[Bioperl-l] Re: 0.9.1 release schedule

[Boris Lenhard]

>
> Can people report in on what code they have in progress - I'll outline
> some of the major components that would be good to get into a dev release
> for extensive testing.  Some of these may need to wait until 0.9.2 which
> should go out at the end of October if we have enough new code.

<snip>

> * TBFS (transcription factor detection and analysis) ( Boris Lenhard)
>

Hello everybody,

here is a quick update on  the status of TFBS modules:

First, I apologize for the delay in documenting the modules, I was just too busy using them :)

Yesterday I made public the 0.1 release of the package, which is available at
http://forkhead.cgb.ki.se/TFBS/ . All public method now have basic POD documentation. There are also
new modules for the Gibbs pattern generator and retrieval of matrices from the TRANSFAC database via
the web. The package now also contains a reasonable suite of regression test scripts.

The major obstacles for seamlessly incorporating  TFBS into bioperl are as follows:

* it still requires PDL for matrix manipulation. I know that it is no sin to depend on external
modules freely available from CPAN, but PDL is huge and apparently not easy to build on some
platforms. It would be possible to replace it with some simpler matrix manipulation package or code
everything manually, but it requires the time I will just not have in the next several weeks.

* it contains a compiled perlxs extension (and a very ugly one)  for scanning of DNA sequences by a
position weight matrix. This should be moved to bioperl-ext or replaced by a perl-only one (which is
slower by a factor of 20, which becomes uncomfortably slow for longer sequences)

* It still does not have proper interface definition  (*I.pm) modules, which has to do with the way
the classes evolved (not by software engineering textbooks, for sure), but that is fairly easy to fix.

* On non-UNIX platforms it is likely to fail miserably due to a few UNIX-specific remnants in the code
(that should also be easy to fix).


If anyone of the bioperl core team can have a look at the modules and give me some directions, I would
be willing to work on making TFBS meet the requirements for inclusion.


Best regards,

Boris


#####################################
 Boris Lenhard, Ph.D.
 Center for Genomics and Bioinformatics
 Karolinska Institutet
 Berzelius väg 37
 171 77 Stockholm, SWEDEN
 Phone: +46 (0)8 728 6142
 FAX: +46 (0)8 33 79 83
 E-mail: Boris.Lenhard@cgb.ki.se
#####################################