[Bioperl-l] NCBI local blast module and psiblast parser now available

[Peter Schattner]

Hello all

I have committed an "alpha" version of my StandAloneBlast.pm bioperl
module to the bioperl-live (main) repository. There is a test driver
script t/StandAloneBlast.t that performs some simple functional checks
as well as an example script, examples/standaloneblast.pl. In addition
there are a couple of modules BPpsilite.pm and BPbl2seq.pm (both based
on the BPlite module) for parsing multiple-iteration psiblast reports
and two-sequence bl2seq reports respectively. There are also test
scripts ( t/BPpsilite.t and t/BPbl2seq) giving examples of usage of the
new parsers. 

The example script demonstrates using the StandAloneBlast module to
monitor the sensitivity of blastall and blastpgp searches to variations
in blast input parameters (eg gap penalties, choice of protein scoring
matrices).  The script also demonstrates comparing alignment-based blast
searches - either using alignments developed by multiple-iteration
psiblast or with a user supplied initial alignment.  

For those of you who have ncbi blast running locally (or want to install
it :-), I'd be happy to hear your thoughts on the utility of these
modules and how they might be improved.  I think they are all reasonably
well documented, but if you find otherwise or run into any bugs or
problems that would be very helpful information for me.  You will need
to have a copy of the blast programs on your machine (as well as
downloading the main trunk from the Bioperl CVS repository, of course). 
To compare results with those in the test scripts you will also need to
have downloaded and installed the swissprot and ecoli.nt databases from
NCBI. 

Even for those who only run remote blasts, the BPpsilite and BPbl2seq
parsers may be useful if you ever run multiple-iteration psiblast or
bl2seq queries.    

Thanks

Peter