[Biojava-l] Alignment with GAPs

[mark.schreiber at novartis.com]

Hi -

I think the DP class should accept a GappedSequence. To get the result you 
want you will probably need to have at least one match state that can emit 
gaps. I'm curious to know why you would want to do that kind of alignment 
though?

- Mark





"Dickson S. Guedes" <guedes at unisul.br>
Sent by: biojava-l-bounces at portal.open-bio.org
01/27/2006 04:49 AM

 
        To:     Matthew Pocock <matthew.pocock at ncl.ac.uk>, Biojava-l at biojava.org
        cc:     (bcc: Mark Schreiber/GP/Novartis)
        Subject:        Re: [Biojava-l] Alignment with GAPs


Thanks Matthew,

To produce a pair-wise alignment from two un-aligned sequences and 
introduce gaps I have used the sample at "BioJava In Anger" and it runs 
successfully.

Now I need an alignment of two sequences where one of them already have 
gaps before I align.

I mean that the DP class don´t accept GappedSequence, it´s right?

[]s
Guedes

Matthew Pocock escreveu:
> If I understand you correctly, one of the two sequences you are aligning 

> contains a gap before you align them? Or do you want to produce a 
pair-wise 
> alignment from two un-aligned sequences and introduce gaps?
> 
> If it is the former, you want a state that emits a gap in one sequence 
and a 
> symbol in the other, and also advances {1,1}. I think that is easy 
enough to 
> set up, but can't remember the exact code. If the worst comes to the 
worst, 
> you can construct the distribution over {gap,Protein} using the classes 
> in .dist and then set up a SimpleState, providing the advance and 
alphabet in 
> the constructor.
> 
> Matthew
-- 
--
:: Dickson S. Guedes (guedes at unisul dot br)
::
:: UNISUL - Universidade do Sul de Santa Catarina
:: ATI - Assessoria de Tecnologia da Informação
:: (0xx48) 621-3200 - http://www.unisul.br
--
"Há 10 tipos de pessoas no mundo: as que entendem
  binário, e as que não entendem"
_______________________________________________
Biojava-l mailing list  -  Biojava-l at biojava.org
http://biojava.org/mailman/listinfo/biojava-l