[GMOD-devel] RE: [Open-bio-l] RE: seqfeature

[Matthew Pocock]

Chris Mungall wrote:

> 
> I'm absolutely in favour of frame columns in relations that are specific
> to alignments; in my original suggestions (3) and (4) I listed "e-value"
> and "score" but this was implicitly meant to be a list of whatever columns
> are most commonly used (in gadfly i have explicit columns for e-val,
> score, % identity, qframe, sframe)

For once, I am all for frame going into an API. To my mind, frame is a 
property of the mapping between a portion of a sequence and an 
alignment/homology. When you talk about exon frame, it is actualy a 
number that tells you something about the relationship between the 
exon's portion of a transcript and the alignment of the codons in that 
transcript to a translation. I hope this viewpoint is not too 
ideosynchratic. I guess exon frames are sort of like maximum-likelihood 
limits on aligning that exon with some frame within some unknown set of 
transcripts to some unknown set of translations.

-M