BioPerl project / Re: I need a parser for BLAST

Georg Fuellen fuellen at TechFak.Uni-Bielefeld.DE
Thu Mar 19 06:46:41 EST 1998


In article <34FDD236.1D0CBFDE at yoga.com>, Gil Kedem <gil at yoga.com> wrote:
> I'm searching for a parser (commercial or freeware) for the Blast
> program.
> Are there any out there?

A couple of Blast parsers written in Perl are linked to at
http://www.techfak.uni-bielefeld.de/bcd/Perl/Bio/welcome.html#blast
under the header ``Blast Modules''.

Let me take this chance to advertise the Bioperl project a little bit--
as a volunteer project we'd like to attract some new people who are
interested to help in designing and coding new modules as well as using,
and critiquing existing modules, e.g. we're looking for someone who is
interested in merging the best of the current Blast modules into one
``unified'' object. Our Homepage:
http://www.techfak.uni-bielefeld.de/bcd/Perl/Bio/
(Most likely, at least some of the core team will be at ISMB'98 and
will present a poster.)

In reference to the recent discussion in bionet.software, in our opinion
Perl has an excellent future as a fast prototyping and scripting
tool for bioinformatics which can interact w/ Java/Corba [1][2].
I tend to agree with the article that ``predicts that the languages of
choice will be Perl and Java with C used where speed is critical, and C++
essentially only existing as legacy code'' (quoted in [3]), although I'm
sure C will continue to be a "language of choice" for many tasks as well,
and I'd be happy if ``cleaner'' languages like python, sather, and even
functional languages would be more popular, but that's not the way it is
right now.

I'm currently writing a major application [4] in 100% Perl (so to say ;-),
with more than 10.000 lines of code spread over 4 modules (objects),
and my experience is positive: maintenance (which I was worried about)
turns out to be easy, and the recent migration from Perl 5.002 to 5.004
resulted only in a couple of new warnings. I think it helps spending
a great deal of time on object design, and on abstraction, trying not to
write any duplicate code (even using closures). I'm also trying to seperate
interface from implementation as much as possible, so that I'm hopeful
that it won't take days to use the PDL package [5] for the matrix-crunching
part of the code at some point. Since Perl gives you the freedom to write
bad code, even small scripts benefit from insisting e.g. on variable
declaration (by adding "use strict;" to the script), on warnings (using
the `-w' switch), etc.

[1] http://www.perl.com/971127.html, http://perl.oreilly.com/prk_index.html
[2] http://www1.lunatech.com/cope/
[3] http://www.bio.net/hypermail/BIO-SOFTWARE/9802/0202.html
[4] http://www.techfak.uni-bielefeld.de/bcd/Perl/Bio/Docs/phylosnapshot.html
[5] http://www.aao.gov.au/local/www/kgb/perldl/

best wishes,
Georg Fuellen,
Univ. Bielefeld, Research Group in Practical Comp. Science
http://www.techfak.uni-bielefeld.de/bcd/welcome.html






More information about the Bio-soft mailing list