On Mar 27, 8:45 pm, "Natacha" <natab... from gmail.com> wrote:
>> Thanks for your answer, I have already try to use geneWise. But I have
> a problem with him.
> I would find the localisation of intron only with similarity, without
> any statistic or model. And geneWise use a lot of model. I have try to
> remove a lot of information contained into the human.gf and
> human.stat file, this improve my result. But it's not perfect and I
> think that there are others file or model, but I don't find them.
> Have you an idea ?
You may skip the frequency files, as described here:
"Use -splice flat -intron tied which switches the splice model to
``start at GT, finish at AG'' with no other information"
Haven't done it myself (I work with vertebrate DNA mostly), so no idea
if it gives any sensible results.
BTW, are you trying to annotate the whole genome or trying to get a
small set (say 1-10) of putative genes in a region?