Fwd: [Bioperl-l] parse Blast Output and Composition Based Statistics parameter

[Hubert Prielinger]

Hi Jason,
I have written to NCBI helpdesk if they can help me with further 
information...
that's the response:

Hello,

I'm sorry, this has recently changed. Instead of "Yes", try using either
'0' '1' or '2', where:

'0' = No Composition-based statistics
'1' = Conditional compositional score matrix adjustment (apply only to
'biased' sequences)
'2' = Universal compositional score matrix adjustment (apply to all).

This works with the URLAPI; I've not tested with the perl module.

Best regards,
Wayne



regards
Hubert



Jason Stajich wrote:

> sorry - i don't really have the time to support this module - lots of 
> people on the list use it so they can hopefully help.
>
> Begin forwarded message:
>
>> *From: *Hubert Prielinger <hubert.prielinger at gmx.at 
>> <mailto:hubert.prielinger at gmx.at>>
>> *Date: *January 16, 2006 3:45:21 PM EST
>> *To: *Jason Stajich <jason.stajich at duke.edu 
>> <mailto:jason.stajich at duke.edu>>
>> *Subject: **Re: [Bioperl-l] parse Blast Output and Composition Based 
>> Statistics parameter*
>>
>> Jason Stajich wrote:
>>
>>> (please don't try and post to bioperl-announce, it is not for  
>>> questions.)
>>>
>>> On Jan 12, 2006, at 6:57 PM, Hubert Prielinger wrote:
>>>
>>>> Hello,
>>>> I want to know, if there is a possibility to get from a Blast  
>>>> Outputfile the whole Sequence of a protein not only the best local  
>>>> alignment...
>>>> for example:
>>>>
>>> No. The parser can only return to you what is in the report file...
>>> use Bio::DB::GenPept to retrieve the sequence via the web or  
>>> (recommended) use a locally indexed sequence database like  
>>> Bio::DB::Fasta
>>>
>>>> >ref|XP_480077.1| hypothetical protein [Oryza sativa (japonica  
>>>> cultivar-group)]
>>>> dbj|BAD33542.1| hypothetical protein [Oryza sativa (japonica  
>>>> cultivar-group)]
>>>>         Length=95
>>>>
>>>> Score = 24.1 bits (47),  Expect =   493
>>>> Identities = 6/7 (85%), Positives = 7/7 (100%), Gaps = 0/7 (0%)
>>>>
>>>> Query  2   KKRRRWW  8
>>>>                K+RRRWW
>>>> Sbjct  87  KRRRRWW  93
>>>>
>>>> and now, if I parse the file, I want to get the whole Sequence of  
>>>> this hypothetical protein....is that possible with hsp for 
>>>> example,  or any other way....
>>>>
>>>> my second question is:
>>>> I do my blast search with bioperl and the remoteblast  
>>>> module.....each parameter is working very well, except the  
>>>> composition based statistics parameter....
>>>> it looks like that:
>>>>
>>>> my $factory = $Bio::Tools::Run::RemoteBlast::HEADER 
>>>> {'COMPOSITION_BASED_STATISTICS'} = 'yes';
>>>>
>>> uh no that is not how you would do it.
>>> You can make it the default for any factories you use in the script  
>>> by doing this
>>>
>>>> $Bio::Tools::Run::RemoteBlast::HEADER 
>>>> {'COMPOSITION_BASED_STATISTICS'} = 'yes';
>>>
>>>
>>> then
>>> $factory = Bio::Tools::Run::RemoteBlast->new();
>>>
>>>
>>>  =OR=
>>> Once you have a factory object you can set the parameter explicitly:
>>> $factory->submit_parameter('COMPOSITION_BASED_STATISTICS', 'yes');
>>>
>>>> it should work like that, but it doesn't....
>>>>
>>>> Thanks for your help in advance......
>>>>
>>>> regards
>>>> Hubert
>>>> _______________________________________________
>>>> Bioperl-l mailing list
>>>> Bioperl-l at portal.open-bio.org <mailto:Bioperl-l at portal.open-bio.org>
>>>> http://portal.open-bio.org/mailman/listinfo/bioperl-l
>>>
>>>
>>>
>>> -- 
>>> Jason Stajich
>>> Duke University
>>> http://www.duke.edu/~jes12 <http://www.duke.edu/%7Ejes12>
>>>
>>>
>>> _______________________________________________
>>> Bioperl-l mailing list
>>> Bioperl-l at portal.open-bio.org <mailto:Bioperl-l at portal.open-bio.org>
>>> http://portal.open-bio.org/mailman/listinfo/bioperl-l
>>>
>>>
>> Hi Jason,
>> I have tried everything that you suggested, but the Composition Based 
>> Statistic parameter isn't still working, every
>> other parameter works using e.g
>>
>> $Bio::Tools::Run::RemoteBlast::HEADER{'DESCRIPTIONS'} = '1000';
>>
>> thanks in advance
>> Hubert
>>
>
> --
> Jason Stajich
> Duke University
> http://www.duke.edu/~jes12 <http://www.duke.edu/%7Ejes12>
>
>