[Bioperl-l] conservative amino acid change?

Aaron J Mackey Aaron J. Mackey" <amackey@virginia.edu
Mon, 6 Aug 2001 10:00:11 -0400 (EDT)


You might also want to consider a special cloning method like "scale" to
return a new copy of the matrix with a different scaling (than the
default).  You might also consider a scale accessor method that reported
the current scale.  It would also be nice to have a method to extract the
information content (bits/residue) of the matrix.  Of course you'd need to
have some letter frequencies lying around to do that - There are a handful
of "standard" amino acid frequencies, it'd be great to get them into the
repository for easy access.

While not applicable to the BLOSUM series, many other matrices have an
evolutionary rate matrix as their foundation (namely, PAM, JTT, WAG, etc).
In this case you need only store one set of numbers, and dynamically
generate the desired matrix at the desired scaling.  I have some code
lying around to do this bit of black magic, if there's interest I'll
send it along.

In summary, try to think of matrices as more living and breathing entities
rather than just an organized bucket of numbers.  I'm sure this will mean
that you'll want to build a Matrix::Factory and all that other stuff that
makes using BioPerl difficult for us simpletons ;) ;)

-Aaron

On Mon, 6 Aug 2001, Heikki Lehvaslaiho wrote:

> OK. I'll write the modules:
>
> Bio::Matrix::SimilarityMatrixI
> 	id()
> 	name()
> 	moltype()
> 	pam()
> 	max()
> 	min()
> 	def()
> 	deletion()
> 	similarity_score()
>
> Bio::Matrix::SimilarityMatrix
> Bio::Matrix::SimilarityMatrixIO
> Bio::Matrix::SimilarityMatrixIO::full
> Bio::Matrix::SimilarityMatrixIO::half
>
> or something like that...
>
>
> We'll need a data directory to place the default matrixes in. What
> would be the best (platform independent) way of refering to files
> within bioperl?
>
> 	-Heikki
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l@bioperl.org
> http://bioperl.org/mailman/listinfo/bioperl-l
>

-- 
 o ~   ~   ~   ~   ~   ~  o
/ Aaron J Mackey           \
\  Dr. Pearson Laboratory  /
 \ University of Virginia  \
 /  (434) 924-2821          \
 \  amackey@virginia.edu    /
  o ~   ~   ~   ~   ~   ~  o