BLAST: Basic Local

Alignment Search
Tool
Understanding Sequence
Alignment and Search

Saicharan, Kartik
 BL
What exactly is
A ST
BLAST?

Sequence

Sequence
1

*Compared*
BLAST from the
past :

Inventor of a different Stephen Altschul,

kind of a Blast Creator of BLAST
No but seriously, What is BLAST?
No but seriously, What is
BLAST?
• Query = New sequence
• Target = Set of many old sequences
• To find : Which (if any) of the target sequences is the query related to ?
No but seriously, What is
BLAST?
• Query = New sequence
• Target = Set of many old sequences
• To find : Which (if any) of the target sequences is the query related to ?

How does it do it?: The BLAST

algorithm
How does it do it?: The BLAST
algorithm
Imagine you walk into a huge library looking for a book (your query sequence).
Instead of checking every book page by page (slow), you:
1. Look for keywords in the catalog
(seeding – breaking your query into small words).
2. Find books that contain those keywords
(word matches in the database).
3. Skim through relevant books to check if they truly match your topic
(extension and scoring).
4. Rank the best books based on how well they match your search
(scoring & E-value).
5. Read the most relevant books first
(final BLAST results).
 How does it do it?: The BLAST
algorithm
• STEP 1 : Seeding the search
Before BLAST can compare your sequence to
a database, it needs to break it into smaller
pieces.

• Example: If your sequence is ATGCGTACG,

and the word size is 3, BLAST will create:
• ATG, TGC, GCG, CGT, GTA, TAC, ACG

This word size (k-mer length) is important:

• Smaller words → More sensitive, finds weak
matches, slower
• Larger words → Faster, but might miss distant
similarities.
Now, BLAST compares these small words against its massive sequence database to find exact
matches—just like searching for keywords in a book.
How does it do it?: The BLAST
algorithm
• STEP 2 : Extending Matches
Once BLAST finds an exact match for one of your words, it extends the alignment in
both directions.

For DNA or protein sequences, BLAST extends the match until:

• The similarity drops below a threshold.
• There are too many mismatches or gaps.

This step finds High-Scoring Segment Pairs (HSPs)—regions where sequences are
highly similar.
How does it do it?: The BLAST
algorithm
• STEP 3 : Scoring Matches
Now that BLAST has found similar regions, it needs to decide which ones are
important.
It does this using two main scores:

1. Bit Score – Measures the quality of the alignment.

• Higher score = Better match.
• Normalized, so it can be compared across different searches.
2. E-value (Expectation Value) – Measures statistical significance.
• Lower E-value = Less likely to be a random match.
• Example: An E-value of 1e-10 means there’s only a 1 in 10 billion chance the
match is random.
How does it do it?: The BLAST
algorithm
• STEP 4 : Filtering Out the Noise – Refining Results
BLAST doesn’t just return every match—it filters out weak and irrelevant
alignments using:

-Substitution Matrices (BLOSUM, PAM)

• These help BLAST decide how much to penalize mismatches.
• Example: BLOSUM62 is commonly used for proteins, rewarding similar amino
acid replacements.

-Gap Penalties
• Small gaps (insertions/deletions) happen naturally in evolution.
• BLAST penalizes too many gaps to keep results biologically meaningful.

These adjustments fine-tune the results, ensuring only strong evolutionary

relationships are detected.
Types of BLASTs :
How to have a BLAST 🥳 :
• Select the type of blast you want to run
How to have a BLAST 🥳 :
• Paste your query sequence(s) in the provided box or upload a FASTA file.

• A : Query Sequence Box: Paste query sequence(s) here in FASTA format.

• B : Choose File: Use this option to upload FASTA file with query sequence
How to have a BLAST 🥳 :
• Choose a program for search optimization and click the BLAST button to run
search.
Fine-Tuning BLAST
BLAST results can be customized using
different parameters to improve accuracy and
relevance:
• E-value – Significance threshold
(lower = better match).
• Bit Score – Measures alignment strength
(higher = better).
• Substitution Matrices – (BLOSUM, PAM)
Used for scoring matches/mismatches.
• Gap Penalties – Penalizes insertions and
deletions.
• Word Size (k-mer) – Adjusts sensitivity vs.
speed.
• Database Selection – Choose specific
databases for focused searches.
 Interpretation of BLAST result
page

Edit Search goes back to the search page to

allow
rerun of the search with adjusted settings.
Save Search saves the search setting used for
later use.
Search Summary A table with summary
statistics and search settings.
Summary: Provides key search details.
Job Title: Defaults to the first query’s
sequence ID
RID: Unique search identifier for
sharing/reporting.
Download All: Options to save results in
XML2, JSON, or CSV.
Program: Specifies BLAST program
(e.g., blastp).
Citation: Provides reference for publications
Database: Target database searched.
See Details: Text description of database
content.
Query ID: Sequence ID, linked if an accession
number is used.
Description: Title from FASTA defline.
Molecule Type: Type of query sequence.
Query Length: Length of the query sequence.
Other Reports: Links to additional report
formats.
 Organism: Filter hits by organism name
exclude option available.
Add Organism: Add more organisms for filtering.
Percent Identity: Filter hits by identity percentage
(e.g., 94.74–94.76%).
E Value: Filter by Expect value significance (e.g., 0.0001–5e-
120)
Query Coverage
Filter
Reset : Restore initial results.

Percent Identity: The percentage of identical matches between the query and subject sequences.
Example:
Query: ATGCTAGGCTA
Subject: ATGCTCGGCTA
Out of 10 positions, 9 are identical. The Percent Identity is:
(9/10)×100=90%
90% of the sequence is an exact match.
Description section
Graphic Summary
Alignment
section Matched Database Sequence:
Shows the database sequence that
matches the query.
Alignment Display:
Mini-tab: Contains alignment statistics.
Middle Line:
Identical match = letter.
Positive match = + (conservative
substitution).
Zero/negative match = space.
Gaps = -.
Related Information: Links to
additional
resources for matched database sequence.
Taxonomy Section
• Lineage Report: Summarizes
BLAST
hits by organism taxonomy.
• Organism Column: Displays
taxonomic
levels with indentation (dots indicate levels)
• Blast Name Column: Lists group/
organism aliases; links to Taxonomy
database.
• Score Column: Shows highest bit
score
for hits from that group.
• Number of Hits Column: Total hits
for
the group; links to Protein database.
• Description Column: Anchored to
detailed hits in the Organism Report.
• Organism Report: Best for
examining
matches for a specific species.
• Subsections: Organized by source
organisms; linked from the Number of Hits
column in other reports.
• Sequence Titles: Lists hit sequence
titles.
• Alignment Quality: Displays Score and
Expect value.
• Hit Titles: Link to the corresponding
alignment in the traditional BLAST results.
• Accessions: Linked to the sequence
record in the source database.
• Taxonomy Report: Similar to Lineage
Report
but focuses on specific organism groups.
• Lineage Display.
• Taxonomy Column: Indicates taxonomic
level
with indentation and dots for quick reference.
• Organism Name Links: Links to Taxonomy
database for additional details.
• Number of Hits Column: Sums hits for
each
group; links to Protein or Nucleotide database.
• Description Column: Links to
Organism Report with details on match quality.
 Advanced BLAST fetures
PSI-BLAST (Position-Specific
Iterated BLAST)
• Used to identify distant homologs.
• First Iteration: Identical to a BLASTp
search.
• Multiple Alignment: Formed from the
highest-scoring sequence pairs above
a preset score (e-value) threshold.
• PSSM (Position-Specific
Scoring Matrix) is calculated from
this multiple alignment.
• PSSM captures conservation patterns
in an alignment with scores for each
position.
• Highly conserved positions get high
scores; weakly conserved ones score
near zero.
• The PSSM replaces the standard
substitution matrix to find more
sequences matching the conservation
pattern.
• Iterative Refinement:
Newly detected sequences above the score threshold are added to the alignment.
The profile is refined and used for another search.
The process repeats until no new sequences meet the threshold (convergence).

• Advantage of PSI-BLAST:
More sensitive than BLASTp for detecting distant sequence similarities.
Uses conservation information from multiple sequences in a single score matrix.
• PHI-BLAST (Pattern Hit Initiated BLAST)
• Combines pattern matching with sequence alignment
to detect homologs sharing specific motifs.

• It performs two types of searches: sequence similarity

using BLAST and motif matching.

• Use when you know protein family has a signature pattern:

active site, structural domain, etc

• PHI-BLAST uses the Prosite pattern

convention:
Any valid residue one-character symbol
• ACDEFGHIKLMNPQRSTVWY (for DNA: GATC)
• •[ ] means any one of the characters in brackets e.g.,
[LFYT]
• means one occurrence of L or F or Y or T
• - means nothing (this is a spacer for human readability)
• x(5) means 5 positions in which any residue is allowed
• x(2,4) means 2 to 4 positions where any residue is allowed
• • [LIVMF]-G-E-x-[GAS]-x(5,11)-R-[STAQ]-A-x-[LIVMA]-x-
[STACV]
DELTA-BLAST (Domain Enhanced Lookup Time Accelerated BLAST)
• Finds protein homologs using conserved domain information.
• Uses CDD (Conserved Domain Database) to improve search accuracy.
• If a homologous protein has low similarity at the sequence level but shares a functional domain, normal BLAST
might miss it.
• Iterative Rounds
• RPS-BLAST (Reverse Position-Specific BLAST):
• Identifies known conserved domains in a protein sequence.
• Uses CDD but does not refine the search iteratively.
• Used mainly for domain annotation and functional analysis.
• Focuses on matching existing domains, rather than discovering new
relationships.
• Searches a database of pre-calculated Position-Specific Scoring Matrices (PSSMs).
• The query sequence is compared against these PSSMs (instead of building a PSSM from the query
sequence, as in PSI-BLAST).
• Displays hits with varying confidence levels, which indicate how strongly the query matches specific
domain profiles.
• MEGA BLAST
• A variant of BLAST designed for highly similar
sequences (e.g., same species or closely related
species) to perform faster searches with higher
sensitivity.
• It aligns long nucleotide sequences efficiently
(e.g., whole-genome sequences).
• Has a default nucleotide size of 28 in search
compared to 11 nucleotides of BLASTN

• Discontiguous MEGA-BLAST
(dMegaBLAST)
• Designed to work with sequences that are less
similar
• Allows for gaps and mismatches within a search,
making it suitable for finding sequences that may
have insertions, deletions, or mutations.
• Better suited for queries with low-complexity or
repetitive regions, where traditional alignment
methods might struggle.
• Primer-BLAST
Used for designing primers for PCR (Polymerase Chain Reaction) experiments.
It helps design primers that match a target sequence and avoid non-specific binding to other
sequences.
• SMART-BLAST:
• Fast and efficient
protein similarity
search.
• Uses well-
annotated
reference proteins
for comparison.
• Combines BLASTP
and HMM-based
methods for
accuracy.
• Provides top hits and
a phylogenetic
tree-like
visualization.
• Best for functional
annotation of
proteins rather than
distant homology
detection
 Applications
• Identifies similar nucleotide sequences in databases.
Used for finding homologous genes, species identification, and evolutionary studies.
Assists in SNP detection by aligning sequences and pinpointing variations at specific nucleotide
positions across individuals or populations.
• Compares protein sequences to find homologous proteins.
Helps in functional annotation, structure prediction, and protein family classification.
• Translates a nucleotide sequence into protein and compares it with protein
databases.
Useful for gene identification, especially in newly sequenced genomes.
• Compares a protein query with translated nucleotide sequences.
Helps in identifying genes in unannotated genomes.
• Compares translated nucleotide sequences against other translated
sequences.
Used for finding distant homologs and studying evolutionary relationships.
• Uses iterative searching to detect distant protein homologs.
Helps in protein family classification and structural prediction.
• Searches for homologous protein domains using conserved domain
databases.
Improves protein functional annotation and evolutionary studies.
• Designs primers for PCR by checking specificity against genomes.
Useful in molecular biology experiments like gene amplification
• Matches a protein sequence against conserved domain profiles.
Helps in identifying protein domains and functional annotations.
• Searches for protein sequences with specific motifs or patterns.
Useful in functional domain analysis and identifying conserved motifs.
 Limitations of BLAST
• Heuristic Sensitivity Trade-Off – BLAST prioritizes speed over exhaustive search, potentially
missing weak but true homologs.
• Inappropriate Default Word Size – Default word lengths may be too long for short queries,
leading to missed matches (BLASTp - 3 amino acids, BLASTn -11 nucleotides.
• Alignment Focus – Emphasizes local matches, potentially missing broader global sequence
similarities.
• Parameter Sensitivity – Accuracy depends on parameter tuning; defaults may not suit all
sequences.
Eg: If the gap penalty is too high, BLAST may fail to align two sequences that match except for a small
missing or extra piece.
• Database Size Dependence – E-value significance changes with database size, affecting result
interpretation.
• Low-Complexity Filtering – Important functional regions may be masked, leading to missed
biological insights.
Eg: Removing repetitive or simple sequences, like long stretches of 'A's or 'G's.
References:
• [Link]
• [Link]
• [Link]
• [Link]
• Bhagwat M, Aravind L. PSI-BLAST Tutorial. In: Bergman NH, editor. Comparative Genomics: Volumes 1
and 2. Totowa (NJ): Humana Press; 2007. Chapter
• Boratyn, G.M., Schäffer, A.A., Agarwala, R. et al. Domain enhanced lookup time accelerated BLAST. Biol
Direct 7, 12 (2012). [Link]
• [Link]
• Ye, J., Coulouris, G., Zaretskaya, I. et al. Primer-BLAST: A tool to design target-specific primers for
polymerase chain reaction. BMC Bioinformatics 13, 134 (2012). [Link]
• [Link]
• [Link]
_Genomes_Networks_and_Evolution_(Kellis_et_al.)/
03%3A_Rapid_Sequence_Alignment_and_Database_Search/
3.05%3A_The_BLAST_algorithm_(Basic_Local_Alignment_Search_Tool)
• [Link]
THANK YOU
:)