Bioperl: RE: Bioperl-guts: codon bias modules - what should be in them?

Gatherer, D. (Derek)
Thu, 23 Mar 2000 08:44:30 +0100

-----Original Message-----
From: John Peden []
Sent: 22 March 2000 18:05
To: Gatherer, D. (Derek); bioperl guts;
Subject: RE: Bioperl-guts: codon bias modules - what should be in them?

>I would be against including Codon Adaptation Index which must be one of
>most misused codon indices around, the index is good but people try to
>it incorrectly to the wrong datasets (and they still manage to get it
>published in Science).


>I would avoid any index that implies prior knowledge of optimal codon
>as these are species specific, and we only known this choice for very few
>species and it including them would seem to imply that someone would have
>maintain the module with lists of optimal codons.

Absolutely, I don't want to have to do that!!

>I consider Nc as superseding Scaled Chi square and derivatives, its just as
>or more accurate (Comeron and Aguade 1998 J Mol Evol 47:268-274) and is
>length dependent, so I would just include Nc.

I must read this article.  I always thought that SCS was length independent,
since the number of codons goes on the bottom line, but maybe Comeron and
Aguade can convince me otherwise.

>The P2 statistic is quite old and I don't see it as having a general
>application to codon bias.


>I am unfamiliar with "Shannon triplet informational entropy and redundancy"
>as applied to codon bias I see it more as a genome survey or gene discovery
>tool than codon bias method.

It's just that highly biased genes exhibit low entropy.

>So I would suggest a single module with light methods.

>these indices are genetic code dependent so this needs to
>be coded for as well.

aha!  Do we have a module that already handles this??

=========== Bioperl Project Mailing List Message Footer =======
Project URL:
For info about how to (un)subscribe, where messages are archived, etc: