[Bioperl-l] hmmer3.pm question re query and hit coordinates

[Chris Fields]

This might be a disconnect between the HMMER3 version in bioperl-live and the one in Kai's bioperl-hmmer3 repo.  I believe the one in bioperl-live is newer.  Scott, can you give that a try?

chris

On Jul 19, 2011, at 9:45 PM, Thomas Sharpton wrote:

> Hi Scott,
> 
> Thanks for writing. I'm on the road at the moment so I have to be briefer and less thorough than I'd like to be.
> 
> What you are observing is not the intended behavior. Oddly, it's not what I recall obtaining in my tests on this software, though I was mostly interested in hmmsearch at the time and may have been sloppier than I should have been when it came to hmmscan.
> 
> What version of HMMER3 you're using? There have been some small formatting changes in the past that might be causing a burp in the parser, though I'm doubting it.
> 
> Kai Blin wrote some test scripts (found here: bioperl-live/t/SearchIO/hmmer.t) that, if I recall correctly, evaluate query/hit coordinates. It might be worth giving this a shot if you haven't already.
> 
> Also, if you don't mind, I'm happy to run your code on your output file on my end. It might help me diagnose the problem.
> 
> Sorry this is being a thorn in your side! I've cc'ed the list in case anyone else has insight into this matter.
> 
> Best,
> Thomas
> 
> On Jul 19, 2011, at 10:43 AM, Givan, Scott A. wrote:
> 
>> Hi Thomas,
>> 
>> I'm using modules in the bipoerl-hmmer3 git repository to parse hmmscan
>> reports. When I parse the files and walk through the HSP's like:
>> 
>> while (my $hit = $rslt->next_model) {
>> 
>> 	while (my $domain = $hit->next_hsp) {
>> 
>> And retrieve the "hit" coordinates like:
>> 
>> 		print "hit coords: ", $domain->start('hit'), "-", $domain->end('hit'),
>> "\n";
>> 
>> The coordinates returned correspond to what I would call the "query",
>> since they are for the sequence I fed to hmmscan to search the profile
>> database. Likewise, when retrieving the query coordinates like
>> $domain->start('query'), I get what I consider the "hit" coordinates,
>> since they are for the domain profile. Is this the intended behavior?
>> 
>> Thanks.
>> 
>> scott
>> 
>> -- 
>> Scott A. Givan
>> Associate Director
>> Informatics Research Core Facility
>> 240e Bond Life Sciences Center
>> Research Assistant Professor
>> Molecular Microbiology and Immunology
>> University of Missouri, Columbia
>> 
>> TEL 573-882-2948
>> FAX 573-884-9676
>> http://ircf.rnet.missouri.edu
>> 
>> 
>> 
> 
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/bioperl-l