[Biojava-l] Translation tables and Met

Emig, Robin Robin.Emig@maxygen.com
Wed, 29 Aug 2001 09:56:33 -0700


looks good, with one for stop as well
-Robin

-----Original Message-----
From: Matthew Pocock [mailto:mrp@sanger.ac.uk]
Sent: Wednesday, August 29, 2001 9:41 AM
To: Emig, Robin
Cc: 'Keith James'; BioJava List
Subject: Re: [Biojava-l] Translation tables and Met


Emig, Robin wrote:

> Sounds good, but I do like the idea of including start/stop codons in the
> class.
> Robin

By all means. How about a:

   // returns a Symbol with a getMatches() that will match
   // exclusively to start codons
   Symbol getStartCodons();

Does that look OK?

Matthew

> 
> -----Original Message-----
> From: Matthew Pocock [mailto:mrp@sanger.ac.uk]
> Sent: Wednesday, August 29, 2001 3:09 AM
> To: Emig, Robin
> Cc: 'Keith James'; BioJava List
> Subject: Re: [Biojava-l] Translation tables and Met
> 
> 
> Emig, Robin wrote:
> 
> 
>> 4) What about the idea of just creating a new object called
> 
> CodonBiasTable,
> 
>> which has all of these ideas added, plus the frequency weights? It could
> 
> be
> 
>> based off of an internal translation table, and distribution.
>> -Robin
> 
> 
> You will get the most mileage out of the CodonBiasTable if it is 
> something as simple as:
> 
> public interface CodonBiasTable {
>    // translates codons to protein
>    TranslationTable getTranslationTable();
> 
>    // codon bias table
>    //
>    // 2nd order distribution of p(aa | codon)
>    Distribution<codonXprotein> getCodonBias();
> 
>    // joint distribition
>    //
>    // 2nd order distribution of p(aa && codon)
>    Distribution<codonXprotein> getJointProbability();
> }
> 
> This way, you can build a pair HMM using the joint probability, or can 
> probablisticaly generate CDS given that you already know the protein 
> sequence using the codonBias table. I guess the two probabilities are 
> related. I will pop in a couple of utility classes into 
> org.biojava.bio.dist
> 
> Does this sound sensible?
> 
> Matthew