[Biopython-dev] [Biopython - Bug #3234] (New) Bio.HMM Viterbi algorithm: initial state probabilities are wrong

[Peter Cock]

On Mon, May 23, 2011 at 4:30 PM,  <redmine at redmine.open-bio.org> wrote:
>
> Issue #3234 has been reported by Walter Gillett.
>
> ----------------------------------------
> Bug #3234: Bio.HMM Viterbi algorithm: initial state probabilities are wrong
> https://redmine.open-bio.org/issues/3234
>
> Author: Walter Gillett
> Status: New
> Priority: Normal
> Assignee: Walter Gillett
> Category:
> Target version:
> URL:
>
>
> Spun off from #2947, see that bug for discussion. Initial state probabilities
> should be set explicitly, rather than using the probability of transitioning
> from a state back to itself, which is incorrect.

Would anyone more familiar with HMMs that I am like to volunteer to review
Walter's changes? e.g. Philip (CC'd)?

Walter's sent a pull request via github:
https://github.com/biopython/biopython/pull/6

This consists of two commits, the first an unrelated minor change to the
ignore file for people using the NetBeans IDE:
https://github.com/wgillett/biopython/commit/50659de2f0cfa3f0bc913b4ea88d6a001b543d98

Secondly his fix for Bug 3234,
https://github.com/wgillett/biopython/commit/a60ac226ceed21fd856ff1ec1dbea2782e2172ae

Thanks,

Peter