[Bioperl-l] Trouble in executing StandAloneBlast on Win2000

Mon Feb 24 09:30:12 EST 2003

Hello,

I installed Bioperl 1.2 along with WrapperBase and tried Blast search again,
but the same result occurred. It took 38 hours for 30,000 search,  initially
search speed was 284 search/min and finally 7 search/min. The first 10,000
search took 5 hours, the second 10,000 search took 12 hours and the last
10,000 search took 21 hours.

In the performance monitor of Win2000, perl process ocuppies approx. 90% of
CPU and blastall.exe is sometimes started and finishes immediately.
Apparently this is not a problem of blastall.exe.

I don't know what I should do next and can anyone help me please?

Takeshi Sasayama

----- Original Message -----
From: "Jason Stajich" <jason at cgt.mc.duke.edu>
To: "Takeshi Sasayama" <sasayama at lmi.hitachi-sk.co.jp>
Cc: "Brian Osborne" <brian_osborne at cognia.com>; <bioperl-l at bioperl.org>
Sent: Thursday, February 20, 2003 11:28 AM
Subject: Re: [Bioperl-l] Trouble in executing StandAloneBlast on Win2000

> See http://bioperl.org/Bugs/ for WrapperBase.pm which was missing from
> 1.2.0  -- it will be included properly in 1.2.1 release in a few days.
>
>
> -jason
> On Thu, 20 Feb 2003, Takeshi Sasayama wrote:
>
> > Hi Brian,
> >
> > Thank you very much for your suggestion. The Win2000 PC has 1 hard
> > disk(115GB) and there is 94.5GB free space. I didn't check free disk
space
> > during the search  because there was still 800MB unused physical memory
and
> > I thought this is not virtual memory trouble due to disk space shortage.
> >
> > Upgrading to 1.2 may be good solution. Actually I once upgraded to
Ver.1.2
> > on Linux machine to test possibility to avoid another bug (which is
already
> > posted bug 1156 and 1240) but failed in running the script (Err like
> > "WrapperBase was not found" was shown. So I thought there might have
been
> > much change between 1.0.2 and 1.2, and decided to use 1.0.2 and
reinstalled.
> > The multiple database bug 1156 seemed not to be fixed in 1.2 and I don't
> > know if this trouble is fixed in 1.2, but maybe I should try Ver.1.2
again.
> >
> > Thanks
> > Takeshi Sasayama
> >
> > ----- Original Message -----
> > From: "Brian Osborne" <brian_osborne at cognia.com>
> > To: "Takeshi Sasayama" <sasayama at lmi.hitachi-sk.co.jp>
> > Cc: <bioperl-l at bioperl.org>
> > Sent: Thursday, February 20, 2003 10:28 AM
> > Subject: Re: [Bioperl-l] Trouble in executing StandAloneBlast on Win2000
> >
> >
> > > Takeshi,
> > >
> > > What is happening to your free disk space as the search progresses?
The
> > > standard answer is upgrade to version 1.2 but in this case the answer
> > > is a good one, there is a known StandAloneBlast bug in older Bioperl's
> > > that sounds a lot like your's.
> > >
> > > Brian O.
> > >
> > >
> > > On Thursday, February 20, 2003, at 08:44 AM, Takeshi Sasayama wrote:
> > >
> > > > Hi, all
> > > >
> > > >
> > > > I'm a newbie in Bioperl and I have a problem in executing Blast
search
> > > > on
> > > > Windows 2000 using StandAloneBlast. During about 30,000 Blast
search,
> > > > the
> > > > search had slowed down gradually as the search went on.
> > > >
> > > > What I am doing is to execute blastn search for about 30,000 times
in a
> > > > loop. The query file is one file of FASTA format (including about
> > > > 30,000
> > > > sequences of all 50mer length) and the database contains 27,000
> > > > entries.
> > > > Each search result is saved in a directory as a file named with its
> > > > sequence
> > > > name.
> > > >
> > > > At first I ran the program on Linux machine (Red Hat Linux 7.2,
PenIII
> > > > 500MHz, 128MB RAM, perl v5.6.0, Bioperl-1.0.2, NCBI-Blast2.2.5(linux
> > > > binary)) and finished successfully in 6 hours. After that I modified
> > > > the
> > > > same program (only PATH handling) and ran on Windows machine
(Windows
> > > > 2000
> > > > professional(SP3), Pentium4 2.53GHz, 1024MB RAM, ActivePerl
> > > > v5.6.1(build
> > > > 628), Bioperl-1.0.2, NCBI-Blast2.2.5(windows binary)). This was
> > > > reproducible
> > > > on the same PC and on another Win 2000 PC. I thought that some kind
of
> > > > resource was exhausted but the memory was left free enough(800MB). I
> > > > don't
> > > > know what causes this trouble .
> > > >
> > > > Does anyone have similar experience?
> > > > Any comments, suggestions, ideas are very much welcome. The script
> > > > that I
> > > > wrote is essentially below.
> > > >
> > > > Thanks in advance!
> > > > Takeshi Sasayama
> > > >
> > > > ###### Code starts here ######
> > > > BEGIN {
> > > > $ENV{BLASTDIR} = 'C:\\blast';
> > > > $ENV{BLASTDATADIR} = 'C:\\db\\Ensembl\\10.30';
> > > > }
> > > >
> > > > use Bio::Tools::Run::StandAloneBlast;
> > > > use Bio::Seq;
> > > > use Bio::SeqIO;
> > > > use strict;
> > > >
> > > > my $database_file = 'Homo_sapiens.cdna.fa';
> > > > my $queryfile = queryfile.fa';
> > > >
> > > > # parameters
> > > > my $expect_value = 1;
> > > > my $filter_query_sequence = 'F';
> > > > my $one_line_description = 100;
> > > > my $alignments = 100;
> > > > my $strands = 1;
> > > > my @params = ('program' => 'blastn', 'database' => $database_file);
> > > > my $progress_interval = 100;
> > > >
> > > > # create factory object and set parameters
> > > > my $factory = Bio::Tools::Run::StandAloneBlast->new(@params);
> > > > $factory->e($expect_value);
> > > > $factory->F($filter_query_sequence);
> > > > $factory->v($one_line_description);
> > > > $factory->b($alignments);
> > > > $factory->S($strands);
> > > >
> > > > # get query
> > > > my ($in, $queryseq, $blast_report, $outfile);
> > > > $in = Bio::SeqIO->new( -file => $queryfile, -format => 'Fasta');
> > > > while ($queryseq = $in->next_seq()) {
> > > >  $outfile = 'outfiledir'."\\".$queryseq->id;
> > > >  $factory->outfile($outfile);
> > > >  $blast_report = $factory->blastall($queryseq);
> > > > }
> > > > ###### Code ends here ######
> > > >
> > > > _______________________________________________
> > > > Bioperl-l mailing list
> > > > Bioperl-l at bioperl.org
> > > > http://bioperl.org/mailman/listinfo/bioperl-l
> > > >
> >
> > _______________________________________________
> > Bioperl-l mailing list
> > Bioperl-l at bioperl.org
> > http://bioperl.org/mailman/listinfo/bioperl-l
> >
>
> --
> Jason Stajich
> Duke University
> jason at cgt.mc.duke.edu