Re: [Gusdev-gusdev] format of sequences for blast

[Steve F. <sfi...@pc...>]

The idea of the plugin is to store in gus a "similarity" association 
between each of the seqs you blasted and the seqs each one hit.  It only 
makes sense if the seqs you blasted against (the blastable db) are 
themselves also in gus.

In blast parlance, the seq you blast is the "query" and the seqs it hits 
in the blastable db are the "subjects".

so, the query table in your case is ExternalNaSequence.

I don't know how you made your blastable db, but, whatever table they 
came out of (and, they must have ids in their definition lines), is the 
subject table

steve

MICHAEL LUCHTAN wrote:

>Hello --
>
>I used the following script to get some data out of GUS:
>#!/usr/bin/perl
>
>use GUS::ObjRelP::DbiDatabase;
>use GUS::Common::GusConfig;
>
>my $gusconfig=
>GUS::Common::GusConfig->new("/home/luchtan/gus.properties");
>my $db= GUS::ObjRelP::DbiDatabase->new($gusconfig->getDbiDsn(),
>
>$gusconfig->getReadOnlyDatabaseLogin(),
>
>$gusconfig->getReadOnlyDatabasePassword,
>                                        1,1,1,
>                                        $gusconfig->getCoreSchemaName);
>my $dbh= $db->getQueryHandle();
>
>my $idStmt= $dbh->prepare("select '>'||na_sequence_id||' '||name||'\n'
>name_str,sequence from dots.externalnasequence");
>$idStmt->execute();
>open(TEMP, ">retrievedSeqForBLAST.tmp");
>while(my (@row)=$idStmt->fetchrow_array()){
>    print TEMP "@row\n";
>}
>close(TEMP);
>
>
>and blasted the resulting retrievedSeqForBLAST file.  Now I want to load
>the results into GUS using the plugin LoadBlastSimilaritiesPK.
>This plugin requires a subject_table and a query_table.  Perhaps it is a
>question of my unfamiliarity of the whole biology aspect of what is going
>on here(I'm just a programmer), but I am not sure what these tables are
>for.  Can anybody help me
>out?
>
>Michael Luchtan
>http://www.cs.uga.edu/~luchtan
>
>
>On Thu, 22 May 2003, Jonathan Crabtree wrote:
>
>  
>
>>Jessie-
>>
>>Jessica Kissinger wrote:
>>    
>>
>>>    If memory serves me correct, once sequences were loaded into GUS, we
>>>then retrieved them along with their GUS ID to submit for blast searches.
>>>      
>>>
>>Yes, I think that's right.
>>
>>    
>>
>>>    When we retrieved the sequences, we created a custom format for the
>>>header line, such that the blast results once generated for these
>>>sequences could be easily parsed and loaded with the existing plug-in.
>>>
>>>    Can someone tell me what the format of the fasta header should be,
>>>i.e is it ">GUSID, External_NA _sequence Name" or the other way around
>>>and should there be any formatting, tabs, spaces etc.  If I remember
>>>correctly, the blast results were loaded by GUSID not "name", but I
>>>don't remember.
>>>      
>>>
>>My recollection is that the defline started as you said, with ">GUSID ".
>>I don't believe that the format is crucial, because (again, from what
>>I remember) when you run the plugin to load the BLAST similarities you
>>supply it with a regular expression that it uses to pick the GUSID
>>(an na_sequence_id for most of the PlasmoDB searches) out of the defline.
>>So as long as the regex matches the defline format, you should be OK,
>>and I don't think that the plugin uses anything on the defline except
>>for the GUSID.
>>
>>Jonathan
>>
>>
>>
>>
>>
>>-------------------------------------------------------
>>This SF.net email is sponsored by: ObjectStore.
>>If flattening out C++ or Java code to make your application fit in a
>>relational database is painful, don't do it! Check out ObjectStore.
>>Now part of Progress Software. http://www.objectstore.net/sourceforge
>>_______________________________________________
>>Gusdev-gusdev mailing list
>>Gus...@li...
>>https://lists.sourceforge.net/lists/listinfo/gusdev-gusdev
>>
>>    
>>
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: ObjectStore.
>If flattening out C++ or Java code to make your application fit in a
>relational database is painful, don't do it! Check out ObjectStore.
>Now part of Progress Software. http://www.objectstore.net/sourceforge
>_______________________________________________
>Gusdev-gusdev mailing list
>Gus...@li...
>https://lists.sourceforge.net/lists/listinfo/gusdev-gusdev
>  
>