[BioC] Gostats and custom list

Marc Carlson mcarlson at fhcrc.org
Thu Feb 26 19:52:19 CET 2009


Hi Iain,

I guess that it depends on what your categories actually are then.  If
they are something like GO, or KEGG, then I don't think you will need to
write any code.  It should be sufficient in that case to just make the
appropriate object type and run hyperGTest() after carefully defining
the genes that meet a certain criteria (presumably the ones in one of
you particular categories).  If however you want to run hypergeometric
tests on an entirely new and still unsupported 'kind' of annotation,
then what the man page says might apply to you and you might want to
start looking at the source code for the Category package to see what is
involved in defining new variants of the functions mentioned.  If you
are thinking of taking the latter approach, then we should probably talk
about what kind of annotation package this is and whether or not we
should consider supporting it here as well?  Also, if it is the latter
case, then it might also be a good idea to migrate this discussion to
the bioc-devel list.

https://stat.ethz.ch/mailman/listinfo/bioc-devel


  Marc




Iain Wallace wrote:
> Hi Marc,
>
> Thanks for getting back to me. I did read the GOstats vignette, but I
> wasn't clear how to define a new set of categories.
> I have no idea how to define a new subclass, as suggested by
> HyperGtest man page which says "To add support for a new FOO category
> type, a developer would need to create a 'FooHyperGParams' subclass
> and then define two methods specialized to the new subclass that get
> called from inside 'hyperGTest': 'universeBuilder' and
> 'categoryToEntrezBuilder'.
>
> Nothing is actually going wrong, hence the lack of sample code,I am
> more just very confused about where to look. I have found some really
> useful examples of how to use the hyperGTest function with Kegg
> pathways, but unfortunately none showing me how define my own very
> limited set of categories.
> You are exactly right, in that I want to see if any of the categories
> are over represented in my set of interesting genes compared to all
> the other genes and categories.
>
> Thanks again for any pointers/tips,
>
> Iain
>
> On Wed, Feb 25, 2009 at 7:49 PM, Marc Carlson <mcarlson at fhcrc.org
> <mailto:mcarlson at fhcrc.org>> wrote:
>
>     Hi Iain,
>
>     Have you read the GOstats vignette?  It is a great place to start
>     if you
>     are doing this.
>
>     It would also be great if you could have a peek at our posting
>     guide and
>     then give us more details about what you are doing and what seems to
>     actually be going wrong???
>
>     http://www.bioconductor.org/docs/postingGuide.html
>
>     Speaking very generally, to use GOstats you need to have a list of
>     genes
>     that you think are interesting.  So if you have a set of categories,
>     perhaps you were hoping to look at just one of the categories as
>     compared to all the other genes?
>
>
>      Marc
>
>
>
>
>     Iain Wallace wrote:
>     > Hi all,
>     >
>     > I was wondering if anybody could give me pointers on how to use
>     GOstats with
>     > a custom list of gene annotations/categories
>     > I have my list of yeast genes associated with one or more of 25
>     different
>     > categories
>     >
>     > My file looks something like this:
>     >
>     > Gene_A Category_1
>     > Gene_A Category_2
>     > Gene_B Category_1
>     > Gene_B Category_2
>     > Gene_C Category_2
>     > Gene_C Category_3
>     >
>     > I am not sure how to convert this into an annotation that
>     GOstats likes.
>     > I tried to figure out how to do it using the AnnotationDbi
>     library, but I
>     > wasn't able to so any pointers would be greatly appreciated.
>     >
>     > My current GOstats command is
>     >
>     params=new("GOHyperGParams",geneIds=sel,universeGeneIds=uni,annotation="org.Sc.sgd.db",ontology=class,pvalueCutoff=as.numeric(pvalue),conditional=FALSE,testDirection="over")
>     >     over=hyperGTest(params)
>     >
>     > Thanks,
>     >
>     > Iain
>     >
>     >       [[alternative HTML version deleted]]
>     >
>     > _______________________________________________
>     > Bioconductor mailing list
>     > Bioconductor at stat.math.ethz.ch
>     <mailto:Bioconductor at stat.math.ethz.ch>
>     > https://stat.ethz.ch/mailman/listinfo/bioconductor
>     > Search the archives:
>     http://news.gmane.org/gmane.science.biology.informatics.conductor
>     >
>     >
>
>



More information about the Bioconductor mailing list