BS982

Welcome to BLAST
Basic Local Alignment Search Tool
Ben Skinner

School of Life Sciences • University of Essex

[email protected]
 Outline

• Why BLAST?

• How BLAST works

• Running BLAST for yourself

 Why BLAST?

• Falling cost of sequencing

• Growing size of databases - 1979 Los Alamos Sequence Database (became GenBank)

https://www.ncbi.nlm.nih.gov/genbank/statistics/
 Why BLAST?

• Imagine – you have sequenced something from an environmental sample

• Has anyone seen this before? Is it new?

Extract DNA

Your Sequence

DNA:
ATGTGATCCGACTATGACA….

PROTEIN:
MRPILMKGHERPLTFLRYNRDG….

https://www.universe-review.ca/I11-50-DNAsequencing1.jpg
 Why BLAST?

• Or you may have a protein/DNA sequence from a database:

NCBI/EMBL/SwissProt/UniProt
• What else is it similar to?

Your Sequence

DNA:
ATGTGATCCGACTATGACA….

PROTEIN:
MRPILMKGHERPLTFLRYNRDG….
 Why BLAST?

BLAST is a tool that helps us find similar

sequences to a query sequence in a
database

BLAST can search for matches to DNA or

protein queries
 Why BLAST?

• Usually the first thing you do when you obtain a

new, unknown sequence
• Provides biological context and helps establish
hypotheses to be tested (by experimentation or
further bioinformatic analysis)
• Part of our lexicon: “To Blast a sequence” as in
“To Google a question”
• Emphasizes speed over sensitivity
• Quick and simple to use, but very often used
sub-optimally
 BLAST can be used for different purposes

• Looking for species. If you are sequencing DNA from unknown

species - identify the correct species or homologous species.
• Looking for domains. If you BLAST a protein sequence (or a
translated nucleotide sequence) – identify known domains
• Looking at phylogeny. You can use the BLAST web pages to
generate a phylogenetic tree of the BLAST result.
• Mapping DNA to a known chromosome. If you are sequencing a
gene from a known species but have no idea of the chromosome
location
• Annotations. BLAST can also be used to map annotations from
one organism to another or look for common genes in two
related species.
 Important concepts

• Similarity: Degree of likeness between two sequences, usually

expressed as a percentage of similar (or identical) residues over
a given length of the alignment. Can usually be easily calculated.
• Homology: Statement about common evolutionary ancestry of
two sequences – hypotheses
• A high degree of similarity implies a high probability of homology
Homologues: genes with a common ancestor

Remember the Hox genes from

Eukaryotic genomes lecture?
 Conservation of sequence implies conservation of function

Key to finding important regions –

functional protein domains

https://ars.els-cdn.com/content/image/1-s2.0-S2001037015000070-gr3.jpg
Searching for similarity and homology
 Outline

• Why BLAST?

• How BLAST works

• Running BLAST for yourself

 Searching for similarity

• Important goal of genomics is to determine if a particular sequence is “like”

another sequence
• Compare new sequences with sequences already stored in a database
• Two alignment types: global and local
 Global alignment

• Compares one whole sequence with another entire sequence (end to

end alignment)
• Suitable for aligning closely related species, for example comparing
two genes with the same function in humans and mouse

See lecture on sequence alignment for details

Needleman-Wunsch algorithm
 Local alignment

• Uses a subset of a sequence to align a subset of other sequences

• Reveals regions that are highly similar, but do not necessarily
provide comparison across the entire two sequences
• Find conserved patterns in DNA sequences or conserved domains in
two proteins
• May uncover regions of homology that are related by descent
between otherwise diverse sequences

Smith-Waterman algorithm
 How does BLAST work? Basic overview

The original BLAST program (Altschul et al 1990 J Mol Biol 215:403 )

• Sequence query is broken into words of length W
• Align all words with sequences in the database
• Calculate a score T for each word that aligns with a sequence in
the database using a substitution matrix
• Discard words whose T value is below a neighbourhood score
threshold
• Extend words in both directions until score drops below the
previous best score
 Using BLAST: the different tools and databases

Query tool to retrieve

homologous genes from a
database BLAST
Sequence
Database
(Target)

Putative Function

Kerfeld and Scott, PLoS Biology 2011 9(2): e1001014.

Using BLAST: https://blast.ncbi.nlm.nih.gov/Blast.cgi
 Outline

• Why BLAST?

• How BLAST works (basics)

• Running BLAST
 Basic search strategy

• (1) Choose your sequence (query)

• (2) Choose the BLAST program
• (3) Choose the database to search
• (4) Choose optional parameters

• Then click “BLAST”

 Getting a query sequence

• We start by finding our query sequence in FASTA format

• e.g. from Genbank, Ensembl, or a paper

Amino acid sequence of a protein, in FASTA format:

>ribosomal protein L7/L12 [Thiomicrospira crunogena XCL-2]
MAITKDDILEAVANMSVMEVVELVEAMEEKFGVSAAAVAVAGPAGDAGAAGEEQTEFDVVLTGAGDNKVAAIKAVRGATG
LGLKEAKSAVESAPFTLKEGVSKEEAETLANELKEAGIEVEVK

Nucleotide sequence of a gene, in FASTA format:

>gi|118139508:333094-333465 Thiomicrospira crunogena XCL-2
ATGGCAATTACAAAAGACGATATTTTAGAAGCAGTTGCTAACATGTCAGTAATGGAAGTTGTTGAACTTGTTGAAGCAAT
GGAAGAGAAGTTTGGTGTTTCTGCAGCAGCAGTTGCGGTTGCAGGTCCTGCAGGTGATGCTGGCGCTGCTGGTGAAGAAC
AAACAGAGTTTGACGTTGTCTTGACTGGTGCTGGTGACAACAAAGTTGCAGCAATCAAAGCCGTTCGTGGCGCAACTGGT
CTTGGGCTTAAAGAAGCGAAAAGTGCAGTTGAAAGTGCACCATTTACGCTTAAAGAGGGTGTTTCTAAAGAAGAAGCAGA
AACTCTTGCAAATGAGCTTAAAGAAGCAGGTATTGAAGTCGAAGTTAAATAA
 What is the FASTA format?

• 1985: FASTP program developed for fast alignment of protein sequences, FASTN for
fast alignment of nucleotide sequences

• 1988: FASTA program for fast alignment of all sequence types - nucleotide or protein
https://fasta.bioch.virginia.edu/fasta_www2/fasta_list2.shtml

• Had a format for input query sequences

• FASTA the program is rarely used now – replaced by BLAST – but the FASTA format lives
on

”description line” (not read as sequence data) > ribosomal proteinL7/L12

• Begins with >
MAITKDDILEAVANMSVMEVVELVEA
• Ends with a new line MEEKFGVSAAAVAVAGPAGDAGAA
GEEQTEFDVVLTGAGDNKVAAIKAVR
Sequence data GATGLGLKEAKSAVESAPFTLKEG
(amino acid in this case) VSKEEAETLANELKEAGIEVEVK
 What is the FASTA format?

• FASTA format files are one of the most common and universally used input files in
bioinformatics and sequence analysis

• Most bioinformatics software expect this format

• For more advanced users there is freely available code to input/output this format in
bioinformatic pipelines, e.g. BioPython and many R packages.

”description line” (not read as sequence data) > ribosomal proteinL7/L12

• Begins with >
MAITKDDILEAVANMSVMEVVELVEA
• Ends with a new line MEEKFGVSAAAVAVAGPAGDAGAA
GEEQTEFDVVLTGAGDNKVAAIKAVR
Sequence data GATGLGLKEAKSAVESAPFTLKEG
(amino acid in this case) VSKEEAETLANELKEAGIEVEVK
 Choose the BLAST program to use

• What type of alignment do you

want?

Basic

• blastn – nucleotide – nucleotide

database

• blastp – protein – protein

database

More complex

• blastx – translated nucleotide –

protein database

• tblastn – protein – translated

nucleotide database

• tblastx – translated nucleotide – https://blast.ncbi.nlm.nih.gov/Blast.cgi

translated nucleotide database
 NCBI BLAST interface (blastp: proteins)

Paste FASTA format

sequence here
Paste FASTA sequence here
NCBI BLAST interface (blastp: proteins)
 Other input options

• When using NCBI’s BLAST, you can also use a GenBank

accession number instead of the FASTA sequence
• e.g. the accession for that ribosomal sequence is
WP_011369709.1
 Choose the database to search

• nr/nt = non-redundant protein and nucleotide databases (most general

database)
• Human G+T – genomic and transcript database for humans
• dbest = database of expressed sequence tags
• gss = genomic survey sequences

Protein databases

Nucleotide databases
BLASTn page
• Default = human
• Next = mouse
• Then = others
• Default others = nr/nt (includes
many of those below, hence nr
= “non-redundant”.
• Many more specific databases
that allow to focus search
BLASTn page
• Search can be focussed on a
particular organism
• Model sequences and those
from uncultured organisms can
be excluded
• Entrez allows key words to be
used to restrict the search, e.g.
“olfactory receptor”
 Choose optional parameters
• To understand optimal parameters you need to look at the output first!

Graphic
display

Descriptions
display

Alignment
display

Kerfeld and Scott, PLoS Biology 2011 9(2): e1001014.

 BLAST results page: potential homologues identified

S‘ (Bit score) a E (expect value) a

measure of overall statistical
sequence similarity measure
Kerfeld and Scott, PLoS Biology 2011 9(2): e1001014.
 How to do BLAST wrong: believing the E-value tells the
whole story

Does it cover the whole length of both the query and subject sequences?

Discovery of a Distant Homolog or

Garbage?

Kerfeld and Scott, PLoS Biology 2011

 Is your result biologically meaningful?

Important considerations
• Understand the output
• Adjust the input
• Treat the analysis like an experiment (not like a google search)
Optional parameters:
Click here to see the list of algorithm
parameters that can be changed
 Summary

• BLAST is not one tool – it is a suite of tools

• Can quickly search for nucleotide or protein sequences – or more

complex queries

• It is not a black box – you should understand how BLAST works

• Tuning the input parameters may be needed to find sequences of

interest

Next time: more complex BLAST & choosing those

optional parameters to improve your search