Re: [Gusdev-gusdev] Loading BLAST results

[Pablo N. M. <pa...@pa...>]

Hi Arnaud,
I don't know if I'll be able to answer all your questions, but I'll try
some to the extent of my knowledge on GUS up to now.

> I have the sequences in FASTA, what regexp shall I need?
> Is only --regex_name parameter required?
> Does it matter if I don't give taxon_id parameter ?

I've run the plugin with these parameters:

[pablo@mkiwi mcl]$ ga GUS::Common::Plugin::InsertNewExternalSequences
--external_database_release_id=3D38
--regex_source_id=3D(.*)
--table_name=3DDoTS::ExternalAASequence
--sequencefile=3Dtbrucei =96commit

The regex_*, as you probably noted, are the regular expressions to
extract the referred info from the FASTA header (name, source_id,
secondary_id, etc.). I've only used the regex_source_id. Also, it seems
not to matter if you don't give the taxon_id parameter. But you
obviously won't make the associations in DoTS::AASequenceTaxon between
sequences and taxa.

> Does LoadBlastSimFast module require generateBlastSimilarity.pl
script? Where can I get this script ?

This module reads similarity results in an especific format like:

>479679 (3 subjects)
  Sum: 479680:1871:1.2e-194:1:353:1:353:1:353:353:353:0:
   HSP1: 479680:353:353:353:1871:1.2e-194:1:353:1:353:0:
  Sum: 488460:1826:7.0e-190:1:353:1:353:1:353:342:348:0:
   HSP1: 488460:342:348:353:1826:7.0e-190:1:353:1:353:0:
>479680 (3 subjects)
  Sum: 479679:1871:1.2e-194:1:353:1:353:1:353:353:353:0:
   HSP1: 479679:353:353:353:1871:1.2e-194:1:353:1:353:0:

The script parseBlastFilesForSimilarity.pl (attached) will do the trick.
I don't know if there are multiple versions of this script traveling
around the list.

[jdai@headnode mclorth]$ ls /scratch/jdai/Cpgus_vs_Pfgus/ | perl
parseBlastFilesForSimilarity.pl
--regex=3D'(\S+)'
--outputFile=3DLm_vs_Lm_parsed=20
--dir=3D/scratch/jdai/Cpgus_vs_Pfgus/

Hope this is useful,
Pablo

On Thu, 2004-08-05 at 08:49, Arnaud Kerhornou wrote:
> Hi
>=20
> I want to load BLAST results in GUS.
> Before running LoadBlastSimFast module, I want to load Uniprot and EMBL=
=20
> databases in DoTS::ExternalAASequence and DoTS::ExternalNASequence.,=20
> just the Ids not the sequences themselves.
>=20
> I need some help to use GUS::Common::Plugin::InsertNewExternalSequences=
=20
> plugin.
>=20
> I have the sequences in FASTA, what regexp shall I need  ? Is only =20
> --regex_name parameter required ? I don't know the taxon attached to th=
e=20
> sequence entries. Does it matter if I don't give taxon_id parameter ?
>=20
> Re. LoadBlastSimFast module, It seems to parse the output in a specific=
=20
> format. Does it require generateBlastSimilarity.pl script ? Where can I=
=20
> get this script ?
>=20
> cheers
> Arnaud
--=20
-----------------------------
Pablo Nascimento Mendes
CTEGD EMF TIPS Fellow
Kissinger Lab
Department of Genetics
University of Georgia
C210 Life Sciences Bldg.
Athens, Georgia 30602
Phone:706 542-1447
E-mail: pa...@ug...