#
# BioPerl module for Bio::SearchIO::blast
#
# Please direct questions and support issues to <bioperl-l@bioperl.org>
#
# Cared for by Jason Stajich <jason@bioperl.org>
#
# Copyright Jason Stajich
#
# You may distribute this module under the same terms as perl itself
# POD documentation - main docs before the code
# 20030409 - sac
# PSI-BLAST full parsing support. Rollout of new
# model which will remove Steve's old psiblast driver
# 20030424 - jason
# Megablast parsing fix as reported by Neil Saunders
# 20030427 - jason
# Support bl2seq parsing
# 20031124 - jason
# Parse more blast statistics, lambda, entropy, etc
# from WU-BLAST in frame-specific manner
# 20060216 - cjf - fixed blast parsing for BLAST v2.2.13 output
# 20071104 - dmessina - added support for WUBLAST -echofilter
# 20071121 - cjf - fixed several bugs (bugs 2391, 2399, 2409)
=head1 NAME
Bio::SearchIO::blast - Event generator for event based parsing of
blast reports
=head1 SYNOPSIS
# Do not use this object directly - it is used as part of the
# Bio::SearchIO system.
use Bio::SearchIO;
my $searchio = Bio::SearchIO->new(-format => 'blast',
-file => 't/data/ecolitst.bls');
while( my $result = $searchio->next_result ) {
while( my $hit = $result->next_hit ) {
while( my $hsp = $hit->next_hsp ) {
# ...
}
}
}
=head1 DESCRIPTION
This object encapsulated the necessary methods for generating events
suitable for building Bio::Search objects from a BLAST report file.
Read the L<Bio::SearchIO> for more information about how to use this.
This driver can parse:
=over 4
=item *
NCBI produced plain text BLAST reports from blastall, this also
includes PSIBLAST, PSITBLASTN, RPSBLAST, and bl2seq reports. NCBI XML
BLAST output is parsed with the blastxml SearchIO driver
=item *
WU-BLAST all reports
=item *
Jim Kent's BLAST-like output from his programs (BLASTZ, BLAT)
=item *
BLAST-like output from Paracel BTK output
=back
=head2 bl2seq parsing
Since I cannot differentiate between BLASTX and TBLASTN since bl2seq
doesn't report the algorithm used - I assume it is BLASTX by default -
you can supply the program type with -report_type in the SearchIO
constructor i.e.
my $parser = Bio::SearchIO->new(-format => 'blast',
-file => 'bl2seq.tblastn.report',
-report_type => 'tblastn');
This only really affects where the frame and strand information are
put - they will always be on the $hsp-E<gt>query instead of on the
$hsp-E<gt>hit part of the feature pair for blastx and tblastn bl2seq
produced reports. Hope that's clear...
=head1 FEEDBACK
=head2 Mailing Lists
User feedback is an integral part of the evolution of this and other
Bioperl modules. Send your comments and suggestions preferably to
the Bioperl mailing list. Your participation is much appreciated.
bioperl-l@bioperl.org - General discussion
http://bioperl.org/wiki/Mailing_lists - About the mailing lists
=head2 Support
Please direct usage questions or support issues to the mailing list:
I<bioperl-l@bioperl.org>
rather than to the module maintainer directly. Many experienced and
reponsive experts will be able look at the problem and quickly
address it. Please include a thorough description of the problem
with code and data examples if at all possible.
=head2 Reporting Bugs
Report bugs to the Bioperl bug tracking system to help us keep track
of the bugs and their resolution. Bug reports can be submitted via the
web:
https://redmine.open-bio.org/projects/bioperl/
=head1 AUTHOR - Jason Stajich
Email Jason Stajich jason-at-bioperl.org
=head1 CONTRIBUTORS
Steve Chervitz sac-at-bioperl.org
=head1 APPENDIX
The rest of the documentation details each of the object methods.
Internal methods are usually preceded with a _
=cut
# Let the code begin...'
package Bio::SearchIO::blast;
use Bio::SearchIO::IteratedSearchResultEventBuilder;
use strict;
use vars qw(%MAPPING %MODEMAP
$DEFAULT_BLAST_WRITER_CLASS
$MAX_HSP_OVERLAP
$DEFAULT_SIGNIF
$DEFAULT_SCORE
$DEFAULTREPORTTYPE
);
use base qw(Bio::SearchIO);
use Data::Dumper;
BEGIN {
# mapping of NCBI Blast terms to Bioperl hash keys
%MODEMAP = (
'BlastOutput' => 'result',
'Iteration' => 'iteration',
'Hit' => 'hit',
'Hsp' => 'hsp'
);
# This should really be done more intelligently, like with
# XSLT
%MAPPING = (
'Hsp_bit-score' => 'HSP-bits',
'Hsp_score' => 'HSP-score',
'Hsp_evalue' => 'HSP-evalue',
'Hsp_n', => 'HSP-n',
'Hsp_pvalue' => 'HSP-pvalue',
'Hsp_query-from' => 'HSP-query_start',
'Hsp_query-to' => 'HSP-query_end',
'Hsp_hit-from' => 'HSP-hit_start',
'Hsp_hit-to' => 'HSP-hit_end',
'Hsp_positive' => 'HSP-conserved',
'Hsp_identity' => 'HSP-identical',
'Hsp_gaps' => 'HSP-hsp_gaps',
'Hsp_hitgaps' => 'HSP-hit_gaps',
'Hsp_querygaps' => 'HSP-query_gaps',
'Hsp_qseq' => 'HSP-query_seq',
'Hsp_hseq' => 'HSP-hit_seq',
'Hsp_midline' => 'HSP-homology_seq',
'Hsp_align-len' => 'HSP-hsp_length',
'Hsp_query-frame' => 'HSP-query_frame',
'Hsp_hit-frame' => 'HSP-hit_frame',
'Hsp_links' => 'HSP-links',
'Hsp_group' => 'HSP-hsp_group',
'Hsp_features' => 'HSP-hit_features',
'Hit_id' => 'HIT-name',
'Hit_len' => 'HIT-length',
'Hit_accession' => 'HIT-accession',
'Hit_def' => 'HIT-description',
'Hit_signif' => 'HIT-significance',
# For NCBI blast, the description line contains bits.
# For WU-blast, the description line contains score.
'Hit_score' => 'HIT-score',
'Hit_bits' => 'HIT-bits',
'Iteration_iter-num' => 'ITERATION-number',
'Iteration_converged' => 'ITERATION-converged',
'BlastOutput_program' => 'RESULT-algorithm_name',
'BlastOutput_version' => 'RESULT-algorithm_version',
'BlastOutput_algorithm-reference' => 'RESULT-algorithm_reference',
'BlastOutput_rid' => 'RESULT-rid',
'BlastOutput_query-def' => 'RESULT-query_name',
'BlastOutput_query-len' => 'RESULT-query_length',
'BlastOutput_query-acc' => 'RESULT-query_accession',
'BlastOutput_query-gi' => 'RESULT-query_gi',
'BlastOutput_querydesc' => 'RESULT-query_description',
'BlastOutput_db' => 'RESULT-database_name',
'BlastOutput_db-len' => 'RESULT-database_entries',
'BlastOutput_db-let' => 'RESULT-database_letters',
'BlastOutput_inclusion-threshold' => 'RESULT-inclusion_threshold',
'Parameters_matrix' => { 'RESULT-parameters' => 'matrix' },
'Parameters_expect' => { 'RESULT-parameters' => 'expect' },
'Parameters_include' => { 'RESULT-parameters' => 'include' },
'Parameters_sc-match' => { 'RESULT-parameters' => 'match' },
'Parameters_sc-mismatch' => { 'RESULT-parameters' => 'mismatch' },
'Parameters_gap-open' => { 'RESULT-parameters' => 'gapopen' },
'Parameters_gap-extend' => { 'RESULT-parameters' => 'gapext' },
'Parameters_filter' => { 'RESULT-parameters' => 'filter' },
'Parameters_allowgaps' => { 'RESULT-parameters' => 'allowgaps' },
'Parameters_full_dbpath' => { 'RESULT-parameters' => 'full_dbpath' },
'Statistics_db-len' => { 'RESULT-statistics' => 'dbentries' },
'Statistics_db-let' => { 'RESULT-statistics' => 'dbletters' },
'Statistics_hsp-len' =>
{ 'RESULT-statistics' => 'effective_hsplength' },
'Statistics_query-len' => { 'RESULT-statistics' => 'querylength' },
'Statistics_eff-space' => { 'RESULT-statistics' => 'effectivespace' },
'Statistics_eff-spaceused' =>
{ 'RESULT-statistics' => 'effectivespaceused' },
'Statistics_eff-dblen' =>
{ 'RESULT-statistics' => 'effectivedblength' },
'Statistics_kappa' => { 'RESULT-statistics' => 'kappa' },
'Statistics_lambda' => { 'RESULT-statistics' => 'lambda' },
'Statistics_entropy' => { 'RESULT-statistics' => 'entropy' },
'Statistics_gapped_kappa' => { 'RESULT-statistics' => 'kappa_gapped' },
'Statistics_gapped_lambda' =>
{ 'RESULT-statistics' => 'lambda_gapped' },
'Statistics_gapped_entropy' =>
{ 'RESULT-statistics' => 'entropy_gapped' },
'Statistics_framewindow' =>
{ 'RESULT-statistics' => 'frameshiftwindow' },
'Statistics_decay' => { 'RESULT-statistics' => 'decayconst' },
'Statistics_hit_to_db' => { 'RESULT-statistics' => 'Hits_to_DB' },
'Statistics_num_suc_extensions' =>
{ 'RESULT-statistics' => 'num_successful_extensions' },
'Statistics_length_adjustment' => { 'RESULT-statistics' => 'length_adjustment' },
'Statistics_number_of_hsps_better_than_expect_value_cutoff_without_gapping' =>
{ 'RESULT-statistics' => 'number_of_hsps_better_than_expect_value_cutoff_without_gapping' },
'Statistics_number_of_hsps_gapped' => { 'RESULT-statistics' => 'number_of_hsps_gapped' },
'Statistics_number_of_hsps_successfully_gapped' => { 'RESULT-statistics' => 'number_of_hsps_successfully_gapped' },
# WU-BLAST stats
'Statistics_DFA_states' => { 'RESULT-statistics' => 'num_dfa_states' },
'Statistics_DFA_size' => { 'RESULT-statistics' => 'dfa_size' },
'Statistics_noprocessors' =>
{ 'RESULT-statistics' => 'no_of_processors' },
'Statistics_neighbortime' =>
{ 'RESULT-statistics' => 'neighborhood_generate_time' },
'Statistics_starttime' => { 'RESULT-statistics' => 'start_time' },
'Statistics_endtime' => { 'RESULT-statistics' => 'end_time' },
);
# add WU-BLAST Frame-Based Statistics
for my $frame ( 0 .. 3 ) {
for my $strand ( '+', '-' ) {
for my $ind (
qw(length efflength E S W T X X_gapped E2
E2_gapped S2)
)
{
$MAPPING{"Statistics_frame$strand$frame\_$ind"} =
{ 'RESULT-statistics' => "Frame$strand$frame\_$ind" };
}
for my $val (qw(lambda kappa entropy )) {
for my $type (qw(used computed gapped)) {
my $key = "Statistics_frame$strand$frame\_$val\_$type";
my $val =
{ 'RESULT-statistics' =>
"Frame$strand$frame\_$val\_$type" };
$MAPPING{$key} = $val;
}
}
}
}
# add Statistics
for my $stats (
qw(T A X1 X2 X3 S1 S2 X1_bits X2_bits X3_bits
S1_bits S2_bits num_extensions
num_successful_extensions
seqs_better_than_cutoff
posted_date
search_cputime total_cputime
search_actualtime total_actualtime
no_of_processors ctxfactor)
)
{
my $key = "Statistics_$stats";
my $val = { 'RESULT-statistics' => $stats };
$MAPPING{$key} = $val;
}
# add WU-BLAST Parameters
for my $param (
qw(span span1 span2 links warnings notes hspsepsmax
hspsepqmax topcomboN topcomboE postsw cpus wordmask
filter sort_by_pvalue sort_by_count sort_by_highscore
sort_by_totalscore sort_by_subjectlength noseqs gi qtype
qres V B Z Y M N)
)
{
my $key = "Parameters_$param";
my $val = { 'RESULT-parameters' => $param };
$MAPPING{$key} = $val;
}
$DEFAULT_BLAST_WRITER_CLASS = 'Bio::SearchIO::Writer::HitTableWriter';
$MAX_HSP_OVERLAP = 2; # Used when tiling multiple HSPs.
$DEFAULTREPORTTYPE = 'BLASTP'; # for bl2seq
}
=head2 new
Title : new
Usage : my $obj = Bio::SearchIO::blast->new(%args);
Function: Builds a new Bio::SearchIO::blast object
Returns : Bio::SearchIO::blast
Args : Key-value pairs:
-fh/-file => filehandle/filename to BLAST file
-format => 'blast'
-report_type => 'blastx', 'tblastn', etc -- only for bl2seq
reports when you want to distinguish between
tblastn and blastx reports (this only controls
where the frame information is put - on the query
or subject object.
-inclusion_threshold => e-value threshold for inclusion in the
PSI-BLAST score matrix model (blastpgp)
-signif => float or scientific notation number to be used
as a P- or Expect value cutoff
-score => integer or scientific notation number to be used
as a blast score value cutoff
-bits => integer or scientific notation number to be used
as a bit score value cutoff
-hit_filter => reference to a function to be used for
filtering hits based on arbitrary criteria.
All hits of each BLAST report must satisfy
this criteria to be retained.
If a hit fails this test, it is ignored.
This function should take a
Bio::Search::Hit::BlastHit.pm object as its first
argument and return true
if the hit should be retained.
Sample filter function:
-hit_filter => sub { $hit = shift;
$hit->gaps == 0; },
(Note: -filt_func is synonymous with -hit_filter)
-overlap => integer. The amount of overlap to permit between
adjacent HSPs when tiling HSPs. A reasonable value is 2.
Default = $Bio::SearchIO::blast::MAX_HSP_OVERLAP.
The following criteria are not yet supported:
(these are probably best applied within this module rather than in the
event handler since they would permit the parser to take some shortcuts.)
-check_all_hits => boolean. Check all hits for significance against
significance criteria. Default = false.
If false, stops processing hits after the first
non-significant hit or the first hit that fails
the hit_filter call. This speeds parsing,
taking advantage of the fact that the hits
are processed in the order they appear in the report.
-min_query_len => integer to be used as a minimum for query sequence length.
Reports with query sequences below this length will
not be processed. Default = no minimum length.
-best => boolean. Only process the best hit of each report;
default = false.
=cut
sub _initialize {
my ( $self, @args ) = @_;
$self->SUPER::_initialize(@args);
# Blast reports require a specialized version of the SREB due to the
# possibility of iterations (PSI-BLAST). Forwarding all arguments to it. An
# issue here is that we want to set new default object factories if none are
# supplied.
my $handler = Bio::SearchIO::IteratedSearchResultEventBuilder->new(@args);
$self->attach_EventHandler($handler);
# 2006-04-26 move this to the attach_handler function in this module so we
# can really reset the handler
# Optimization: caching
# the EventHandler since it is used a lot during the parse.
# $self->{'_handler_cache'} = $handler;
my ($rpttype ) = $self->_rearrange(
[
qw(
REPORT_TYPE)
],
@args
);
defined $rpttype && ( $self->{'_reporttype'} = $rpttype );
}
sub attach_EventHandler {
my ($self,$handler) = @_;
$self->SUPER::attach_EventHandler($handler);
# Optimization: caching the EventHandler since it is used a lot
# during the parse.
$self->{'_handler_cache'} = $handler;
return;
}
=head2 next_result
Title : next_result
Usage : my $hit = $searchio->next_result;
Function: Returns the next Result from a search
Returns : Bio::Search::Result::ResultI object
Args : none
=cut
sub next_result {
my ($self) = @_;
my $v = $self->verbose;
my $data = '';
my $flavor = '';
$self->{'_seentop'} = 0; # start next report at top
my ( $reporttype, $seenquery, $reportline, $reportversion );
my ( $seeniteration, $found_again );
my $incl_threshold = $self->inclusion_threshold;
my $bl2seq_fix;
$self->start_document(); # let the fun begin...
my (@hit_signifs);
my $gapped_stats = 0; # for switching between gapped/ungapped
# lambda, K, H
local $_ = "\n"; #consistency
PARSER:
while ( defined( $_ = $self->_readline ) ) {
next if (/^\s+$/); # skip empty lines
next if (/CPU time:/);
next if (/^>\s*$/);
next if (/[*]+\s+No hits found\s+[*]+/);
if (
/^((?:\S+?)?BLAST[NPX]?)\s+(.+)$/i # NCBI BLAST, PSIBLAST
# RPSBLAST, MEGABLAST
|| /^(P?GENEWISE|HFRAME|SWN|TSWN)\s+(.+)/i #Paracel BTK
)
{
($reporttype, $reportversion) = ($1, $2);
# need to keep track of whether this is WU-BLAST
if ($reportversion && $reportversion =~ m{WashU$}) {
$self->{'_wublast'}++;
}
$self->debug("blast.pm: Start of new report: $reporttype, $reportversion\n");
if ( $self->{'_seentop'} ) {
# This handles multi-result input streams
$self->_pushback($_);
last PARSER;
}
$self->_start_blastoutput;
if ($reporttype =~ /RPS-BLAST/) {
$reporttype .= '(BLASTP)'; # default RPS-BLAST type
}
$reportline = $_; # to fix the fact that RPS-BLAST output is wrong
$self->element(
{
'Name' => 'BlastOutput_program',
'Data' => $reporttype
}
);
$self->element(
{
'Name' => 'BlastOutput_version',
'Data' => $reportversion
}
);
$self->element(
{
'Name' => 'BlastOutput_inclusion-threshold',
'Data' => $incl_threshold
}
);
}
# parse the BLAST algorithm reference
elsif(/^Reference:\s+(.*)$/) {
# want to preserve newlines for the BLAST algorithm reference
my $algorithm_reference = "$1\n";
$_ = $self->_readline;
# while the current line, does not match an empty line, a RID:, a
# Database:, or a query definition line (Query=) we are still
# looking at the algorithm_reference, append it to what we parsed so
# far
while($_ !~ /^$/ && $_ !~ /^RID:/ && $_ !~ /^Database:/ && $_ !~ /^Query=/) {
$algorithm_reference .= "$_";
$_ = $self->_readline;
}
# if we exited the while loop, we saw an empty line, a RID:, or a
# Database:, so push it back
$self->_pushback($_);
$self->element(
{
'Name' => 'BlastOutput_algorithm-reference',
'Data' => $algorithm_reference
}
);
}
# parse BLAST RID (Request ID)
elsif(/^RID:\s+(.*)$/) {
my $rid = $1;
$self->element(
{
'Name' => 'BlastOutput_rid',
'Data' => $rid
}
);
}
# added Windows workaround for bug 1985
elsif (/^(Searching|Results from round)/) {
next unless $1 =~ /Results from round/;
$self->debug("blast.pm: Possible psi blast iterations found...\n");
$self->in_element('hsp')
&& $self->end_element( { 'Name' => 'Hsp' } );
$self->in_element('hit')
&& $self->end_element( { 'Name' => 'Hit' } );
if ( defined $seeniteration ) {
$self->within_element('iteration')
&& $self->end_element( { 'Name' => 'Iteration' } );
$self->start_element( { 'Name' => 'Iteration' } );
}
else {
$self->start_element( { 'Name' => 'Iteration' } );
}
$seeniteration = 1;
}
elsif (/^Query=\s*(.*)$/) {
my $q = $1;
$self->debug("blast.pm: Query= found...$_\n");
my $size = 0;
if ( defined $seenquery ) {
$self->_pushback($_);
$self->_pushback($reportline) if $reportline;
last PARSER;
}
if ( !defined $reporttype ) {
$self->_start_blastoutput;
if ( defined $seeniteration ) {
$self->in_element('iteration')
&& $self->end_element( { 'Name' => 'Iteration' } );
$self->start_element( { 'Name' => 'Iteration' } );
}
else {
$self->start_element( { 'Name' => 'Iteration' } );
}
$seeniteration = 1;
}
$seenquery = $q;
$_ = $self->_readline;
while ( defined($_) ) {
if (/^Database:/) {
$self->_pushback($_);
last;
}
# below line fixes length issue with BLAST v2.2.13; still works
# with BLAST v2.2.12
if ( /\((\-?[\d,]+)\s+letters.*\)/ || /^Length=(\-?[\d,]+)/ ) {
$size = $1;
$size =~ s/,//g;
last;
}
else {
# bug 2391
$q .= ($q =~ /\w$/ && $_ =~ /^\w/) ? " $_" : $_;
$q =~ s/\s+/ /g; # this catches the newline as well
$q =~ s/^ | $//g;
}
$_ = $self->_readline;
}
chomp($q);
my ( $nm, $desc ) = split( /\s+/, $q, 2 );
$self->element(
{
'Name' => 'BlastOutput_query-def',
'Data' => $nm
}
) if $nm;
$self->element(
{
'Name' => 'BlastOutput_query-len',
'Data' => $size
}
);
defined $desc && $desc =~ s/\s+$//;
$self->element(
{
'Name' => 'BlastOutput_querydesc',
'Data' => $desc
}
);
my ( $gi, $acc, $version ) = $self->_get_seq_identifiers($nm);
$version = defined($version) && length($version) ? ".$version" : "";
$self->element(
{
'Name' => 'BlastOutput_query-acc',
'Data' => "$acc$version"
}
) if $acc;
# these elements are dropped with some multiquery reports; add
# back here
$self->element(
{
'Name' => 'BlastOutput_db-len',
'Data' => $self->{'_blsdb_length'}
}
) if $self->{'_blsdb_length'};
$self->element(
{
'Name' => 'BlastOutput_db-let',
'Data' => $self->{'_blsdb_letters'}
}
) if $self->{'_blsdb_letters'};
$self->element(
{
'Name' => 'BlastOutput_db',
'Data' => $self->{'_blsdb'}
}
) if $self->{'_blsdb_letters'};
}
# added check for WU-BLAST -echofilter option (bug 2388)
elsif (/^>Unfiltered[+-]1$/) {
# skip all of the lines of unfiltered sequence
while($_ !~ /^Database:/) {
$self->debug("Bypassing features line: $_");
$_ = $self->_readline;
}
$self->_pushback($_);
}
elsif (/Sequences producing significant alignments:/) {
$self->debug("blast.pm: Processing NCBI-BLAST descriptions\n");
$flavor = 'ncbi';
# PSI-BLAST parsing needs to be fixed to specifically look
# for old vs new per iteration, as sorting based on duplication
# leads to bugs, see bug 1986
# The next line is not necessarily whitespace in psiblast reports.
# Also note that we must look for the end of this section by testing
# for a line with a leading >. Blank lines occur with this section
# for psiblast.
if ( !$self->in_element('iteration') ) {
$self->start_element( { 'Name' => 'Iteration' } );
}
# changed 8/28/2008 to exit hit table if blank line is found after an
# appropriate line
my $h_regex;
my $seen_block;
DESCLINE:
while ( defined( my $descline = $self->_readline() ) ) {
if ($descline =~ m{^\s*$}) {
last DESCLINE if $seen_block;
next DESCLINE;
}
# any text match is part of block...
$seen_block++;
# GCG multiline oddness...
if ($descline =~ /^(\S+)\s+Begin:\s\d+\s+End:\s+\d+/xms) {
my ($id, $nextline) = ($1, $self->_readline);
$nextline =~ s{^!}{};
$descline = "$id $nextline";
}
# NCBI style hit table (no N)
if ($descline =~ /(?<!cor) # negative lookahead
(\d*\.?(?:[\+\-eE]+)?\d+) # number (float or scientific notation)
\s+ # space
(\d*\.?(?:[\+\-eE]+)?\d+) # number (float or scientific notation)
\s*$/xms) {
my ( $score, $evalue ) = ($1, $2);
# Some data clean-up so e-value will appear numeric to perl
$evalue =~ s/^e/1e/i;
# This to handle no-HSP case
my @line = split ' ',$descline;
# we want to throw away the score, evalue
pop @line, pop @line;
# and N if it is present (of course they are not
# really in that order, but it doesn't matter
if ($3) { pop @line }
# add the last 2 entries s.t. we can reconstruct
# a minimal Hit object at the end of the day
push @hit_signifs, [ $evalue, $score, shift @line, join( ' ', @line ) ];
} elsif ($descline =~ /^CONVERGED/i) {
$self->element(
{
'Name' => 'Iteration_converged',
'Data' => 1
}
);
} else {
$self->_pushback($descline); # Catch leading > (end of section)
last DESCLINE;
}
}
}
elsif (/Sequences producing High-scoring Segment Pairs:/) {
# This block is for WU-BLAST, so we don't have to check for psi-blast stuff
# skip the next line
$self->debug("blast.pm: Processing WU-BLAST descriptions\n");
$_ = $self->_readline();
$flavor = 'wu';
if ( !$self->in_element('iteration') ) {
$self->start_element( { 'Name' => 'Iteration' } );
}
while ( defined( $_ = $self->_readline() )
&& !/^\s+$/ )
{
my @line = split;
pop @line; # throw away first number which is for 'N'col
# add the last 2 entries to array s.t. we can reconstruct
# a minimal Hit object at the end of the day
push @hit_signifs,
[ pop @line, pop @line, shift @line, join( ' ', @line ) ];
}
}
elsif (/^Database:\s*(.+?)\s*$/) {
$self->debug("blast.pm: Database: $1\n");
my $db = $1;
while ( defined( $_ = $self->_readline ) ) {
if (
/^\s+(\-?[\d\,]+|\S+)\s+sequences\;
\s+(\-?[\d,]+|\S+)\s+ # Deal with NCBI 2.2.8 OSX problems
total\s+letters/ox
)
{
my ( $s, $l ) = ( $1, $2 );
$s =~ s/,//g;
$l =~ s/,//g;
$self->element(
{
'Name' => 'BlastOutput_db-len',
'Data' => $s
}
);
$self->element(
{
'Name' => 'BlastOutput_db-let',
'Data' => $l
}
);
# cache for next round in cases with multiple queries
$self->{'_blsdb'} = $db;
$self->{'_blsdb_length'} = $s;
$self->{'_blsdb_letters'} = $l;
last;
}
else {
chomp;
$db .= $_;
}
}
$self->element(
{
'Name' => 'BlastOutput_db',
'Data' => $db
}
);
}
# move inside of a hit
elsif (/^>\s*(\S+)\s*(.*)?/) {
chomp;
$self->debug("blast.pm: Hit: $1\n");
$self->in_element('hsp')
&& $self->end_element( { 'Name' => 'Hsp' } );
$self->in_element('hit')
&& $self->end_element( { 'Name' => 'Hit' } );
# special case when bl2seq reports don't have a leading
# Query=
if ( !$self->within_element('result') ) {
$self->_start_blastoutput;
$self->start_element( { 'Name' => 'Iteration' } );
}
elsif ( !$self->within_element('iteration') ) {
$self->start_element( { 'Name' => 'Iteration' } );
}
$self->start_element( { 'Name' => 'Hit' } );
my $id = $1;
my $restofline = $2;
$self->debug("Starting a hit: $1 $2\n");
$self->element(
{
'Name' => 'Hit_id',
'Data' => $id
}
);
my ($gi, $acc, $version ) = $self->_get_seq_identifiers($id);
$self->element(
{
'Name' => 'Hit_accession',
'Data' => $acc
}
);
# add hit significance (from the hit table)
# this is where Bug 1986 went awry
# Changed for Bug2409; hit->significance and hit->score/bits derived
# from HSPs, not hit table unless necessary
HITTABLE:
while (my $v = shift @hit_signifs) {
my $tableid = $v->[2];
if ($tableid !~ m{\Q$id\E}) {
$self->debug("Hit table ID $tableid doesn't match current hit id $id, checking next hit table entry...\n");
next HITTABLE;
} else {
last HITTABLE;
}
}
while ( defined( $_ = $self->_readline() ) ) {
next if (/^\s+$/);
chomp;
if (/Length\s*=\s*([\d,]+)/) {
my $l = $1;
$l =~ s/\,//g;
$self->element(
{
'Name' => 'Hit_len',
'Data' => $l
}
);
last;
}
else {
if ($restofline !~ /\s$/) { # bug #3235
s/^\s(?!\s)/\x01/; #new line to concatenate desc lines with <soh>
}
$restofline .= $_;
}
}
$restofline =~ s/\s+/ /g;
$self->element(
{
'Name' => 'Hit_def',
'Data' => $restofline
}
);
}
elsif (/\s+(Plus|Minus) Strand HSPs:/i) {
next;
}
elsif (
( $self->in_element('hit') || $self->in_element('hsp') )
&& # paracel genewise BTK
m/Score\s*=\s*(\S+)\s*bits\s* # Bit score
(?:\((\d+)\))?, # Raw score
\s+Log\-Length\sScore\s*=\s*(\d+) # Log-Length score
/ox
)
{
$self->in_element('hsp')
&& $self->end_element( { 'Name' => 'Hsp' } );
$self->start_element( { 'Name' => 'Hsp' } );
$self->debug( "Got paracel genewise HSP score=$1\n");
# Some data clean-up so e-value will appear numeric to perl
my ( $bits, $score, $evalue ) = ( $1, $2, $3 );
$evalue =~ s/^e/1e/i;
$self->element(
{
'Name' => 'Hsp_score',
'Data' => $score
}
);
$self->element(
{
'Name' => 'Hsp_bit-score',
'Data' => $bits
}
);
$self->element(
{
'Name' => 'Hsp_evalue',
'Data' => $evalue
}
);
}
elsif (
( $self->in_element('hit') || $self->in_element('hsp') )
&& # paracel hframe BTK
m/Score\s*=\s*([^,\s]+), # Raw score
\s*Expect\s*=\s*([^,\s]+), # E-value
\s*P(?:\(\S+\))?\s*=\s*([^,\s]+) # P-value
/ox
)
{
$self->in_element('hsp')
&& $self->end_element( { 'Name' => 'Hsp' } );
$self->start_element( { 'Name' => 'Hsp' } );
$self->debug( "Got paracel hframe HSP score=$1\n");
# Some data clean-up so e-value will appear numeric to perl
my ( $score, $evalue, $pvalue ) = ( $1, $2, $3 );
$evalue = "1$evalue" if $evalue =~ /^e/;
$pvalue = "1$pvalue" if $pvalue =~ /^e/;
$self->element(
{
'Name' => 'Hsp_score',
'Data' => $score
}
);
$self->element(
{
'Name' => 'Hsp_evalue',
'Data' => $evalue
}
);
$self->element(
{
'Name' => 'Hsp_pvalue',
'Data' => $pvalue
}
);
}
elsif (
( $self->in_element('hit') || $self->in_element('hsp') )
&& # wublast
m/Score\s*=\s*(\S+)\s* # Bit score
\(([\d\.]+)\s*bits\), # Raw score
\s*Expect\s*=\s*([^,\s]+), # E-value
\s*(?:Sum)?\s* # SUM
P(?:\(\d+\))?\s*=\s*([^,\s]+) # P-value
(?:\s*,\s+Group\s*\=\s*(\d+))? # HSP Group
/ox
)
{ # wu-blast HSP parse
$self->in_element('hsp')
&& $self->end_element( { 'Name' => 'Hsp' } );
$self->start_element( { 'Name' => 'Hsp' } );
# Some data clean-up so e-value will appear numeric to perl
my ( $score, $bits, $evalue, $pvalue, $group ) =
( $1, $2, $3, $4, $5 );
$evalue =~ s/^e/1e/i;
$pvalue =~ s/^e/1e/i;
$self->element(
{
'Name' => 'Hsp_score',
'Data' => $score
}
);
$self->element(
{
'Name' => 'Hsp_bit-score',
'Data' => $bits
}
);
$self->element(
{
'Name' => 'Hsp_evalue',
'Data' => $evalue
}
);
$self->element(
{
'Name' => 'Hsp_pvalue',
'Data' => $pvalue
}
);
if ( defined $group ) {
$self->element(
{
'Name' => 'Hsp_group',
'Data' => $group
}
);
}
}
elsif (
( $self->in_element('hit') || $self->in_element('hsp') )
&& # ncbi blast, works with 2.2.17
m/^\sFeatures\s\w+\sthis\spart/xmso
# If the line begins with "Features in/flanking this part of subject sequence:"
)
{
$self->in_element('hsp')
&& $self->end_element( { 'Name' => 'Hsp' } );
my $featline;
$_ = $self->_readline;
while($_ !~ /^\s*$/) {
chomp;
$featline .= $_;
$_ = $self->_readline;
}
$self->_pushback($_);
$featline =~ s{(?:^\s+|\s+^)}{}g;
$featline =~ s{\n}{;}g;
$self->start_element( { 'Name' => 'Hsp' } );
$self->element(
{
'Name' => 'Hsp_features',
'Data' => $featline
}
);
$self->{'_seen_hsp_features'} = 1;
}
elsif (
( $self->in_element('hit') || $self->in_element('hsp') )
&& # ncbi blast, works with 2.2.17
m/Score\s*=\s*(\S+)\s*bits\s* # Bit score
(?:\((\d+)\))?, # Missing for BLAT pseudo-BLAST fmt
\s*Expect(?:\((\d+\+?)\))?\s*=\s*([^,\s]+) # E-value
/ox
)
{ # parse NCBI blast HSP
if( !$self->{'_seen_hsp_features'} ) {
$self->in_element('hsp')
&& $self->end_element( { 'Name' => 'Hsp' } );
$self->start_element( { 'Name' => 'Hsp' } );
}
# Some data clean-up so e-value will appear numeric to perl
my ( $bits, $score, $n, $evalue ) = ( $1, $2, $3, $4 );
$evalue =~ s/^e/1e/i;
$self->element(
{
'Name' => 'Hsp_score',
'Data' => $score
}
);
$self->element(
{
'Name' => 'Hsp_bit-score',
'Data' => $bits
}
);
$self->element(
{
'Name' => 'Hsp_evalue',
'Data' => $evalue
}
);
$self->element(
{
'Name' => 'Hsp_n',
'Data' => $n
}
) if defined $n;
$score = '' unless defined $score; # deal with BLAT which
# has no score only bits
$self->debug("Got NCBI HSP score=$score, evalue $evalue\n");
}
elsif (
$self->in_element('hsp')
&& m/Identities\s*=\s*(\d+)\s*\/\s*(\d+)\s*[\d\%\(\)]+\s*
(?:,\s*Positives\s*=\s*(\d+)\/(\d+)\s*[\d\%\(\)]+\s*)? # pos only valid for Protein alignments
(?:\,\s*Gaps\s*=\s*(\d+)\/(\d+))? # Gaps
/oxi
)
{
$self->element(
{
'Name' => 'Hsp_identity',
'Data' => $1
}
);
$self->element(
{
'Name' => 'Hsp_align-len',
'Data' => $2
}
);
if ( defined $3 ) {
$self->element(
{
'Name' => 'Hsp_positive',
'Data' => $3
}
);
}
else {
$self->element(
{
'Name' => 'Hsp_positive',
'Data' => $1
}
);
}
if ( defined $6 ) {
$self->element(
{
'Name' => 'Hsp_gaps',
'Data' => $5
}
);
}
$self->{'_Query'} = { 'begin' => 0, 'end' => 0 };
$self->{'_Sbjct'} = { 'begin' => 0, 'end' => 0 };
if (/(Frame\s*=\s*.+)$/) {
# handle wu-blast Frame listing on same line
$self->_pushback($1);
}
}
elsif ( $self->in_element('hsp')
&& /Strand\s*=\s*(Plus|Minus)\s*\/\s*(Plus|Minus)/i )
{
# consume this event ( we infer strand from start/end)
if (!defined($reporttype)) {
$self->{'_reporttype'} = $reporttype = 'BLASTN';
$bl2seq_fix = 1; # special case to resubmit the algorithm
# reporttype
}
next;
}
elsif ( $self->in_element('hsp')
&& /Links\s*=\s*(\S+)/ox )
{
$self->element(
{
'Name' => 'Hsp_links',
'Data' => $1
}
);
}
elsif ( $self->in_element('hsp')
&& /Frame\s*=\s*([\+\-][1-3])\s*(\/\s*([\+\-][1-3]))?/ )
{
# this is for bl2seq only
unless ( defined $reporttype ) {
$bl2seq_fix = 1;
if ( $1 && $2 ) { $reporttype = 'TBLASTX' }
else {
$reporttype = 'BLASTX';
# we can't distinguish between BLASTX and TBLASTN straight from the report }
}
$self->{'_reporttype'} = $reporttype;
}
my ( $queryframe, $hitframe );
if ( $reporttype eq 'TBLASTX' ) {
( $queryframe, $hitframe ) = ( $1, $2 );
$hitframe =~ s/\/\s*//g;
}
elsif ( $reporttype eq 'TBLASTN' || $reporttype eq 'PSITBLASTN') {
( $hitframe, $queryframe ) = ( $1, 0 );
}
elsif ( $reporttype eq 'BLASTX' || $reporttype eq 'RPS-BLAST(BLASTP)') {
( $queryframe, $hitframe ) = ( $1, 0 );
# though NCBI doesn't report it, this is a special BLASTX-like
# RPS-BLAST; should be handled differently
if ($reporttype eq 'RPS-BLAST(BLASTP)') {
$self->element(
{
'Name' => 'BlastOutput_program',
'Data' => 'RPS-BLAST(BLASTX)'
}
);
}
}
$self->element(
{
'Name' => 'Hsp_query-frame',
'Data' => $queryframe
}
);
$self->element(
{
'Name' => 'Hsp_hit-frame',
'Data' => $hitframe
}
);
}
elsif (/^Parameters:/
|| /^\s+Database:\s+?/
|| /^\s+Subset/
|| /^\s*Lambda/
|| /^\s*Histogram/
|| ( $self->in_element('hsp') && /WARNING|NOTE/ ) )
{
# Note: Lambda check was necessary to parse
# t/data/ecoli_domains.rpsblast AND to parse bl2seq
$self->debug("blast.pm: found parameters section \n");
$self->in_element('hsp')
&& $self->end_element( { 'Name' => 'Hsp' } );
$self->in_element('hit')
&& $self->end_element( { 'Name' => 'Hit' } );
# This is for the case when we specify -b 0 (or B=0 for WU-BLAST)
# and still want to construct minimal Hit objects
$self->_cleanup_hits(\@hit_signifs) if scalar(@hit_signifs);
$self->within_element('iteration')
&& $self->end_element( { 'Name' => 'Iteration' } );
next if /^\s+Subset/;
my $blast = (/^(\s+Database\:)|(\s*Lambda)/) ? 'ncbi' : 'wublast';
if (/^\s*Histogram/) {
$blast = 'btk';
}
my $last = '';
# default is that gaps are allowed
$self->element(
{
'Name' => 'Parameters_allowgaps',
'Data' => 'yes'
}
);
while ( defined( $_ = $self->_readline ) ) {
if (
/^((?:\S+)?BLAST[NPX]?)\s+(.+)$/i # NCBI BLAST, PSIBLAST
# RPSBLAST, MEGABLAST
|| /^(P?GENEWISE|HFRAME|SWN|TSWN)\s+(.+)/i #Paracel BTK
)
{
$self->_pushback($_);
# let's handle this in the loop
last;
}
elsif (/^Query=/) {
$self->_pushback($_);
$self->_pushback($reportline) if $reportline;
last PARSER;
}
# here is where difference between wublast and ncbiblast
# is better handled by different logic
if ( /Number of Sequences:\s+([\d\,]+)/i
|| /of sequences in database:\s+(\-?[\d,]+)/i )
{
my $c = $1;
$c =~ s/\,//g;
$self->element(
{
'Name' => 'Statistics_db-len',
'Data' => $c
}
);
}
elsif (/letters in database:\s+(\-?[\d,]+)/i) {
my $s = $1;
$s =~ s/,//g;
$self->element(
{
'Name' => 'Statistics_db-let',
'Data' => $s
}
);
}
elsif ( $blast eq 'btk' ) {
next;
}
elsif ( $blast eq 'wublast' ) {
# warn($_);
if (/E=(\S+)/) {
$self->element(
{
'Name' => 'Parameters_expect',
'Data' => $1
}
);
}
elsif (/nogaps/) {
$self->element(
{
'Name' => 'Parameters_allowgaps',
'Data' => 'no'
}
);
}
elsif (/ctxfactor=(\S+)/) {
$self->element(
{
'Name' => 'Statistics_ctxfactor',
'Data' => $1
}
);
}
elsif (
/(postsw|links|span[12]?|warnings|notes|gi|noseqs|qres|qype)/
)
{
$self->element(
{
'Name' => "Parameters_$1",
'Data' => 'yes'
}
);
}
elsif (/(\S+)=(\S+)/) {
$self->element(
{
'Name' => "Parameters_$1",
'Data' => $2
}
);
}
elsif ( $last =~ /(Frame|Strand)\s+MatID\s+Matrix name/i ) {
my $firstgapinfo = 1;
my $frame = undef;
while ( defined($_) && !/^\s+$/ ) {
s/^\s+//;
s/\s+$//;
if ( $firstgapinfo
&& s/Q=(\d+),R=(\d+)\s+//x )
{
$firstgapinfo = 0;
$self->element(
{
'Name' => 'Parameters_gap-open',
'Data' => $1
}
);
$self->element(
{
'Name' => 'Parameters_gap-extend',
'Data' => $2
}
);
my @fields = split;
for my $type (
qw(lambda_gapped
kappa_gapped
entropy_gapped)
)
{
next if $type eq 'n/a';
if ( !@fields ) {
warn "fields is empty for $type\n";
next;
}
$self->element(
{
'Name' =>
"Statistics_frame$frame\_$type",
'Data' => shift @fields
}
);
}
}
else {
my ( $frameo, $matid, $matrix, @fields ) =
split;
if ( !defined $frame ) {
# keep some sort of default feature I guess
# even though this is sort of wrong
$self->element(
{
'Name' => 'Parameters_matrix',
'Data' => $matrix
}
);
$self->element(
{
'Name' => 'Statistics_lambda',
'Data' => $fields[0]
}
);
$self->element(
{
'Name' => 'Statistics_kappa',
'Data' => $fields[1]
}
);
$self->element(
{
'Name' => 'Statistics_entropy',
'Data' => $fields[2]
}
);
}
$frame = $frameo;
my $ii = 0;
for my $type (
qw(lambda_used
kappa_used
entropy_used
lambda_computed
kappa_computed
entropy_computed)
)
{
my $f = $fields[$ii];
next unless defined $f; # deal with n/a
if ( $f eq 'same' ) {
$f = $fields[ $ii - 3 ];
}
$ii++;
$self->element(
{
'Name' =>
"Statistics_frame$frame\_$type",
'Data' => $f
}
);
}
}
# get the next line
$_ = $self->_readline;
}
$last = $_;
}
elsif ( $last =~ /(Frame|Strand)\s+MatID\s+Length/i ) {
my $frame = undef;
while ( defined($_) && !/^\s+/ ) {
s/^\s+//;
s/\s+$//;
my @fields = split;
if ( @fields <= 3 ) {
for my $type (qw(X_gapped E2_gapped S2)) {
last unless @fields;
$self->element(
{
'Name' =>
"Statistics_frame$frame\_$type",
'Data' => shift @fields
}
);
}
}
else {
for my $type (
qw(length
efflength
E S W T X E2 S2)
)
{
$self->element(
{
'Name' =>
"Statistics_frame$frame\_$type",
'Data' => shift @fields
}
);
}
}
$_ = $self->_readline;
}
$last = $_;
}
elsif (/(\S+\s+\S+)\s+DFA:\s+(\S+)\s+\((.+)\)/) {
if ( $1 eq 'states in' ) {
$self->element(
{
'Name' => 'Statistics_DFA_states',
'Data' => "$2 $3"
}
);
}
elsif ( $1 eq 'size of' ) {
$self->element(
{
'Name' => 'Statistics_DFA_size',
'Data' => "$2 $3"
}
);
}
}
elsif (
m/^\s+Time to generate neighborhood:\s+
(\S+\s+\S+\s+\S+)/x
)
{
$self->element(
{
'Name' => 'Statistics_neighbortime',
'Data' => $1
}
);
}
elsif (/processors\s+used:\s+(\d+)/) {
$self->element(
{
'Name' => 'Statistics_noprocessors',
'Data' => $1
}
);
}
elsif (
m/^\s+(\S+)\s+cpu\s+time:\s+# cputype
(\S+\s+\S+\s+\S+) # cputime
\s+Elapsed:\s+(\S+)/x
)
{
my $cputype = lc($1);
$self->element(
{
'Name' => "Statistics_$cputype\_cputime",
'Data' => $2
}
);
$self->element(
{
'Name' => "Statistics_$cputype\_actualtime",
'Data' => $3
}
);
}
elsif (/^\s+Start:/) {
my ( $junk, $start, $stime, $end, $etime ) =
split( /\s+(Start|End)\:\s+/, $_ );
chomp($stime);
$self->element(
{
'Name' => 'Statistics_starttime',
'Data' => $stime
}
);
chomp($etime);
$self->element(
{
'Name' => 'Statistics_endtime',
'Data' => $etime
}
);
}
elsif (/^\s+Database:\s+(.+)$/) {
$self->element(
{
'Name' => 'Parameters_full_dbpath',
'Data' => $1
}
);
}
elsif (/^\s+Posted:\s+(.+)/) {
my $d = $1;
chomp($d);
$self->element(
{
'Name' => 'Statistics_posted_date',
'Data' => $d
}
);
}
}
elsif ( $blast eq 'ncbi' ) {
if (m/^Matrix:\s+(.+)\s*$/oxi) {
$self->element(
{
'Name' => 'Parameters_matrix',
'Data' => $1
}
);
}
elsif (/^Gapped/) {
$gapped_stats = 1;
}
elsif (/^Lambda/) {
$_ = $self->_readline;
s/^\s+//;
my ( $lambda, $kappa, $entropy ) = split;
if ($gapped_stats) {
$self->element(
{
'Name' => "Statistics_gapped_lambda",
'Data' => $lambda
}
);
$self->element(
{
'Name' => "Statistics_gapped_kappa",
'Data' => $kappa
}
);
$self->element(
{
'Name' => "Statistics_gapped_entropy",
'Data' => $entropy
}
);
}
else {
$self->element(
{
'Name' => "Statistics_lambda",
'Data' => $lambda
}
);
$self->element(
{
'Name' => "Statistics_kappa",
'Data' => $kappa
}
);
$self->element(
{
'Name' => "Statistics_entropy",
'Data' => $entropy
}
);
}
}
elsif (m/effective\s+search\s+space\s+used:\s+(\d+)/oxi) {
$self->element(
{
'Name' => 'Statistics_eff-spaceused',
'Data' => $1
}
);
}
elsif (m/effective\s+search\s+space:\s+(\d+)/oxi) {
$self->element(
{
'Name' => 'Statistics_eff-space',
'Data' => $1
}
);
}
elsif (
m/Gap\s+Penalties:\s+Existence:\s+(\d+)\,
\s+Extension:\s+(\d+)/ox
)
{
$self->element(
{
'Name' => 'Parameters_gap-open',
'Data' => $1
}
);
$self->element(
{
'Name' => 'Parameters_gap-extend',
'Data' => $2
}
);
}
elsif (/effective\s+HSP\s+length:\s+(\d+)/) {
$self->element(
{
'Name' => 'Statistics_hsp-len',
'Data' => $1
}
);
}
elsif (/Number\s+of\s+HSP's\s+better\s+than\s+(\S+)\s+without\s+gapping:\s+(\d+)/) {
$self->element(
{
'Name' => 'Statistics_number_of_hsps_better_than_expect_value_cutoff_without_gapping',
'Data' => $2
}
);
}
elsif (/Number\s+of\s+HSP's\s+gapped:\s+(\d+)/) {
$self->element(
{
'Name' => 'Statistics_number_of_hsps_gapped',
'Data' => $1
}
);
}
elsif (/Number\s+of\s+HSP's\s+successfully\s+gapped:\s+(\d+)/) {
$self->element(
{
'Name' => 'Statistics_number_of_hsps_successfully_gapped',
'Data' => $1
}
);
}
elsif (/Length\s+adjustment:\s+(\d+)/) {
$self->element(
{
'Name' => 'Statistics_length_adjustment',
'Data' => $1
}
);
}
elsif (/effective\s+length\s+of\s+query:\s+([\d\,]+)/i) {
my $c = $1;
$c =~ s/\,//g;
$self->element(
{
'Name' => 'Statistics_query-len',
'Data' => $c
}
);
}
elsif (/effective\s+length\s+of\s+database:\s+([\d\,]+)/i) {
my $c = $1;
$c =~ s/\,//g;
$self->element(
{
'Name' => 'Statistics_eff-dblen',
'Data' => $c
}
);
}
elsif (
/^(T|A|X1|X2|X3|S1|S2):\s+(\d+(\.\d+)?)\s+(?:\(\s*(\d+\.\d+) bits\))?/
)
{
my $v = $2;
chomp($v);
$self->element(
{
'Name' => "Statistics_$1",
'Data' => $v
}
);
if ( defined $4 ) {
$self->element(
{
'Name' => "Statistics_$1_bits",
'Data' => $4
}
);
}
}
elsif (
m/frameshift\s+window\,
\s+decay\s+const:\s+(\d+)\,\s+([\.\d]+)/x
)
{
$self->element(
{
'Name' => 'Statistics_framewindow',
'Data' => $1
}
);
$self->element(
{
'Name' => 'Statistics_decay',
'Data' => $2
}
);
}
elsif (m/^Number\s+of\s+Hits\s+to\s+DB:\s+(\S+)/ox) {
$self->element(
{
'Name' => 'Statistics_hit_to_db',
'Data' => $1
}
);
}
elsif (m/^Number\s+of\s+extensions:\s+(\S+)/ox) {
$self->element(
{
'Name' => 'Statistics_num_extensions',
'Data' => $1
}
);
}
elsif (
m/^Number\s+of\s+successful\s+extensions:\s+
(\S+)/ox
)
{
$self->element(
{
'Name' => 'Statistics_num_suc_extensions',
'Data' => $1
}
);
}
elsif (
m/^Number\s+of\s+sequences\s+better\s+than\s+
(\S+):\s+(\d+)/ox
)
{
$self->element(
{
'Name' => 'Parameters_expect',
'Data' => $1
}
);
$self->element(
{
'Name' => 'Statistics_seqs_better_than_cutoff',
'Data' => $2
}
);
}
elsif (/^\s+Posted\s+date:\s+(.+)/) {
my $d = $1;
chomp($d);
$self->element(
{
'Name' => 'Statistics_posted_date',
'Data' => $d
}
);
}
elsif ( !/^\s+$/ ) {
#$self->debug( "unmatched stat $_");
}
}
$last = $_;
}
} elsif ( $self->in_element('hsp') ) {
$self->debug("blast.pm: Processing HSP\n");
# let's read 3 lines at a time;
# bl2seq hackiness... Not sure I like
$self->{'_reporttype'} ||= $DEFAULTREPORTTYPE;
my %data = (
'Query' => '',
'Mid' => '',
'Hit' => ''
);
my $len;
for ( my $i = 0 ; defined($_) && $i < 3 ; $i++ ) {
# $self->debug("$i: $_") if $v;
if ( ( $i == 0 && /^\s+$/) ||
/^\s*(?:Lambda|Minus|Plus|Score)/i )
{
$self->_pushback($_) if defined $_;
$self->end_element( { 'Name' => 'Hsp' } );
last;
}
chomp;
if (/^((Query|Sbjct):?\s+(\-?\d+)?\s*)(\S+)\s+(\-?\d+)?/) {
my ( $full, $type, $start, $str, $end ) =
( $1, $2, $3, $4, $5 );
if ( $str eq '-' ) {
$i = 3 if $type eq 'Sbjct';
}
else {
$data{$type} = $str;
}
$len = length($full);
$self->{"\_$type"}->{'begin'} = $start
unless $self->{"_$type"}->{'begin'};
$self->{"\_$type"}->{'end'} = $end;
} elsif (/^((Query|Sbjct):?\s+(\-?0+)\s*)/) {
# Bug from NCBI's BLAST: unaligned output
$_ = $self->_readline() for 0..1;
last;
} else {
$self->throw("no data for midline $_")
unless ( defined $_ && defined $len );
$data{'Mid'} = substr( $_, $len );
}
$_ = $self->_readline();
}
$self->characters(
{
'Name' => 'Hsp_qseq',
'Data' => $data{'Query'}
}
);
$self->characters(
{
'Name' => 'Hsp_hseq',
'Data' => $data{'Sbjct'}
}
);
$self->characters(
{
'Name' => 'Hsp_midline',
'Data' => $data{'Mid'}
}
);
}
else {
#$self->debug("blast.pm: unrecognized line $_");
}
}
$self->debug("blast.pm: End of BlastOutput\n");
if ( $self->{'_seentop'} ) {
$self->within_element('hsp')
&& $self->end_element( { 'Name' => 'Hsp' } );
$self->within_element('hit')
&& $self->end_element( { 'Name' => 'Hit' } );
# cleanup extra hits
$self->_cleanup_hits(\@hit_signifs) if scalar(@hit_signifs);
$self->within_element('iteration')
&& $self->end_element( { 'Name' => 'Iteration' } );
if ($bl2seq_fix) {
$self->element(
{
'Name' => 'BlastOutput_program',
'Data' => $reporttype
}
);
}
$self->end_element( { 'Name' => 'BlastOutput' } );
}
return $self->end_document();
}
# Private method for internal use only.
sub _start_blastoutput {
my $self = shift;
$self->start_element( { 'Name' => 'BlastOutput' } );
$self->{'_seentop'} = 1;
$self->{'_result_count'}++;
$self->{'_handler_rc'} = undef;
}
=head2 _will_handle
Title : _will_handle
Usage : Private method. For internal use only.
if( $self->_will_handle($type) ) { ... }
Function: Provides an optimized way to check whether or not an element of a
given type is to be handled.
Returns : Reference to EventHandler object if the element type is to be handled.
undef if the element type is not to be handled.
Args : string containing type of element.
Optimizations:
=over 2
=item 1
Using the cached pointer to the EventHandler to minimize repeated
lookups.
=item 2
Caching the will_handle status for each type that is encountered so
that it only need be checked by calling
handler-E<gt>will_handle($type) once.
=back
This does not lead to a major savings by itself (only 5-10%). In
combination with other optimizations, or for large parse jobs, the
savings good be significant.
To test against the unoptimized version, remove the parentheses from
around the third term in the ternary " ? : " operator and add two
calls to $self-E<gt>_eventHandler().
=cut
sub _will_handle {
my ( $self, $type ) = @_;
my $handler = $self->{'_handler_cache'};
my $will_handle =
defined( $self->{'_will_handle_cache'}->{$type} )
? $self->{'_will_handle_cache'}->{$type}
: ( $self->{'_will_handle_cache'}->{$type} =
$handler->will_handle($type) );
return $will_handle ? $handler : undef;
}
=head2 start_element
Title : start_element
Usage : $eventgenerator->start_element
Function: Handles a start element event
Returns : none
Args : hashref with at least 2 keys 'Data' and 'Name'
=cut
sub start_element {
my ( $self, $data ) = @_;
# we currently don't care about attributes
my $nm = $data->{'Name'};
my $type = $MODEMAP{$nm};
if ($type) {
my $handler = $self->_will_handle($type);
if ($handler) {
my $func = sprintf( "start_%s", lc $type );
$self->{'_handler_rc'} = $handler->$func( $data->{'Attributes'} );
}
#else {
#$self->debug( # changed 4/29/2006 to play nice with other event handlers
# "Bio::SearchIO::InternalParserError ".
# "\nCan't handle elements of type \'$type.\'"
#);
#}
unshift @{ $self->{'_elements'} }, $type;
if ( $type eq 'result' ) {
$self->{'_values'} = {};
$self->{'_result'} = undef;
} else {
# cleanup some things
if ( defined $self->{'_values'} ) {
foreach my $k (
grep { /^\U$type\-/ }
keys %{ $self->{'_values'} }
)
{
delete $self->{'_values'}->{$k};
}
}
}
}
}
=head2 end_element
Title : end_element
Usage : $eventgenerator->end_element
Function: Handles an end element event
Returns : hashref with an element's worth of data
Args : hashref with at least 2 keys 'Data' and 'Name'
=cut
sub end_element {
my ( $self, $data ) = @_;
my $nm = $data->{'Name'};
my $type;
my $rc;
# cache these (TODO: we should probably cache all cross-report data)
if ( $nm eq 'BlastOutput_program' ) {
if ( $self->{'_last_data'} =~ /(t?blast[npx])/i ) {
$self->{'_reporttype'} = uc $1;
}
$self->{'_reporttype'} ||= $DEFAULTREPORTTYPE;
}
if ( $nm eq 'BlastOutput_version' ) {
$self->{'_reportversion'} = $self->{'_last_data'};
}
# Hsps are sort of weird, in that they end when another
# object begins so have to detect this in end_element for now
if ( $nm eq 'Hsp' ) {
foreach (qw(Hsp_qseq Hsp_midline Hsp_hseq Hsp_features)) {
$self->element(
{
'Name' => $_,
'Data' => $self->{'_last_hspdata'}->{$_}
}
) if defined $self->{'_last_hspdata'}->{$_};
}
$self->{'_last_hspdata'} = {};
$self->element(
{
'Name' => 'Hsp_query-from',
'Data' => $self->{'_Query'}->{'begin'}
}
);
$self->element(
{
'Name' => 'Hsp_query-to',
'Data' => $self->{'_Query'}->{'end'}
}
);
$self->element(
{
'Name' => 'Hsp_hit-from',
'Data' => $self->{'_Sbjct'}->{'begin'}
}
);
$self->element(
{
'Name' => 'Hsp_hit-to',
'Data' => $self->{'_Sbjct'}->{'end'}
}
);
# } elsif( $nm eq 'Iteration' ) {
# Nothing special needs to be done here.
}
if ( $type = $MODEMAP{$nm} ) {
my $handler = $self->_will_handle($type);
if ($handler) {
my $func = sprintf( "end_%s", lc $type );
$rc = $handler->$func( $self->{'_reporttype'}, $self->{'_values'} );
}
shift @{ $self->{'_elements'} };
}
elsif ( $MAPPING{$nm} ) {
if ( ref( $MAPPING{$nm} ) =~ /hash/i ) {
# this is where we shove in the data from the
# hashref info about params or statistics
my $key = ( keys %{ $MAPPING{$nm} } )[0];
$self->{'_values'}->{$key}->{ $MAPPING{$nm}->{$key} } =
$self->{'_last_data'};
}
else {
$self->{'_values'}->{ $MAPPING{$nm} } = $self->{'_last_data'};
}
}
else {
#$self->debug("blast.pm: unknown nm $nm, ignoring\n");
}
$self->{'_last_data'} = ''; # remove read data if we are at
# end of an element
$self->{'_result'} = $rc if ( defined $type && $type eq 'result' );
$self->{'_seen_hsp_features'} = 0;
return $rc;
}
=head2 element
Title : element
Usage : $eventhandler->element({'Name' => $name, 'Data' => $str});
Function: Convenience method that calls start_element, characters, end_element
Returns : none
Args : Hash ref with the keys 'Name' and 'Data'
=cut
sub element {
my ( $self, $data ) = @_;
# Note that start element isn't needed for character data
# Not too SAX-y, though
#$self->start_element($data);
$self->characters($data);
$self->end_element($data);
}
=head2 characters
Title : characters
Usage : $eventgenerator->characters($str)
Function: Send a character events
Returns : none
Args : string
=cut
sub characters {
my ( $self, $data ) = @_;
if ( $self->in_element('hsp')
&& $data->{'Name'} =~ /^Hsp\_(qseq|hseq|midline)$/ )
{
$self->{'_last_hspdata'}->{ $data->{'Name'} } .= $data->{'Data'}
if defined $data->{'Data'};
}
return unless ( defined $data->{'Data'} && $data->{'Data'} !~ /^\s+$/ );
$self->{'_last_data'} = $data->{'Data'};
}
=head2 within_element
Title : within_element
Usage : if( $eventgenerator->within_element($element) ) {}
Function: Test if we are within a particular element
This is different than 'in' because within can be tested
for a whole block.
Returns : boolean
Args : string element name
See Also: L<in_element>
=cut
sub within_element {
my ( $self, $name ) = @_;
return 0
if ( !defined $name && !defined $self->{'_elements'}
|| scalar @{ $self->{'_elements'} } == 0 );
foreach ( @{ $self->{'_elements'} } ) {
if ( $_ eq $name ) {
return 1;
}
}
return 0;
}
=head2 in_element
Title : in_element
Usage : if( $eventgenerator->in_element($element) ) {}
Function: Test if we are in a particular element
This is different than 'within_element' because within
can be tested for a whole block.
Returns : boolean
Args : string element name
See Also: L<within_element>
=cut
sub in_element {
my ( $self, $name ) = @_;
return 0 if !defined $self->{'_elements'}->[0];
return ( $self->{'_elements'}->[0] eq $name );
}
=head2 start_document
Title : start_document
Usage : $eventgenerator->start_document
Function: Handle a start document event
Returns : none
Args : none
=cut
sub start_document {
my ($self) = @_;
$self->{'_lasttype'} = '';
$self->{'_values'} = {};
$self->{'_result'} = undef;
$self->{'_elements'} = [];
}
=head2 end_document
Title : end_document
Usage : $eventgenerator->end_document
Function: Handles an end document event
Returns : Bio::Search::Result::ResultI object
Args : none
=cut
sub end_document {
my ( $self, @args ) = @_;
#$self->debug("blast.pm: end_document\n");
return $self->{'_result'};
}
sub write_result {
my ( $self, $blast, @args ) = @_;
if ( not defined( $self->writer ) ) {
$self->warn("Writer not defined. Using a $DEFAULT_BLAST_WRITER_CLASS");
$self->writer( $DEFAULT_BLAST_WRITER_CLASS->new() );
}
$self->SUPER::write_result( $blast, @args );
}
sub result_count {
my $self = shift;
return $self->{'_result_count'};
}
sub report_count { shift->result_count }
=head2 inclusion_threshold
Title : inclusion_threshold
Usage : my $incl_thresh = $isreb->inclusion_threshold;
: $isreb->inclusion_threshold(1e-5);
Function: Get/Set the e-value threshold for inclusion in the PSI-BLAST
score matrix model (blastpgp) that was used for generating the reports
being parsed.
Returns : number (real)
Default value: $Bio::SearchIO::IteratedSearchResultEventBuilder::DEFAULT_INCLUSION_THRESHOLD
Args : number (real) (e.g., 0.0001 or 1e-4 )
=cut
=head2 max_significance
Usage : $obj->max_significance();
Purpose : Set/Get the P or Expect value used as significance screening cutoff.
This is the value of the -signif parameter supplied to new().
Hits with P or E-value above this are skipped.
Returns : Scientific notation number with this format: 1.0e-05.
Argument : Scientific notation number or float (when setting)
Comments : Screening of significant hits uses the data provided on the
: description line. For NCBI BLAST1 and WU-BLAST, this data
: is P-value. for NCBI BLAST2 it is an Expect value.
=cut
=head2 signif
Synonym for L<max_significance()|max_significance>
=cut
=head2 min_score
Usage : $obj->min_score();
Purpose : Set/Get the Blast score used as screening cutoff.
This is the value of the -score parameter supplied to new().
Hits with scores below this are skipped.
Returns : Integer or scientific notation number.
Argument : Integer or scientific notation number (when setting)
Comments : Screening of significant hits uses the data provided on the
: description line.
=cut
=head2 min_query_length
Usage : $obj->min_query_length();
Purpose : Gets the query sequence length used as screening criteria.
This is the value of the -min_query_len parameter supplied to new().
Hits with sequence length below this are skipped.
Returns : Integer
Argument : n/a
=cut
=head2 best_hit_only
Title : best_hit_only
Usage : print "only getting best hit.\n" if $obj->best_hit_only;
Purpose : Set/Get the indicator for whether or not to process only
: the best BlastHit.
Returns : Boolean (1 | 0)
Argument : Boolean (1 | 0) (when setting)
=cut
=head2 check_all_hits
Title : check_all_hits
Usage : print "checking all hits.\n" if $obj->check_all_hits;
Purpose : Set/Get the indicator for whether or not to process all hits.
: If false, the parser will stop processing hits after the
: the first non-significance hit or the first hit that fails
: any hit filter.
Returns : Boolean (1 | 0)
Argument : Boolean (1 | 0) (when setting)
=cut
# general private method used to make minimal hits from leftover
# data in the hit table
sub _cleanup_hits {
my ($self, $hits) = @_;
while ( my $v = shift @{ $hits }) {
next unless defined $v;
$self->start_element( { 'Name' => 'Hit' } );
my $id = $v->[2];
my $desc = $v->[3];
$self->element(
{
'Name' => 'Hit_id',
'Data' => $id
}
);
my ($gi, $acc, $version ) = $self->_get_seq_identifiers($id);
$self->element(
{
'Name' => 'Hit_accession',
'Data' => $acc
}
);
if ( defined $v ) {
$self->element(
{
'Name' => 'Hit_signif',
'Data' => $v->[0]
}
);
if (exists $self->{'_wublast'}) {
$self->element(
{
'Name' => 'Hit_score',
'Data' => $v->[1]
}
);
} else {
$self->element(
{
'Name' => 'Hit_bits',
'Data' => $v->[1]
}
);
}
}
$self->element(
{
'Name' => 'Hit_def',
'Data' => $desc
}
);
$self->end_element( { 'Name' => 'Hit' } );
}
}
1;
__END__
Developer Notes
---------------
The following information is added in hopes of increasing the
maintainability of this code. It runs the risk of becoming obsolete as
the code gets updated. As always, double check against the actual
source. If you find any discrepencies, please correct them.
[ This documentation added on 3 Jun 2003. ]
The logic is the brainchild of Jason Stajich, documented by Steve
Chervitz. Jason: please check it over and modify as you see fit.
Question:
Elmo wants to know: How does this module unmarshall data from the input stream?
(i.e., how does information from a raw input file get added to
the correct Bioperl object?)
Answer:
This answer is specific to SearchIO::blast, but may apply to other
SearchIO.pm subclasses as well. The following description gives the
basic idea. The actual processing is a little more complex for
certain types of data (HSP, Report Parameters).
You can think of blast::next_result() as faking a SAX XML parser,
making a non-XML document behave like its XML. The overhead to do this
is quite substantial (~650 lines of code instead of ~80 in
blastxml.pm).
0. First, add a key => value pair for the datum of interest to %MAPPING
Example:
'Foo_bar' => 'Foo-bar',
1. next_result() collects the datum of interest from the input stream,
and calls element().
Example:
$self->element({ 'Name' => 'Foo_bar',
'Data' => $foobar});
2. The element() method is a convenience method that calls start_element(),
characters(), and end_element().
3. start_element() checks to see if the event handler can handle a start_xxx(),
where xxx = the 'Name' parameter passed into element(), and calls start_xxx()
if so. Otherwise, start_element() does not do anything.
Data that will have such an event handler are defined in %MODEMAP.
Typically, there are only handler methods for the main parts of
the search result (e.g., Result, Iteration, Hit, HSP),
which have corresponding Bioperl modules. So in this example,
there was an earlier call such as $self->element({'Name'=>'Foo'})
and the Foo_bar datum is meant to ultimately go into a Foo object.
The start_foo() method in the handler will typically do any
data initialization necessary to prepare for creating a new Foo object.
Example: SearchResultEventBuilder::start_result()
4. characters() takes the value of the 'Data' key from the hashref argument in
the elements() call and saves it in a local data member:
Example:
$self->{'_last_data'} = $data->{'Data'};
5. end_element() is like start_element() in that it does the check for whether
the event handler can handle end_xxx() and if so, calls it, passing in
the data collected from all of the characters() calls that occurred
since the start_xxx() call.
If there isn't any special handler for the data type specified by 'Name',
end_element() will place the data saved by characters() into another
local data member that saves it in a hash with a key defined by %MAPPING.
Example:
$nm = $data->{'Name'};
$self->{'_values'}->{$MAPPING{$nm}} = $self->{'_last_data'};
In this case, $MAPPING{$nm} is 'Foo-bar'.
end_element() finishes by resetting the local data member used by
characters(). (i.e., $self->{'_last_data'} = '';)
6. When the next_result() method encounters the end of the Foo element in the
input stream. It will invoke $self->end_element({'Name'=>'Foo'}).
end_element() then sends all of the data in the $self->{'_values'} hash.
Note that $self->{'_values'} is cleaned out during start_element(),
keeping it at a resonable size.
In the event handler, the end_foo() method takes the hash from end_element()
and creates a new hash containing the same data, but having keys lacking
the 'Foo' prefix (e.g., 'Foo-bar' becomes '-bar'). The handler's end_foo()
method then creates the Foo object, passing in this new hash as an argument.
Example: SearchResultEventBuilder::end_result()
7. Objects created from the data in the search result are managed by
the event handler which adds them to a ResultI object (using API methods
for that object). The ResultI object gets passed back to
SearchIO::end_element() when it calls end_result().
The ResultI object is then saved in an internal data member of the
SearchIO object, which returns it at the end of next_result()
by calling end_document().
(Technical Note: All objects created by end_xxx() methods in the event
handler are returned to SearchIO::end_element(), but the SearchIO object
only cares about the ResultI objects.)
(Sesame Street aficionados note: This answer was NOT given by Mr. Noodle ;-P)