Synchronisation Problem in GCG with PIR : Solution & description

PARSONS_A at snd01.pcr.co.uk PARSONS_A at snd01.pcr.co.uk
Thu Jan 7 05:58:17 EST 1993


Hi Netfolk,

David Mathog (Caltech) mailed me to post to the net what the solution to my
problem was, as had thanked all those who had helped me.  Reinhard Doelz
(Basel) and Liz Cowe (Oxford) both gave me the answer to fix the problem but Liz
also provided an in depth analysis of the cause - with her agreement I am
posting her analysis and am also alerting MIPS who provided the copy of PIR
in question.

Viva la (le?) Net!

Tony p.


==============================================================================
To: parsons_a at uk.co.pcr.snd01
Message-ID: <009662A5.93B2E8A0.15248 at vax.path.ox.ac.uk>
Subject: GCG syncoronization problem

Tony,

	I have also seen this problem with the PIR files that I get from MIPS
- is this where yours come from too ? The problem seems to be that an extra
character is added to the end of some sequence lines - this can't be seen (or
may appear as an odd control character) and the PIR programs all cope with it
fine. The GCG FASTA program decides the sequence and index don't match and so
stop - if you fetch the sequence that is dying (find this by runing fasta with
/monitor=1) then you will see an extra residue added at random to the end of
the sequence - well I did. But some GCG programs can skip this character so my
solution was to recreate the GCG index files - I first used Chopup so I could 
get short enough lines to check there were no funnies but actually thing this 
step isn't necessary and you can just use dataset and dbindex. 

	I tried getting help from GCG and PIR but both suggested that the 
funny character was being included by MIPS when they make the GCG index so i 
should contact them. By this time I'd done all the reformating so didn't 
bother!

Liz Cowe
Oxford University Molecular Biology Data Centre
liz at vax.path.ox.ac.uk




More information about the Bio-soft mailing list