submit blast script for creating submissions that are split

README

@@ -0,0 +1,2 @@
+The scripts in here are intended as either standalone or examples for job submissions on the UCR bioinformatics cluster (biocluster.ucr.edu)
+

make-jobs/submit_blast_simple.pl

@@ -0,0 +1,135 @@
+#!/usr/bin/perl -w
+use strict;
+use File::Spec;
+use Bio::SeqIO;
+use Getopt::Long;
+
+
+my $home = $ENV{'HOME'};
+my $jobdir = File::Spec->catfile($home,'jobs');
+my $input  = File::Spec->catfile($home,'input');
+my $outdir = File::Spec->catfile($home,'output');
+my $tmpdir = File::Spec->catfile($home, 'tmp');
+my $homebase = 'head3';
+my $qsize = 5000;
+my ($query,$dbin,$params,$exe);
+$exe = 'blastall'; # default
+my $formatdb = 'formatdb';
+my $prefix = 'blast';
+my $usescript = 0;
+$params = '-p blastp -e 1e-3 -m 8';
+GetOptions(
+	   'i|q|query:s' => \$query,
+	   'd|db:s'      => \$dbin,
+	   'p|params:s'  => \$params,
+	   's|size:i'    => \$qsize,
+	   'usescript!'  => \$usescript,
+	   'e|exe:s'     => \$exe,
+	   'indir|input:s'  => \$input,
+	   'o|out|outdir:s' => \$outdir,
+	   'j|job|jobdir:s' => \$jobdir,
+	   't|tmp|tmpdir:s' => \$tmpdir,
+	   );
+die( "need a dbname\n" ) unless (defined $dbin );
+my @dblst;
+for my $dbpart ( split(/\s+/,$dbin) ) {
+    my $rel = File::Spec->rel2abs($dbpart);
+    my (undef,undef,$dbname) = File::Spec->splitpath($rel);
+    push @dblst, $rel;# File::Spec->catfile($scratchdir,$prefix,$dbname);    
+} 
+my $db = join(" ", @dblst);
+warn("db is $db\n");
+die( "need an inputfile\n" ) unless defined $query && -e $query;
+die( "need some params\n" ) unless defined $params;
+
+
+for my $dir( $jobdir,$input,$outdir,$tmpdir ) {
+    mkdir $dir unless -d $dir;
+}
+
+$query = File::Spec->rel2abs($query);
+my (undef,undef,$query_file) = File::Spec->splitpath($query);
+my $uid = $query_file;
+my $in = Bio::SeqIO->new(-file   => $query,
+			 -format => "fasta");
+
+my ($ct,$setct) = (0,0);
+my $infile = File::Spec->catfile($input,"$uid.$setct.fa");
+my $input_seqfh = Bio::SeqIO->new(-format => 'fasta',
+				  -file   => ">$infile");
+
+my @infiles = ($infile);
+my @infilenames = ("$uid.$setct.fa");
+
+while( my $s = $in->next_seq ) {
+	next if $s->seq =~ /^[Xx]+$/;
+    $input_seqfh->write_seq($s);
+    if( ++$ct >= $qsize ) {
+	$input_seqfh->close();
+	$setct++;
+	$infile = File::Spec->catfile($input,"$uid.$setct.fa");
+	$input_seqfh = Bio::SeqIO->new(-format => 'fasta',
+				       -file   => ">$infile");
+	push @infiles, $infile;
+	push @infilenames, "$uid.$setct.fa";
+	$ct = 0;
+    }
+}
+
+
+for( my $i = 0; $i <= $setct; $i++ ) {
+    my $jobfile = File::Spec->catfile($jobdir,"$uid.$i.sh");
+    open my $fh => ">$jobfile" || die $!;
+    chmod(0700,$jobfile);
+    my $ofile = File::Spec->catfile($outdir,"$uid.$i.output");
+    my $out   = File::Spec->catfile($tmpdir,"$uid.$i.out");
+    my $ifile = $infiles[$i];#File::Spec->catfile($wkdir,$infilenames[$i]);
+
+    print $fh ( "#!/bin/bash\n",
+		'#PBS'." -N $prefix.$i\n",
+		"source ~/.bash_profile\n",
+		);
+
+    print $fh "$exe $params -i $ifile -d \"$db\" -o $ofile\n";    
+
+}
+
+__END__
+
+=head1 NAME
+
+submit_blast_simple.pl - a script for submitting BLAST jobs for parallel submission by splitting the input into chunks
+
+=head1 SYNOPSIS
+
+submit_blast_simple.pl -i peptides.fa -d blast_database -s 1000
+
+submit_blast_simple.pl -i queryscaffolds.fa -d genome -s 1000 --params "-p blastn -e 1e-5 -m 8"
+
+=head1 DESCRIPTION
+
+Assumes you will be running NCBI blast (blastall) and the default
+parameter options are:
+
+ -p blastp -e 1e-3 -m 8
+
+It will create the directories
+ ~/jobs ~/output ~/input
+
+The job files will be saved in ~/jobs and you will submit them like
+
+ for job in jobs/*.sh
+ do
+  qsub jobs/$job
+ done
+
+The output will be written to ~/output. When your jobs are done you
+can safely remove the job files and input files saved in ~/jobs and
+~/input respectively.
+
+
+=head1 FEEDBACK
+
+jason.stajich [at] ucr.edu
+
+=cut