[Bioperl-l] Finding significant hits with BPlite

Hilmar Lapp hlapp@gmx.net
Mon, 11 Sep 2000 13:02:57 +0200

gert thijs wrote:
> Just, before I went on holidays myself, I found what was causing the problem.
> In "Bio::Tools::BPlite::Sbjct::nextHSP" there is some code to parse a
> scoreline and to assign the value to $p.
> The code looks like:
> ($p) = $scoreline =~ /Expect = (\S+)/ ;
> But in a Blast report there can be another form than the "Expect = Score": eg.
> "Expect(2) = Score". So I only have to add an extra check for this kind of
> scoreline.

Interesting. The stack trace as reported in your initial mail indicated a
bug in Similarity.pm, which I fixed (i.e., there indeed was one)
meanwhile. The only condition this bug could have been exposed is by
retrieving $sbjct->significance() without having it set before, which I
did not worry about, but which indicates a bug in BPLite. So, the
condition that exposed the bug should now no longer trigger a throw() in
Similarity.pm (instead, you'll get undef returned), even though there's
almost certainly still a bug present in the BPLite code.


Hilmar Lapp                                email: hlapp@gmx.net
NFI Vienna, IFD/Bioinformatics             phone: +43 1 86634 631
A-1235 Vienna                                fax: +43 1 86634 727