds donors and acceptors
SBDD
Structure based drug Design. Effective: 3-D structure of inhibitor with target known. Large no of
medicines. Molecular recognition in protein ligand complexes
Drug Like properties

• ADMET in phase 1
• Filters
• Bioavailability, PK, CNS
Pre- Clinical Pharma-cology & Toxicity

• Animals testing
• Xenograft models
• ADME/T testing and validation

MUHAMMAD IMRAN 196

Downloaded by Muhammad Muzammal Idrees (clpreaction@gmail.com)