PAIRWISE ALIGNMENT USING BLAST

Dr. Gopa Mitra

1
 The Basic Local Alignment Search Tool (BLAST) finds regions of
local similarity between protein or nucleotide sequences.
 The program compares nucleotide or protein sequences to
sequence in a database and calculates the statistical significance of
the matches.
 This chapter first provides an introduction to BLAST and then
describes the practical application of different BLAST programs
based on the BLAST Quick Start mini-course
([Link]/Class/minicourses).
 In each example, emphasis is placed on practical step-by-step
procedures, although relevant theory is also given where it affects
the choice of BLAST program, parameters, and database.

2
  INTRODUCTION
 BLAST is an acronym for Basic Local Alignment Search Tool and
refers to a suite of programs used to generate alignments between
a nucleotide or protein sequence, referred to as a “query” and
nucleotide or protein sequences within a database, referred to as
“subject” sequences.
 The original BLAST program used a protein “query” sequence to
scan a protein sequence database.
 A version operating on nucleotide query” sequences and a
nucleotide sequence database soon followed.
 The introduction of an intermediate layer in which nucleotide
sequences are translated into their corresponding protein
sequences according to a specified genetic code allows cross-
comparisons between nucleotide and protein sequences.
 Specialized variants of BLAST allow fast searches of nucleotide
databases with very large query sequences, or the generation of
alignments between a single pair of sequences.
 Both the standalone and web version of BLAST are available from
the National Center for Biotechnology Information
([Link]).

3
 The web version provides searches of the complete genomes
of Homo sapiens as well as those of many model organisms,
including mouse, rat, fruit fly, and Arabidopsis thaliana, allowing
BLAST alignments to be seen in a full genomic context.
 Query and Database Sequence (Subject sequence)Formats
 BLAST “query” sequences are given as character strings of single
letter nucleotide or amino acid codes, preceded by a definition
line, beginning with a “>” symbol and containing identifiers and
descriptive information.
 This format is known as FASTA.
 BLAST databases are constructed from concatenated FASTA
formatted sequences using a program called “formatdb” that
produces a mixture of binary- and ascii-encoded files containing
the sequences and indexing information used during the BLAST
search.

4
BLAST PROGRAMS

5
  Scoring of Alignments and Substitution Matrices
 A BLAST alignment consists of a pair of sequences, in which every
letter in one sequence is paired with, or “aligned to,” exactly one
letter or a gap in the other.
 The alignment score is computed by assigning a value to each
aligned pair of letters and then summing these values over the
length of the alignment.
 For protein sequence alignments, scores for every possible amino
acid letter pair are given in a “substitution matrix” where likely
substitutions have positive values and unlikely substitutions have
negative values.
 By default, BLAST uses the “blosum62” matrix, a member of the
most commonly used series of substitution matrix, however,
several members of the PAM series are also available.
 For nucleotide alignments, BLAST uses a reward of +2 for aligned
pairs of identical letters and a penalty of −3 for each nonidentical
aligned pair.
 The creation of a gap in an alignment results in a negative “gap-
creation” penalty, with each extension of a preexisting gap
incurring a lesser penalty. For a detailed treatment of the theory
of alignment scoring .

6
  OVERVIEW OF THE ALGORITHM
 BLAST begins a search by indexing all character strings of a certain
length within the “query” by their starting position in the query.
 The length of the string to index, called the “wordsize” is configurable by
the user.
 The allowable range for the “wordsize” varies according to the BLAST
program used; typical values are 3 for protein-to-protein sequence
searches and 11 for nucleotide to nucleotide searches.
 BLAST then scans the database looking for matches between the “words”
indexed in the “query” and strings found within the database sequences.
 For nucleotide-to-nucleotide searches, these matches must be exact; for
protein-to-protein searches, the score of the match as determined using
a substitution matrix, must exceed a specified threshold.
 When a word match is found, two nearby words in the case of protein
searches, BLAST attempts to extend both forward and backward from the
match to produce an alignment.
 BLAST will continue this extension as long as the alignment score
continues to increase or until it drops by a critical amount owing to the
negative scores given by mismatches.
 This critical amount is known as the “dropoff.”

7
  Statistical Significance
 The alignments found by BLAST during a search are scored, as
previously described, and assigned a statistical value, called the
“Expect Value.”
 The “Expect Value” is the number of times that an alignment as
good or better than that found by BLAST would be expected to
occur by chance, given the size of the database searched.
 An “Expect Value” threshold, set by the user, determines which
alignments will be reported.
 A higher “Expect Value” threshold is less stringent and the BLAST
default of “10” is designed to ensure that no biologically
significant alignment is missed.
 However, “Expect Values” in the range of 0.001 to 0.0000001 are
commonly used to restrict the alignments shown to those of high
quality.
 Lower = more biologically meaningful
 Available as an online web server ,
 at NCBI (hGp://[Link]/[Link])
8
 E Value How many random alignments justas good?
 1 1 in 1
 .2 1 in 5
 1e-5 1 in 100,000
 1e-9 1 in 1,000,000,000
 0 0%

9
10