Re: [Gusdev-gusdev] BlastAnal.pm and LoadBlastSimFast.pm

[Arnaud K. <ax...@sa...>]

Hi

Steve Fischer wrote:

> arnaud-
>
> i will make the change to BlastAnal and release a new .tar for the 
> CBIL code.
>
> go ahead and commit your fix to LoadBlastSimFast.pm
>
I have commited an updated version of LoadBlastSimFast plugin. It can 
also parse the real subject and query identifiers and get the primary 
keys associated with them.
It shouldn't affect the parsing if the primary keys are already in the 
BLAST formated output. If not let me know.

cheers
Arnaud

> steve
>
> Arnaud Kerhornou wrote:
>
>> Hi
>>
>> I've got a script from Pablo Mendes called 
>> parseBlastFilesForSimilarity.pl that uses CBIL::Bio::Blast::BlastAnal 
>> module part of the CBIL distribution. To make the blast parsing work 
>> for me (to parse the score), I had to do a minor modification, line 131:
>>    if (/... \s(\d+) .../) {
>> by
>>    if (/... \s+(\d+) .../) {
>>
>> ie match at least one space before parsing the score value, instead 
>> of matching one and only one space.
>>
>> Also, in the BLAST output, the query and subject ids are meant to be 
>> GUS Primary keys, I have instead the real subject and query identifiers.
>> In LoadBlastSimFast.pm, could we make the parsing more generic so it 
>> can accept both ? I have made the changes locally, I can commit an 
>> update if nobody minds.
>>
>> Shall I make sourceforge entries to log them ?
>>
>> cheers
>> Arnaud
>>