tfscan blues

[Charles]

> > I have a related question : do you think that it would be possible to build
> > fake transfac databases from a simple file?
> > Currently I'm adding into an array (in the GCG findpattern format) any
> > binding site of my interest that i find in the litterature. (with a name and
> > a reference)
> > The goal would be to use existing programs to do searches within a set of
> > home-selected transcription factors.
>
> It woul dbe nice to have a public front end for such a database so that
> submissions could be sent to a curator. Then we can return the information
> to the public domain (all literature referenced of course so we cannot be
> accused of stealing TRANSFAC).

Well, i bet that tools allowing public contribution could be a CGI form,
or a CVS archive, but setting up those interfaces is far beyond my
capacities.

Currently i have betveen 40 and 50 entries, very focused on molecular
biology of development in early vertebrate embryos. Some are complex and
other degenerate. I could not get something interesting of it for
the moment, using GCG findpatterns : i get either close to no sites or
plenty if i allow mismatches.

I can easily imagine a way to store more information in a separate array,
and then build a findpattern data file using a perl script. But i'd like
to try some programs that give a score to the matches, in order to search
for high complexity binding sites more efficiently.

Charles