Hi All
Here in Oxford are in the process of migrating to srs5; we are currently
running both srs4.08 and srs5.05. One of the primary justifications for
duplicating the SRS indexes locally is the ability to use getz for
heavy-duty database manipulation, generating specialised blast databases
and most importantly in hundreds of scripts to retrieve sequences
automatically from or for various analyses.
My problem with srs5 is because the syntax of getz has radically
changed; there are 100's of scripts that will break when we migrate.
Gratuitous example of srs4 and srs5 getz syntax:
SRS4:
getz -sf fasta -f "def" -d -l embl -l emnew -l genbank -l genbanknew \
"[sequence-sl#1:] !([sequence-org:metazoa]|[sequence-org:escherich*])";
SRS5.05:
getz -sf fasta -f "des seq" \
"[libs={embl emnew genbankgenbanknew}-sl#1:] \
!([libs-org metazoa]|[libs-org:escherich*])";
(The outputs are not identical as the srs4 output has an extra
"entry: EMBL:
.." output field)
Has anyone started to or considered writing a wrapper for getz to
convert srs4 syntax to srs5?
Is there a general interest in such a wrapper?
Are there plans to make SRS5 syntax backwardly compatible thus negating
the need for the wrapper?
One more question
how do you retrieve the annotation part of an entry
using SRS5 getz? I am currently using
SRS4:
getz -t '[embl-id:ecreca]'
SRS5:
getz -sf fasta -e '[embl-id:ecreca]' | sed '/^>/,$ d'
I am sure there is a better way ...
Many thanks
John
_____________________________________________________________________
John Peden | johnP at molbiol.ox.ac.uk
Molecular Biology Data Centre | Tel. 01865 2-22992
Oxford University | If its not broke -- then don't fix it
----------------------------------------------------------------------
