[Bioperl-l] Getting IC & Consensus with Bio::Matrix::PSM::SiteMatrix

[Stefan Kirov]

Hi Edwars,

Edward Wijaya wrote:

> Dear Stefan and James,
>
> Thanks so much for answering
>
> On Mon, 14 Mar 2005 21:15:43 +0800, Stefan Kirov <skirov at utk.edu> wrote:
>
>> The rules for too low are:
>
> [snip]
> Got it Stef.
>
>> I can fix this, but do you really want to get in your consensus a  
>> position with proba less than 0.5?
>
>
> Yes, Don't you think by default it should be that way?
> Besides it'll be nice to have an *option* of how are we going to get 
> the  Consensus.
>
I'll commit the code tomorrow, just update your bioperl-live

>> We can easily add IC calculating method if you really need it.
>
> Yes, we definitely need that. I think naturally we would need the  
> computation,
> same as e-value.

You mean you want SiteMatrix to compute the e-value? Hmm... we are 
getting out a bit out of scope here. Essentially PSM modules were 
supposed only to provide data structure and parsers. And if IC is 
generally straightforward and contained in the PFM/PSM, this is not the 
case with e-val or p-val. Therefore I am reluctant to put it in PSM 
collection.

>
> Actually I have the subroutine to compute the IC given the frequency  
> matrices.
> You can use it to incorporate it to the module if you want, although 
> it  isn't a great piece of work.
> I just thought it may save you time.
>
Sure, that would be great. Just send it and I will optimize it if I can 
and put it in. But maybe it should go to Bio::Tools... Any thoughts from 
anyone else?

>> Please let me know if you have further questions.
>
> I'll save them till next time Stef ;-)
>
>
Stefan