[BioC] how to build a GOstats-compatible annotation package for plasmodium falciparum?

[Marc Carlson]

Paul Shannon wrote:
> Hi Marc,
>
> It sounds like you have your hands full!
>
> Would it be crazy if we were to undertake this ourselves, at the SBRI?
> I am thinking of an organism-based package (like YEAST) rather than
> a chip-specific package.
>
> Does the new pipeline create annotation packages which work with
> GOstats and GSEA?
>
>  - Paul
>
>
> On Nov 6, 2007, at 4:41 PM, Marc Carlson wrote:
>
>> Well I plan to do this.  But it's just not going to happen overnight.  I
>> am booked right now with a homology implementation and I also have to
>> finish getting part of the promised pipeline in place for our other
>> annotation package collaborators so that they can update their packages
>> to the newer format for the upcoming release.  This means that I may not
>> get to start this for a couple of months, but it has been added to my
>> todo list.
>>
>> For now, I recommend that you try to use the AnnBuilder package.  We are
>> planning to retire this package along with the style of annotation
>> package that it spawns so this is definitely NOT a good long term
>> solution and is to be used for the short term ONLY.  But I think that it
>> will probably get you the quick fix that you need.
>>
>> http://bioconductor.org/packages/2.1/bioc/html/AnnBuilder.html
>>
>>
>>     Marc

Yes I am a very busy guy.  I would love to collaborate with you on
this.  But I don't think that making a new package from scratch would be
a very efficient use of your time.  That is, there are good reasons why
its going to take me a little while to get it to you.  There are a lot
of things to do.  I agree that an organism based package is what is
called for here and that is what I was planning to work on. 

As for your immediate needs, all you should need for GO stats or GSEA is
an environment which you could make for yourself from the appropriate
information.  I can give you those parts in an unformated form if you
want them.   To format them into a proper environment you should only
need to wrap them up in one.

#Lets suppose that we rip off some of the info from the YEAST package to
see how this would work:
library(YEAST)
res=mget(ls(YEASTGO), YEASTGO)

#Then we could quickly make a couple quick fakey environments:
MYGO=new.env(parent=emptyenv())
for (nm in names(res)) MYGO[[nm]] <- res[[nm]]

MYENTREZID <- new.env()
for (nm in ls(MYGO)) MYENTREZID[[nm]] <- paste("fauxId", nm)

#Then we could package them up into a local environments:
MYpkg <- new.env(parent=emptyenv())
MYpkg[["MYGO"]] <- MYGO
MYpkg[["MYENTREZID"]] <- MYENTREZID

#And attach them
attach(MYpkg, 2, "package:MY")


#At this point we should be able to do with these environments whatever
we need to.

I have the relevant information here from NCBI for falciparum to make
both of these environments (for real).  If you send me a personal email,
I can arrange to get it to you...

    Marc