Course: MDMBIO1: Introduction to Bioinformatics

Notes of Module 3
Sequence Analysis

• Contents:

Types of biological sequences: DNA, RNA, Protein, Pairwise and

Multiple
Sequence Alignment (MSA), Scoring matrices (PAM, BLOSUM),
Tools: BLAST,
FASTA, ClustalW, Applications: gene finding, phylogeny, structure
prediction

--

Types of biological sequences: DNA, RNA, Protein:

Biological sequences are the fundamental building blocks of genetic

information and the functional molecules that carry out cellular processes.
These sequences are essentially a linear arrangement of monomers, and
their specific order determines their structure and function. The three
primary types of biological sequences are DNA, RNA, and protein.

DNA (Deoxyribonucleic Acid)

DNA is a double-stranded molecule that serves as the genetic blueprint for
all cellular life. It's found in the nucleus of eukaryotic cells and the
cytoplasm of prokaryotic cells. The information in DNA is encoded in its
sequence of nucleotides.

Structure: DNA is a polymer made up of monomers called nucleotides. Each

nucleotide
has three components: a deoxyribose sugar, a phosphate group, and a
nitrogenous base. There are four types of nitrogenous bases: adenine (A),
guanine (G), cytosine (C), and thymine (T).

Double Helix: DNA exists as a double helix, with two strands wound
around each other. The strands are held together by hydrogen bonds
between the nitrogenous bases, following specific base-pairing rules: A
always pairs with T, and C always pairs with G. The two strands are also
antiparallel, meaning they run in opposite directions.

Function: The primary function of DNA is to store and transmit genetic

information. It contains the instructions for building and maintaining an
organism. DNA also serves as a template for its own replication, ensuring
that genetic information is passed on to new cells during cell division.

RNA (Ribonucleic Acid)

RNA is a single-stranded nucleic acid that plays a critical role in gene
expression. It acts as an intermediary, carrying genetic information from
DNA to the protein-making machinery of the cell.

Structure: RNA is also a polymer of nucleotides. However, it differs from

DNA in three key ways:

The sugar in RNA is ribose, not deoxyribose.

The nitrogenous base uracil (U) replaces thymine (T).

RNA is typically single-stranded, which allows it to fold into complex three-

dimensional structures.

Function: RNA has diverse functions, with several types playing distinct
roles:

Messenger RNA (mRNA): Carries the genetic code from DNA in the nucleus
to the ribosomes in the cytoplasm, where proteins are synthesized.

Transfer RNA (tRNA): Acts as an adaptor molecule, bringing the correct

amino acid to the ribosome during protein synthesis based on the mRNA
code.

Ribosomal RNA (rRNA): A major structural and catalytic component of

ribosomes, the molecular machines that assemble proteins.

DNA and RNA structure

Protein (Polypeptides)
Proteins are large, complex molecules that carry out most of the work in
cells. They are essential for the structure, function, and regulation of the
body's tissues and organs.
Structure: Proteins are polymers made from monomers called amino
acids. There are 20 common types of amino acids. The specific sequence
of these amino acids is called the primary structure of the protein. This
sequence folds into a unique three-dimensional shape, which is essential
for its function.

Folding: Protein structure is described at four levels:

Primary Structure: The linear sequence of amino acids.

Secondary Structure: Local folding of the polypeptide chain into structures

like alpha-helices and beta-pleated sheets, stabilized by hydrogen bonds.

Tertiary Structure: The overall three-dimensional shape of a single

polypeptide chain, resulting from interactions between amino acid side
chains.

Quaternary Structure: The arrangement of multiple polypeptide chains

(subunits) to form a larger, functional protein complex.

Function: Proteins perform a vast array of functions, including:

Enzymes: Catalyze biochemical reactions.

Structural proteins: Provide support and shape to cells and tissues (e.g.,
collagen).

Transport proteins: Carry molecules across membranes or throughout the

body (e.g., hemoglobin).

Hormones: Act as chemical messengers (e.g., insulin).

Mechanism of Enzyme Action

The Central Dogma of Molecular Biology

The relationship between these three sequences is described by the
Central Dogma of Molecular Biology, which states that genetic information
flows from DNA to RNA to protein. This process involves two main steps:

Transcription: The process where a DNA sequence is copied into a

complementary RNA sequence (specifically mRNA) by an enzyme called
RNA polymerase.

Translation: The process where the mRNA sequence is used as a template

by ribosomes to synthesize a specific protein. tRNA molecules "read" the
mRNA code and bring the corresponding amino acids to the ribosome to
be linked together.

Protei
Feature DNA RNA n
Nucleotides (A, T, Nucleotides (A, U, G,
Monomers G, C) Amino acids (20
C) types)
Deoxyribos
Sugar e Ribose —
polypeptid
Single e
Strands Double-stranded Single-stranded
chain
geneti Transfer &
Store c expression Perform cellular
Function
function
information of info s
single-
Database 20 AA letter
A,T,G,C A,U,G,C
Letters codes

2. Pairwise and Multiple Sequence Alignment (MSA)

Sequence alignment is a fundamental process in bioinformatics used to

identify regions of similarity between biological sequences (DNA, RNA, or
protein). These similarities often indicate functional, structural, or
evolutionary relationships. There are two main types: pairwise and
multiple sequence alignment.

Pairwise Sequence Alignment

Pairwise sequence alignment compares two sequences to each other to
find the best possible alignment. This is done by lining up the sequences
and introducing gaps to maximize a similarity score. The core idea is that
the most similar arrangement of two sequences will have the highest
score.

Types of Pairwise Alignment

 • Global Alignment: This method attempts to align the entire length
of both sequences. It's best used when the sequences are of similar
length and are expected to be homologous (evolutionarily related)
along their whole length. The Needleman-Wunsch algorithm is a
classic example of a global alignment method that uses dynamic
programming to find the optimal alignment.
• Local Alignment: This method finds the most similar regions
between two sequences, even if they are otherwise very different.
This is useful for sequences that are long but only share a small
conserved domain or motif. The Smith-Waterman algorithm is
the standard for local alignment, also using dynamic programming.

How it Works

Both algorithms use a scoring matrix to evaluate the alignment.

• Match: A positive score is given for matching characters (e.g., A-A).

• Mismatch: A negative score is given for mismatched characters
(e.g., A-C).
• Gap Penalty: A negative penalty is given for introducing a gap to
align a sequence (e.g., A-). This penalty discourages the algorithm
from introducing too many gaps.
Example:
AATGCATGGATAT
TT
ACTGCACGGCTTT
AT Score=05
Q2. Write a python program to take 2 sequences as input for
pairwise analysis. If mismatch at some position, then specify the
position as well.
Q3. Write a python program to take 2 sequences as input for
pairwise analysis. If match then add 1 to score and if mismatch at
some position then subtract 1 from the score, then specify the
position as well as the score.

The algorithm then uses dynamic programming to fill a matrix and find
the path with the highest cumulative score, which represents the optimal
alignment.

Multiple Sequence Alignment (MSA)

Multiple sequence alignment (MSA) is the alignment of three or more

biological sequences. It's a more complex process than pairwise alignment
but provides richer information. MSA is used to infer evolutionary
relationships, identify highly conserved regions, and predict protein
structure and function.
How it Works
Aligning multiple sequences simultaneously is computationally
challenging. Most MSA algorithms use heuristic approaches to find an
approximate solution rather than an optimal one. The most common
method is progressive alignment.

• Progressive Alignment: This approach builds the final alignment

in a series of steps:
1. Pairwise Alignments: All sequences are aligned in pairs to
calculate a similarity score for each pair.
2. Guide Tree Construction: These scores are used to build a
guide tree (a type of phylogenetic tree) that represents the
evolutionary relationships between the sequences.
3. Progressive Alignment: The most closely related sequences
(the "leaves" of the tree) are aligned first. The resulting
alignments are then treated as a single "profile" and are
progressively aligned with other sequences or profiles
according to the order defined by the guide tree.

Key Tools

Several software tools are widely used for MSA, including:

• Clustal Omega: A popular tool that uses a progressive alignment

approach.
• MUSCLE: Known for its speed and accuracy, particularly with large
datasets.
• MAFFT: Another fast and powerful tool that uses Fast Fourier
Transforms.

Summary and Key Differences

Sequenc Multiple Sequence Alignment

Feature Pairwise e (MSA)
Alignment
Number of Two Three or more
Sequences
To identify similarity
and To find conserved regions, infer
homology between a evolutionary relationships, and
Purpose pair predict
of structure/function for a group of
sequences.
sequences.
Less complex.
Exact
Computatio Highly complex. Heuristic
nal methods (dynamic methods are
Complexity programmin are used to find approximate
 g) solutions.
feasible.
Needleman-Wunsch
Algorith Smith Progressive alignment (e.g.,
ms (Global), - Clustal
Waterman Omega), Iterative alignment
(Local)
 Sequenc Multiple Sequence Alignment
Feature Pairwise e (MSA)
Alignment
A single alignment of A single alignment of all
two sequences,
Primary often with a consensus sequence
Output sequences with a and a
similarity visual representation of
score. conservation.

3. Scoring Matrices (PAM, BLOSUM)

Scoring matrices are tables used in bioinformatics to assign scores to the

alignment of two characters (e.g., amino acids or nucleotides) during
sequence alignment. They are essential for determining the quality of an
alignment and for identifying homologous sequences, as they reflect the
likelihood of one character mutating into another over evolutionary time.
The two most common types for protein sequences are PAM and BLOSUM.

How Scoring Matrices Work

Scoring matrices are typically 20×20 tables for amino acids, where the
value at the intersection of a row and a column represents the score for
aligning one amino acid with another. A positive score indicates that the
substitution is considered likely (or "conservative"), while a negative score
indicates it is unlikely (or "non-conservative").

The score is often a log-odds ratio, calculated as:

where:

• Sij is the score for aligning amino acid i with amino acid j.
• Pij is the observed probability of amino acid i and j aligning in a set
of related sequences.
• qi and qj are the background probabilities (or frequencies) of amino
acids i and j occurring by chance.

A positive log-odds score means the substitution occurs more often than
expected by chance, suggesting an evolutionary or functional relationship.
A negative score means it occurs less often than expected.
PAM (Point Accepted Mutation) Matrices

PAM matrices, also known as Dayhoff matrices, were the first widely used
set of substitution matrices. They were developed by Margaret Dayhoff
and colleagues in the 1970s.

• Derivation: PAM matrices are based on an evolutionary model.

They were calculated by analyzing substitutions in globally
aligned, highly similar protein sequences (at least 85%
identity). This small evolutionary distance allowed researchers to
assume that most substitutions were single, one-step mutations.
• The PAM Unit: A "1 PAM" unit of evolutionary distance is defined
as the amount of evolutionary time in which 1% of the amino
acids are expected to have been replaced.
• Creating Different PAM Matrices: A PAM1 matrix is generated from
the initial data. To create matrices for greater evolutionary distances
(e.g., PAM250), the PAM1 matrix is mathematically extrapolated by
multiplying it by itself a specific number of times. For example,
PAM2=(PAM1)2, PAM250=(PAM1)250.
• Numbering Convention: Higher PAM numbers correspond to
greater evolutionary distance and are used for aligning more
distantly related sequences. A PAM250 matrix is used for very
divergent sequences, while PAM30 is for closely related ones.

BLOSUM (Blocks Substitution Matrix) Matrices

BLOSUM matrices were developed by Steven and Jorja Henikoff in the

early 1990s as an alternative to PAM.

• Derivation: BLOSUM matrices are derived directly from the analysis of

locally aligned conserved blocks of amino acids in a large
database. These "blocks" are ungapped segments from multiple
sequence alignments of related proteins.
• Clustering: To avoid bias from closely related sequences, the
Henikoffs clustered sequences within a block that had more than a
certain percentage of identity. They then calculated substitution
frequencies from these clusters, not from individual sequences.
• Numbering Convention: Unlike PAM, the numbering in BLOSUM
matrices relates to the clustering threshold. A BLOSUM62 matrix,
for example, is derived from sequences with a maximum of 62%
identity.
• Usage: Lower BLOSUM numbers (e.g., BLOSUM45) are used for
aligning more distantly related sequences, as they were created
from blocks with lower
 sequence identity thresholds. Higher BLOSUM numbers (e.g.,
BLOSUM80) are used for aligning closely related sequences.
BLOSUM62 is the most commonly used matrix because it works well
for a wide range of alignments, striking a balance between detecting
close and distant relationships.

Key Differences and Applications

PAM
Feature Matrices BLOSUM Matrices
Local alignments of
conserved
Data Globa alignment ver sequenc
Source l s of y similar blocks of es with
sequences (>85% varying
identity). identities.
explic evolutionar Based on empirical data
Based on an it y from
Evolutionar model matrice greate
y ; s for r alignments; each matrix is
distance extrapolate directl calculate
Model s are d from a y d from a
matrix for short specific
distances. dataset.
number
Lower s mean more
Highe number evolutiona distanc
r s mean more ry e (e.g.,
evolutiona distanc
Numbering ry e (e.g., BLOSUM45 is for distant
PAM250 is for distant
sequences). sequences).
Historically significant, but less Widel today
used y used , with
Common today for general purposes due BLOSUM62 being the
Use to its standard
age and for general-purpose sequence
assumptions.
alignment.

4. Tools: BLAST, FASTA, ClustalW

BLAST (Basic Local Alignment Search Tool)

BLAST is a powerful and widely used bioinformatics tool for comparing a
query sequence (DNA, RNA, or protein) against a large database of
sequences. Unlike global alignment algorithms that try to align the entire
sequence, BLAST is a heuristic algorithm that finds high-scoring, locally
similar regions between a query and sequences in the database.

How BLAST Works: A Heuristic Approach

BLAST's speed comes from a clever "seed-and-extend" strategy that

avoids a full dynamic programming alignment for every sequence in the
database. Here's a simplified breakdown:
 1. Word Generation: The algorithm breaks down the query sequence
into short "words" (typically 3 amino acids for proteins, 11 for DNA).
For example, the protein sequence GATGC would be broken into
words like GAT, ATC, TGC.
2. Neighboring Words: For each word, BLAST generates a list of
"neighboring words" that score above a certain threshold when
aligned with the original word using a scoring matrix like
BLOSUM62. This step accounts for conservative amino acid
substitutions.
3. High-Scoring Pairs (HSPs): BLAST searches the database for
exact matches to these words and their neighbors. When a match is
found, it's considered a "hit." These hits are called High-Scoring
Pairs (HSPs).
4. Extending the HSPs: The algorithm then extends the HSPs in both
directions along the sequences, adding more aligned characters as
long as the alignment score continues to increase. The extension
stops when the score drops below a certain threshold.
5. Statistical Significance: BLAST uses statistics to determine if an
alignment is significant. The E-value (Expectation value) is the
key metric. It represents the number of alignments with a score as
good as or better than the observed score that would be expected
to occur by chance in a database of that size. A lower E-value
means a more significant and less random match. For example, an
E-value of 10−5 means there's a 1 in 100,000 chance that the
observed similarity is random.

Types of BLAST Programs

There are different versions of BLAST tailored for specific types of

sequence comparisons:

• blastn: Compares a nucleotide query sequence against a

nucleotide database.
• blastp: Compares an amino acid query sequence against a protein
database.
• blastx: Translates a nucleotide query into all six possible protein
frames and compares them against a protein database. This is
useful for identifying potential protein-coding genes in an unknown
DNA sequence.
• tblastn: Compares an amino acid query sequence against a
nucleotide database that has been translated in all six frames.
Useful for finding protein homologs in genomic DNA.
• tblastx: Translates both the nucleotide query and the nucleotide
database into all six reading frames and compares them. This is
computationally intensive and used for comparing highly divergent
nucleotide sequences.
FASTA
FASTA is both a file format and a sequence alignment tool. As a tool, it was
one of the first widely used heuristic methods for sequence database
searching. The algorithm is similar to BLAST's but predates it and has some
key differences in its approach.

FASTA File Format

The FASTA file format is a simple, text-based format for representing

biological sequences. It's now the standard for sequence submission and
retrieval. A FASTA file consists of:

1. A single-line header that begins with a > symbol, followed by a

sequence identifier and description.
2. One or more lines of sequence data, which can be wrapped to fit a
certain line length.

Example: >sequence_identifier_1 A short description of the sequence

GATTACA AATTAGGA

The FASTA Algorithm

The FASTA algorithm also uses a heuristic approach for speed. It's a three-
step process:

1. Finding High-Scoring Regions: It identifies regions of high

similarity between the query and database sequences by finding
short, identical "words" (k-tuples). It calculates a score for these
initial matches.
2. Scoring and Rescoring: The algorithm then joins the highest-
scoring regions together, allowing for gaps. It recalculates a more
accurate alignment score for these joined regions using a
substitution matrix.
3. Dynamic Programming: Finally, it performs a full dynamic
programming alignment (like the Smith-Waterman algorithm) on a
narrow band around the highest-scoring alignment from the
previous step. This "banded" approach makes the final step
computationally efficient.

FASTA is generally considered less sensitive and slower than BLAST, but it
remains a foundational tool in bioinformatics.

ClustalW

ClustalW is a popular and widely used program for performing multiple

sequence alignment (MSA). Its purpose is to align three or more
biological sequences (protein or DNA) to find conserved regions and infer
evolutionary relationships.

The ClustalW Method: Progressive Alignment

ClustalW uses a progressive alignment method, which is a heuristic
approach to solve the computationally difficult problem of aligning
multiple sequences. The process is as follows:

1. Pairwise Alignment: ClustalW first performs a pairwise alignment

of every possible pair of sequences in the input set. These
alignments are scored to determine how similar each pair is.
2. Guide Tree Construction: The scores from the pairwise
alignments are then used to build a guide tree. This tree is a
hierarchical representation of the sequences' relationships, where
the most similar sequences are clustered together. A commonly
used method for building this tree is the Neighbor-Joining
method.
3. Progressive Alignment: The guide tree dictates the order of
alignment. The program starts by aligning the most closely related
pair of sequences at the leaves of the tree. This alignment is treated
as a "profile."
4. Profile Alignment: The program then progressively aligns the next
closest sequence or profile to the growing alignment. For example,
it might align a third sequence to the profile of the first two, and
then a fourth sequence to the profile of the first three, and so on,
until all sequences are included.

Features and Limitations of ClustalW

• Versatility: ClustalW can align both protein and DNA sequences

and offers a variety of scoring matrices (e.g., BLOSUM, PAM) and
gap penalties for customization.
• Output: The output is a single alignment of all the input sequences,
often with a "consensus" line that highlights conserved residues.
• Limitations: Progressive alignment is a greedy algorithm. A
potential drawback is that errors made early in the alignment (when
aligning the most similar sequences) cannot be corrected later. This
can lead to inaccuracies, especially with large datasets of distantly
related sequences. More modern tools like Clustal Omega and
MUSCLE have improved on this by using iterative refinement
strategies.

[Link]: Gene Finding, Phylogeny, Structure Prediction

Applications of bioinformatics and sequence analysis: gene finding,

phylogeny, and structure prediction.

Gene Finding (Gene Prediction)

Gene finding, or gene prediction, is the process of identifying the
sequences of DNA that are biologically functional. This includes identifying
protein-coding genes, which contain the instructions for building proteins,
as well as RNA genes (tRNA, rRNA, etc.) and regulatory regions. Gene
finding is a crucial first step in understanding the function of a genome.

Challenges of Gene Finding

The human genome is vast, and only a small portion of it codes for
proteins. Identifying these specific regions is difficult for several reasons:

• Introns and Exons: In eukaryotes, protein-coding genes are

interrupted by non-coding sequences called introns. The coding
regions, or exons, must be spliced together to form the final
protein. Predicting the boundaries of these introns and exons is a
major challenge.
• Regulatory Sequences: Promoters, enhancers, and other
regulatory elements that control gene expression are often short
and can be located far from the genes they regulate, making them
hard to identify.
• Variability: The signals that mark the beginning and end of genes
(e.g., start and stop codons) are not always unique and can occur
randomly in non-coding DNA.

Computational Approaches

Gene finding tools use a combination of different approaches to solve

these challenges:

1. Ab Initio Methods (First Principles): These methods predict

genes based on the intrinsic properties of the DNA sequence itself.
They look for specific statistical patterns that are characteristic of
genes, such as:
o Codon Usage Bias: Different organisms have a preference for
certain codons (three-nucleotide sequences that code for an
amino acid). Ab initio methods use this bias to distinguish
coding from non-coding regions.
o Promoter and Splice Site Signals: They identify conserved
sequences that mark the beginning of a gene (promoters) and
the boundaries of introns and exons (splice sites).
o Open Reading Frames (ORFs): A long stretch of DNA without
a stop codon is a strong indicator of a protein-coding region.
2. Homology-Based Methods: This approach relies on comparing a
new, unknown sequence to a database of known, annotated genes
from other organisms. If the new sequence has a high degree of
similarity to a known gene, it's a strong indicator that it's also a
gene. Tools like BLAST are fundamental to this method.
 3. Transcriptome-Based Methods (RNA-Seq): This is the most
direct and accurate method. RNA is isolated from a cell, and the
RNA transcripts (which represent the expressed genes) are
sequenced. The resulting sequences are then mapped back to the
genome to identify the exact locations of exons. This method
provides direct experimental evidence for gene expression.

Most modern gene finders use a hybrid approach, combining ab initio

predictions with evidence from homology and RNA-Seq data to produce
highly accurate gene models.

Phylogeny (Evolutionary Relationships)

Phylogeny is the study of evolutionary relationships among groups of

organisms (or sequences) and is a cornerstone of evolutionary biology. By
comparing sequences, we can reconstruct the evolutionary history of life
and build phylogenetic trees. These trees depict the branching patterns
of descent from a common ancestor.

The Process of Phylogenetic Analysis

Building a phylogenetic tree from molecular data involves several steps:

1. Sequence Collection: Collect a set of homologous sequences

(DNA, RNA, or protein) from the organisms or genes of interest.
Homology implies that the sequences share a common evolutionary
origin.
2. Multiple Sequence Alignment (MSA): This is the most crucial
step. The sequences are aligned to identify regions of similarity and
highlight differences. The alignment ensures that homologous
positions (positions that evolved from a common ancestral position)
are compared correctly. Errors in the alignment can lead to an
incorrect tree. Tools like ClustalW or MUSCLE are used for this
purpose.
3. Model of Evolution: A statistical model is chosen to describe how
the sequences are expected to have evolved. For example, a model
might account for the different rates of transition and transversion
mutations in DNA.
4. Tree Building (Phylogenetic Inference): A computational
method is used to build the tree based on the alignment and the
evolutionary model. The most common methods are:
o Distance-based methods (e.g., Neighbor-Joining): These
methods calculate a pairwise "evolutionary distance" for all
sequences and then use this distance matrix to build the tree.
They are computationally fast but can be less accurate.
o Character-based methods (e.g., Maximum Parsimony,
Maximum Likelihood, Bayesian Inference): These
methods evaluate the tree
 based on the actual character states (e.g., A, T, C, G) at each
position in the alignment.
▪ Maximum Parsimony: Aims to find the tree that
requires the fewest evolutionary changes to explain the
data.
▪ Maximum Likelihood and Bayesian Inference:
These are model-based methods that use statistical
probabilities to find the most likely tree given the data
and the evolutionary model. They are the most accurate
but also the most computationally intensive.
5. Tree Visualization and Interpretation: The final tree is
visualized and interpreted to understand the evolutionary
relationships. Branch lengths often represent the amount of
evolutionary change, and branching points (nodes) represent
inferred common ancestors.

Structure Prediction

Structure prediction is the process of predicting the three-dimensional

(3D) structure of a protein from its amino acid sequence. The "protein
folding problem" is a major
challenge in molecular biology because a protein's function is intimately
linked to its 3D shape. A protein sequence folds into a specific shape to carry
out its biological role.

Computational Approaches

Predicting protein structure is one of the most computationally

challenging problems in bioinformatics. The main methods are:

1. Homology Modeling (Comparative Modeling): This is the most

reliable and widely used method. It's based on the principle that if
two proteins share a significant degree of sequence similarity
(typically >30%), they are likely to have a very similar 3D structure.
The steps are:
o Find a Template: Find a known protein structure (the
template) in a database like the Protein Data Bank (PDB) that
is homologous to the target sequence.
o Sequence Alignment: Align the target sequence to the
template sequence.
o Model Building: Use the alignment to build a 3D model of the
target protein, copying the backbone structure from the
template and adjusting the side chains and loop regions.
2. Threading (Fold Recognition): This method is used when no
homologous template with a known structure can be found. It tries
to "thread" or fit the target sequence into a library of known protein
folds. It assesses how well the sequence fits into each fold by
calculating a score that considers factors like residue interactions
and solvent accessibility. If the sequence fits a known fold well, it
suggests that the protein adopts that structure.
3. Ab Initio (De Novo) Modeling: This is the most difficult method
and is used when neither a homologous template nor a suitable fold
can be found. Ab initio methods attempt to build a protein structure
from scratch, based on the physical and chemical principles of
protein folding. They rely on complex algorithms to search for the
most energetically favorable conformation of the amino acid chain.
This method is computationally expensive and is typically only
feasible for very small proteins.

Impact of Structure Prediction

Accurate structure prediction is revolutionizing biology and medicine. It

can be used to:

• Predict protein function based on its shape.

• Design new drugs by targeting the active site of a protein.
• Understand diseases caused by misfolded proteins (e.g.,
Alzheimer's, Parkinson's).
• Engineer novel enzymes with improved catalytic activity.
Question Bank:

Part A: 2-Mark Questions

1. What are the three main types of biological sequences studied in

bioinformatics?
2. What is the primary purpose of a pairwise sequence alignment?
3. Define a substitution matrix.
4. What does the acronym BLAST stand for?
5. What is the key difference between global and local alignment?
6. What does PAM stand for, and what is its primary use?
7. What is an indel?
8. Define a gap penalty.
9. What is the purpose of Multiple Sequence Alignment (MSA)?
10. What does the E-value in a BLAST search result represent?
11. How does the FASTA algorithm differ from BLAST in its initial
search step?
12. What is the role of the ClustalW tool?
13. What is the purpose of a phylogenetic tree?
14. What are homologous sequences?
15. What information does a BLOSUM matrix provide?
16. How is a BLOSUM matrix created?
17. Name one application of sequence alignment.
18. What is the difference between a DNA and an RNA sequence?
19. What is a query sequence in a BLAST search?
20. What is a target database in a BLAST search?
21. What is the main principle behind gene finding?
22. Define homology modeling in the context of structure
prediction.
23. Why are scoring matrices essential for sequence alignment?
24. What is a "seed" in the FASTA and BLAST algorithms?
25. What is the key output of a ClustalW alignment?

Part B: 5-Mark Questions

26. Explain the difference between global alignment and local

alignment, and provide a scenario where each would be more
appropriate.
27. Describe the underlying principle of a scoring matrix, using
either a PAM or BLOSUM matrix as an example.
28. Explain the steps of a typical BLAST search and the
information you can glean from its output.
29. Describe how the FASTA algorithm works, highlighting its key
steps.
30. Explain the role of Multiple Sequence Alignment (MSA) in
identifying conserved regions and inferring evolutionary
relationships.
 31. Differentiate between the PAM and BLOSUM scoring matrices,
explaining how they are constructed and their appropriate uses.
32. Discuss how gene finding is performed using bioinformatics
tools and what types of sequence features are analyzed.
33. Describe how sequence alignment can be used to construct a
phylogenetic tree to show evolutionary relationships.
34. Explain the concept of homology modeling for protein
structure prediction, and what information is required to perform it.
35. How do gap penalties affect the outcome of a sequence
alignment, and why are they necessary?
36. Describe the key differences between DNA, RNA, and Protein
sequences in terms of their alphabet and functional diversity.
37. Explain the function of ClustalW and how it produces a
multiple sequence alignment.
38. What is the significance of the E-value and bit score in a
BLAST result?
39. Discuss the limitations of pairwise sequence alignment for
analyzing a large family of proteins.
40. Explain how sequence similarity can be used to infer the
function of an unknown protein.

Part C: 10-Mark Questions

41. Describe in detail the concepts of pairwise and Multiple

Sequence Alignment (MSA). Explain their differences, the algorithms
used, and their respective applications in bioinformatics.
42. Explain the construction and application of both PAM and
BLOSUM scoring matrices. Discuss why different matrices are used
for different evolutionary distances.
43. Describe the functionality of the BLAST tool in detail. Explain its
search algorithm, the significance of its output parameters, and its
various applications.
44. Discuss how bioinformatics tools are used for gene finding.
Explain the different types of evidence (e.g., codon usage,
regulatory signals) that are used and the challenges involved.
45. Explain how a Multiple Sequence Alignment (MSA) can be
used to reconstruct a phylogenetic tree. Describe the underlying
assumptions and the information gained from such an analysis.
46. Describe the process of protein structure prediction through
homology modeling. Explain the steps involved and the limitations
of this method.
47. Compare and contrast the FASTA and BLAST search
algorithms. Discuss their historical context, their underlying
principles, and their advantages and disadvantages.
48. Discuss the role of scoring matrices and gap penalties in
optimizing a sequence alignment. Explain the different types of gap
penalties and their impact on the alignment result.
49. Elaborate on the different types of biological sequences (DNA,
RNA, Protein). Discuss how their unique properties (alphabet size,
structure, function) influence the choice of bioinformatics tools and
alignment strategies.
50. Explain the concept of sequence homology and its importance
in bioinformatics. Describe how both a pairwise and a multiple
sequence alignment can be used to infer homology and functional
relationships.