[Bioperl-l] newbie tries StandAloneBlast and receives "cant openBLOSUM62"

[Chris Fields]

You're running a pretty old version of bioperl.  Bioperl 1.5.1 is the latest
and should work with BLAST output from v 2.2.13 (NCBI never upgraded the
formatting for netblast client and local blast, so it's text output is
exactly like v. 2.2.12; only web output has changed).  

I suggest upgrading to the latest CVS using the following instructions:

http://cvs.open-bio.org/cgi-bin/viewcvs/viewcvs.cgi/bioperl-live/

http://www.bioperl.org/wiki/Installing_Bioperl_on_Windows#Beyond_the_Core

Don't worry if 'nmake test' fails a number of times; it's a bit flaky with
bioperl-live.  This should install over the old version.  

If you're really wanting to stick with PPM and are especially brave, I put
up a VERY preliminary online method of making a custom PPM version of
bioperl:

http://www.bioperl.org/wiki/Create_a_Bioperl_PPM_Package

It's a bit longish. But the nice thing about this is you get very pretty,
browsable autogenerated HTML docs from POD for all of the installed modules,
including bioperl!  New versions of PPM do this automagically. 

If installing from CVS doesn't fix it drop a line here again.

Christopher Fields
Postdoctoral Researcher - Switzer Lab
Dept. of Biochemistry
University of Illinois Urbana-Champaign

> -----Original Message-----
> From: bioperl-l-bounces at lists.open-bio.org [mailto:bioperl-l-
> bounces at lists.open-bio.org] On Behalf Of Harald
> Sent: Wednesday, March 01, 2006 9:44 AM
> To: bioperl-l at lists.open-bio.org
> Subject: [Bioperl-l] newbie tries StandAloneBlast and receives "cant
> openBLOSUM62"
> 
> Hi all.
> 
> I am a Bioinformatics Newbie and want to use BioPerl for doing BLASTs on
> a local protein-sequence file. Unfortunately something went wrong, which
> gives me the error message: "...Unable to open BLOSUM62...".
> 
> I am using Win2000 on a normal Desktop PC with 1.4 GHz AMD and 256 MB
> RAM (I know, this is not much, but my sequence database is not so big).
> My perl-interpreter is from ActiveState and in the version 5.8.7 built
> for MSWin32-x86-multi-thread
> BioPerl is version 1.4
> Blast has version 2.2.13
> 
> 
> (1) I have downloaded the blast distribution from
> http://www.ncbi.nlm.nih.gov/blast/download.shtml
> 
> (2) I have installed it and put the file ncbi.ini into my
> WINNT-directory, which is pointing to blast-directory/data
> ==============C:/WINNT/ncbi.ini========================
> [NCBI]
> Data="D:/BLAST/BLAST-2.2.13/data"
> ===================================================
> 
> (3) I downloaded a FASTA-formatted protein-sequence database file
> (ftp://ftp.ncbi.nih.gov/pub/COG/KOG/kyva) [48MB]
> 
> (4) I used formatdb -i kyva and received the files kyva.phr, kyva.pin
> and kyva which I put into the directory, where my perl-script resides.
> 
> (5) Now I tried to BLAST this database for a sequence, I have copied out
> of it. Therefore I just modified the BLAST-script from the Beginner-HowTo:
> ==================D:/BLAST/StandAloneBlast.pl======================
> use strict;
> 
> use Bio::Seq;
> use Bio::Tools::Run::StandAloneBlast;
> 
> my @params = (program  => 'blastp', database => 'kyva' );
> 
> my $blast_obj = Bio::Tools::Run::StandAloneBlast->new(@params);
> 
> my $seq_obj = Bio::Seq->new(-id  =>"test_query", -seq
> =>"SYFICPISQEVMREPRVAADGFTYEAESLREWLDNGHETSPMTNLKLAHNNLVPNHALRSAIQEWLQRNS"
> );
> 
> 
> my $report_obj = $blast_obj->blastall($seq_obj);
> 
> my $result_obj = $report_obj->next_result;
> 
> print $result_obj->num_hits;
> 
> =============================================================
> 
> (6) But if I trigger this script, I receive:
> ==============================================================
> [NULL_Caption] WARNING:  [000.000]  test_query: Unable to open BLOSUM62
> [NULL_Caption] WARNING:  [000.000]  test_query: BlastScoreBlkMatFill
> returned no
> n-zero status
> [NULL_Caption] WARNING:  [000.000]  test_query: SetUpBlastSearch failed.
> 0
> ==============================================================
> 
> Does anyone has a clue, what is going wrong? Since my ncbi.ini - file
> points to the directory, where blosum62 is found, I can not understand
> this error-message.
> 
> With kind regards,
> frustrated Harald
> 
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/bioperl-l