Parse and read blast files
use Boulder::Blast;
# parse from a single file
$blast = Boulder::Blast->parse('run3.blast');
# parse and read a set of blast output files
$stream = Boulder::Blast->new('run3.blast','run4.blast');
while ($blast = $stream->get) {
# do something with $blast object
}
# parse and read a whole directory of blast runs
$stream = Boulder::Blast->new(<*.blast>);
while ($blast = $stream->get) {
# do something with $blast object
}
# parse and read from STDIN
$stream = Boulder::Blast->new;
while ($blast = $stream->get) {
# do something with $blast object
}
# parse and read as a filehandle
$stream = Boulder::Blast->newFh(<*.blast>);
while ($blast = <$stream>) {
# do something with $blast object
}
# once you have a $blast object, you can get info about it:
$query = $blast->Blast_query;
@hits = $blast->Blast_hits;
foreach $hit (@hits) {
$hit_sequence = $hit->Name; # get the ID
$significance = $hit->Signif; # get the significance
@hsps = $hit->Hsps; # list of HSPs
foreach $hsp (@hsps) {
$query = $hsp->Query; # query sequence
$subject = $hsp->Subject; # subject sequence
$signif = $hsp->Signif; # significance of HSP
}
}
The Boulder::Blast class parses the output of the Washington University (\s-1WU\s0) or National Cenber for Biotechnology Information (\s-1NCBI\s0) series of \s-1BLAST\s0 programs and turns them into Stone records. You may then use the standard Stone access methods to retrieve information about the \s-1BLAST\s0 run, or add the information to a Boulder stream.
The parser works equally well on the contents of a static file, or on information read dynamically from a filehandle or pipe.
$stone = Boulder::Blast->parse($file_path); $stone = Boulder::Blast->parse($filehandle);
The parse() method accepts a path to a file or a filehandle, parses its contents, and returns a Boulder Stone object. The file path may be absolute or relative to the current directgly. The filehandle may be specified as an IO::File object, a FileHandle object, or a reference to a glob (\*(C`\*FILEHANDLE\*(C' notation). If you call parse() without any arguments, it will try to parse the contents of standard input.
$stream = Boulder::Blast->new; $stream = Boulder::Blast->new($file [,@more_files]); $stream = Boulder::Blast->new(\*FILEHANDLE);
If you wish, you may create the parser first with Boulder::Blast new(), and then invoke the parser object's parse() method as many times as you wish to, producing a Stone object each time.
The following tags are defined in the parsed Blast Stone object:
These top-level tags provide information about the version of the \s-1BLAST\s0 program itself.
The name of the algorithm used to run the analysis. Possible values include: blastn blastp blastx tblastn tblastx fasta3 fastx3 fasty3 tfasta3 tfastx3 tfasty3
This gives the version of the program in whatever form appears on the banner page, e.g. \*(L"2.0a19-WashU\*(R".
This gives the date at which the program was compiled, if and only if it appears on the banner page.
These top-level tags give information about the particular run, such as the parameters that were used for the algorithm.
This gives the date and time at which the similarity analysis was run, in the format \*(L"Fri Jul 6 09:32:36 1998\*(R"
This points to a subrecord containing information about the algorithm's runtime parameters. The following subtags are used. Others may be added in the future: Hspmax the value of the -hspmax argument Expectation the value of E Matrix the matrix in use, e.g. BLOSUM62 Ctxfactor the value of the -ctxfactor argument Gapall The value of the -gapall argument
Thse top-level tags give information about the query sequence and the database that was searched on.
The identifier for the search sequence, as defined by the \s-1FASTA\s0 format. This will be the first set of non-whitespace characters following the \*(L">\*(R" character. In other words, the search sequence \*(L"name\*(R".
The length of the query sequence, in base pairs.
The Unix filesystem path to the subject database.
The title of the subject database.
Each \s-1BLAST\s0 hit is represented by the tag Blast_hits. There may be zero, one, or many such tags. They will be presented in reverse sorted order of significance, i.e. most significant hit first.
Each Blast_hits tag is a Stone subrecord containing the following subtags:
The name/identifier of the sequence that was hit.
The total length of the sequence that was hit
The significance of the hit. If there are multiple HSPs in the hit, this will be the most significant (smallest) value.
The percent identity of the hit. If there are multiple HSPs, this will be the one with the highest percent identity.
The expectation value for the hit. If there are multiple HSPs, this will be the lowest expectation value in the set.
One or more sub-sub-tags, pointing to a nested record containing information about each high-scoring segment pair (\s-1HSP\s0). See the next section for details.
Each Blast_hit tag will have at least one, and possibly several Hsps tags, each one corresponding to a high-scoring segment pair (\s-1HSP\s0). These records contain detailed information about the hit, including the alignments. Tags are as follows:
The significance (P value) of this \s-1HSP\s0.
The number of bits of significance.
Expectation value for this \s-1HSP\s0.
Percent identity.
Percent positive matches.
The Smith-Waterman alignment score.
The word \*(L"plus\*(R" or \*(L"minus\*(R". This tag is only present for nucleotide searches, when the reverse complement match may be present.
Depending on algorithm used, indicates complementarity of match and possibly the reading frame. This is copied out of the blast report. Possibilities include: "Plus / Minus" "Plus / Plus" -- blastn algorithm "+1 / -2" "+2 / -2" -- blastx, tblastx
Position at which the \s-1HSP\s0 starts in the query sequence (1-based indexing).
Position at which the \s-1HSP\s0 stops in the query sequence.
Position at which the \s-1HSP\s0 starts in the subject (target) sequence.
Position at which the \s-1HSP\s0 stops in the subject (target) sequence.
These three tags contain strings which, together, create the gapped alignment of the query sequence with the subject sequence. For example, to print the alignment of the first \s-1HSP\s0 of the first match, you might say: $hsp = $blast->Blast_hits->Hsps; print join("\n",$hsp->Query,$hsp->Alignment,$hsp->Subject),"\n";
See the bottom of this manual page for an example \s-1BLAST\s0 run.
This module has been extensively tested with \s-1WUBLAST\s0, but very little with \s-1NCBI\s0 \s-1BLAST\s0. It probably will not work with \s-1PSI\s0 Blast or other variants.
The author plans to adapt this module to parse other formats, as well as non-BLAST formats such as the output of Fastn.
Boulder, Boulder::GenBank
Lincoln Stein <[email protected]>.
Copyright (c) 1998-1999 Cold Spring Harbor Laboratory
This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself. See \s-1DISCLAIMER\s0.txt for disclaimers of warranty.
This output was generated by the quickblast.pl program, which is located in the eg/ subdirectory of the Boulder distribution directory. It is a typical blastn (nucleotide->nucleotide) run; however long lines (usually \s-1DNA\s0 sequences) have been truncated. Also note that per the Boulder protocol, the percent sign (%) is escaped in the usual way. It will be unescaped when reading the stream back in.
Blast_run_date=Fri Nov 6 14:40:41 1998 Blast_db_date=2:40 PM EST Nov 6, 1998 Blast_parms={ Hspmax=10 Expectation=10 Matrix=+5,-4 Ctxfactor=2.00 } Blast_program_date=05-Feb-1998 Blast_db= /usr/tmp/quickblast18202aaaa Blast_version=2.0a19-WashU Blast_query=BCD207R Blast_db_title= test.fasta Blast_query_length=332 Blast_program=blastn Blast_hits={ Signif=3.5e-74 Expect=3.5e-74, Name=BCD207R Identity=100%25 Length=332 Hsps={ Subject=GTGCTTTCAAACATTGATGGATTCCTCCCCTTGACATATATATATACTTTGGGTTCCCGCAA... Signif=3.5e-74 Length=332 Bits=249.1 Query_start=1 Subject_end=332 Query=GTGCTTTCAAACATTGATGGATTCCTCCCCTTGACATATATATATACTTTGGGTTCCCGCAA... Positives=100%25 Expect=3.5e-74, Identity=100%25 Query_end=332 Orientation=plus Score=1660 Strand=Plus / Plus Subject_start=1 Alignment=||||||||||||||||||||||||||||||||||||||||||||||||||||||||||... } } =
Here is the output from a typical blastp (protein->protein) run. Long lines have again been truncated.
Blast_run_date=Fri Nov 6 14:37:23 1998 Blast_db_date=2:36 PM EST Nov 6, 1998 Blast_parms={ Hspmax=10 Expectation=10 Matrix=BLOSUM62 Ctxfactor=1.00 } Blast_program_date=05-Feb-1998 Blast_db= /usr/tmp/quickblast18141aaaa Blast_version=2.0a19-WashU Blast_query=YAL004W Blast_db_title= elegans.fasta Blast_query_length=216 Blast_program=blastp Blast_hits={ Signif=0.95 Expect=3.0, Name=C28H8.2 Identity=30%25 Length=51 Hsps={ Subject=HMTVEFHVTSQSW---FGFEDHFHMIIR-AVNDENVGWGVRYLSMAF Signif=0.95 Length=46 Bits=15.8 Query_start=100 Subject_end=49 Query=HLTQD-HGGDLFWGKVLGFTLKFNLNLRLTVNIDQLEWEVLHVSLHF Positives=52%25 Expect=3.0, Identity=30%25 Query_end=145 Orientation=plus Score=45 Subject_start=7 Alignment=H+T + H W GF F++ +R VN + + W V ++S+ F } } Blast_hits={ Signif=0.99 Expect=4.7, Name=ZK896.2 Identity=24%25 Length=340 Hsps={ Subject=FSGKFTTFVLNKDQATLRMSSAEKTAEWNTAFDSRRGFF----TSGNYGL... Signif=0.99 Length=101 Bits=22.9 Query_start=110 Subject_end=243 Query=FWGKVLGFTL-KFNLNLRLTVNIDQLEWEVLHVSLHFWVVEVSTDQTLSVE... Positives=41%25 Expect=4.7, Identity=24%25 Query_end=210 Orientation=plus Score=65 Subject_start=146 Alignment=F GK F L K LR++ EW S + T +... } } =