[Biopython-dev] "Features" of Bio.Clustalw

Brad Chapman chapmanb at arches.uga.edu
Sat Jul 28 15:03:10 EDT 2001

Hi Davide;
> I wanted to point out some (minor) problem I had using Bio.Clustalw.

Thanks for sending the comments!

> Thanks for the comments.  We're all getting back from ISMB and 
> catching up on email, so the proper developer will comment soon.

I think I'm the "proper developer" since I wrote the Clustalw stuff, 
although this is the first time I've had someone refer to me like that
> In an alignment run clustalw generates two files:
> - the output file (with default extension aln), and
> - the tree file (with default extension dnd)
> the MultipleAlignCL permits to change the output file with method
> set_output but there is no way to change the tree file.

Yup, this is bad. Thanks for catching it!
> Is it possible to add a method set_tree that add to the command line the
> -newtree option and the -align option?

Definately -- this sounds exactly right.
> The do_alignment function executes
>   print "executing %s..." % command_line
> and I think not everybody want it printed (I do not). Yes, one could
> temporary redirect sys.stdout, but why?

Ooop, another good catch. It is definately bad for libraries to print
> Even if one sets the alignment type to PROTEIN (via set_type method of
> MultipleAlignCL) the resulting alignment has DNA alphabet.

Bleah. You caught all of my bugs :-). Bad me!

> PS: Is it possible to submit patches to biopython? How can I contribute?

Definately. We loooooove contributions. If you have fixes for the
above problems, the best way to submit the patch is to send a context
or unified diff ('diff -c' or 'diff -u') to this list, and I'll be
happy to apply it and make sure things are worked out. You can check
out the CVS tree anonymously (instructions for this are at
http://cvs.biopython.org/ ), and submitting patches against this would
be extra-great.

Thanks for your feedback on this. If you can send patches for whatever
you can fix, I'll take care of the rest of the bugs, and beef up the
test suite to catch them. 

Thanks again!

More information about the Biopython-dev mailing list