[Bioperl-l] Yet another 'what version's supposed to work?' question?

George Hartzell hartzell at kestrel.alerce.com
Wed Sep 28 18:24:51 EDT 2005


I'm digging into bioperl-run on FreeBSD.

I've built a port for exonerate-1.0.0, and it passes all of it's tests
(but not with --enable-assertions, I've asked Guy about that...).

t/Exonerate.t blows up all over the place, but I think that the
problem is that the cDNA<->genomic alignments that are coming back
aren't quite what the test expect to see.  They seem reasonable
though, leading and/or trailing hsp's seem to have been broken into
smaller parts.

Sim4 gives answers that generally agree with both the test expected
and the calculated data.

Actually, looking at the exonerate text output and what I'm seeing in
the perl debugger, it looks like some part of the current bioperl
parser is treating this:

    4 : GTTGTGCT-GGGGGGGTCCGCGCGCCCTGCAGTCCCGAACCCGCTACGCGCTCCGTCGGCCC :   64
        |||||||| |||||||||||||||||||||||||||||||||||||||||||||||||||||
   29 : GTTGTGCTGGGGGGGGTCCGCGCGCCCTGCAGTCCCGAACCCGCTACGCGCTCCGTCGGCCC :   90

As two hsps, one from 4->11/29->36, then 12->/38->....  The exonerate
expected data expects it as a single HSP.

Does anyone out there have any experience with the exonerate and how
well it works?

g.



More information about the Bioperl-l mailing list