[Bioc-devel] A geneSet data class for facilitating GSEA

Fri Mar 16 14:18:43 CET 2007

Hi,

Vincent Carey 525-2265 wrote:
>> Dear bioc-developers,
>>
>> would it be useful to introduce an additional slot for the direction and/or
>> magnitude of expression change of each gene in the gene set?
> 
> My understanding is that we are currently trying to get a
> structure that identifies a group of genes in a coherent way.
> Connecting a group of genes to a specific experimental result is outside
> the scope of this task.

   That is a good question, but I would like to point out that it is 
almost surely the case that notions of direction of change, and 
magnitude are with respect to a comparison of phenotype (eg disease to 
healthy, or stage I vs stage IV) and hence are not properties of the 
gene set.

   While that information is important and useful in a particular 
analysis, it should not be stored with the gene set, in my opinion. We 
will need some easy way for users to specify it and use it in practice, 
but as Vince has said, it is probably not what we want here.

> 
> Designing an extension of the group class that incorporates
> qualitative or quantitative information on gene behaviors under
> certain conditions seems worthwhile but should be kept separate
> from the original design problem -- I think.
> 
>> It seems that GSEA and GSEA-like methods use sets of genes that are
>> homogeneously down- or upregulated (correct me if I am wrong, I am far from
>> being up to date on GSEA methods).
>>
>> This seems to be reflected in the example presented in the PGSEA vignette
>> where target genes of Ras and Myc are separated into 'UP' and 'DN' regulated
>> genes.

    Hopefully we will use UP and DOWN, the savings by using 
abbreviations are almost never worth it, especially when for many users 
English is not their first language.

   best wishes
     Robert

>>
>> However, (alternative?) methods could actually use the quantitative
>> information about expression changes to score each gene set. Adding a
>> corresponding slot in the geneSet class would allow to accommodate such
>> methods.
>>
>> Best,
>> Alexandre
>>
>>
>> -----Original Message-----
>> From: bioc-devel-bounces at stat.math.ethz.ch
>> [mailto:bioc-devel-bounces at stat.math.ethz.ch] On Behalf Of Dykema, Karl
>> Sent: mercredi, 14. mars 2007 16:15
>> To: bioc-devel at stat.math.ethz.ch
>> Subject: Re: [Bioc-devel] A geneSet data class for facilitating GSEA
>>
>> Sorry I forgot to attach the str()
>>
>> $ 15-delta prostaglandin J2 10 uM  DOWN : list()
>>   ..- attr(*, "reference")= chr "15-delta prostaglandin J2 10 uM  DOWN "
>>   ..- attr(*, "desc")= chr "DOWN "
>>   ..- attr(*, "source")= chr "PubMed"
>>   ..- attr(*, "design")= chr "????"
>>   ..- attr(*, "identifier")= chr "17008526"
>>   ..- attr(*, "species")= chr "human"
>>   ..- attr(*, "data")= chr "raw"
>>   ..- attr(*, "private")= chr "no"
>>   ..- attr(*, "creator")= chr "Karl Dykema <karl.dykema at vai.org>"
>>   ..- attr(*, "ids")= chr [1:75] "171392" "5680" "2149" "54557" ...
>>   ..- attr(*, "class")= atomic [1:1] smc
>>   .. ..- attr(*, "package")= chr "PGSEA"
>>
>>
>> This closely mirrors the geneSet proposed and we will be happy to adopt
>> a consensus structure.
>>
>> The only significant difference is a "creator" to let folk know who
>> curated the gene list... This may help if groups are collaborating to
>> the collect gene sets.
>>
>>
>> -------------------------------
>> Karl Dykema
>> Bioinformatics Programmer/Analyst
>> Laboratory of Computational Biology
>> Van Andel Research Institute
>> 333 Bostwick Ave. NE
>> Grand Rapids, MI 49503
>> (616) 234-5554
>>
>>
>>
>> -----Original Message-----
>> From: Vincent Carey 525-2265 <stvjc at channing.harvard.edu>
>> Date: Wed, 14 Mar 2007 10:19:36 -0400 (EDT)
>> To: Sean Davis <sdavis2 at mail.nih.gov>
>> Cc: <bioc-devel at stat.math.ethz.ch>, Ross Lazarus
>> <rerla at channing.harvard.edu>
>> Subject: Re: [Bioc-devel] A geneSet data class for facilitating GSEA
>>
>> i like this idea in principle.  the RGenetics folks may have done
>> something in this direction.
>>
>> you might want to have geneList as an abstract class, and then extend to
>> EntrezGeneList, RefseqGeneList and so forth so that dispatch could work
>> without looking into the idType ...
>>
>> a version or date field might also be important
>>
>> ---
>> Vince Carey, PhD
>> Assoc. Prof Med (Biostatistics)
>> Harvard Medical School
>> Channing Laboratory - ph 6175252265 fa 6177311541
>> 181 Longwood Ave Boston MA 02115 USA
>> stvjc at channing.harvard.edu
>>
>> On Wed, 14 Mar 2007, Sean Davis wrote:
>>
>>> GSEA, both the specific method and the general concept, is becoming
>>> more prevalent and important in data analysis.  There have been
>>> several mentions of including various "gene lists" for use with
>>> Category or other methods.  Is there interest in making a generic
>>> geneSet class for storing such information?  (Or does it already exist
>>> and I just haven't seen it?)  I bring this up because I think it could
>>> be quite useful to have a general solution for the community (like the
>>> eSet class has become).  A class could be as simple as a vector of
>>> Entrez Gene IDs to something more complicated (but perhaps a bit more
>> useful for general consumption) like:
>>> identifier: an identifier for the set (perhaps from a public database
>>> like
>>> MSigDB)
>>> title:  One line title
>>> description: free text description
>>> species: The species to which the dataset applies
>>> URL: from where the data were derived
>>> MIAME: class "MIAME" object
>>> protocol: (could be in MIAME, also) description of methods to produce
>>> genelist from raw data source
>>> idType:  What type of ID is stored (Entrez, Refseq, Ensembl, etc)?
>>> geneList: vector of IDs
>>>
>>> A simple wrapper data structure (even just a list) could then be used
>>> to distribute the geneSets.  Some methods could then be defined for
>>> converting to an incidence matrix for use by Category, etc.  But I
>>> think the most important points from above are 1) maintaining some
>>> metadata about the genelists and 2) standardization to reduce
>>> duplicated work.  Individual groups would then instantiate the
>>> geneSets using whatever means they see fit (parsing MSigDB, IPI files,
>> etc.).
>>> Any thoughts?
>>>
>>> Sean
>>>
>>> _______________________________________________
>>> Bioc-devel at stat.math.ethz.ch mailing list
>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>
>> _______________________________________________
>> Bioc-devel at stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>
>> This email message, including any attachments, is for the so...{{dropped}}
>>
>> _______________________________________________
>> Bioc-devel at stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>
>> _______________________________________________
>> Bioc-devel at stat.math.ethz.ch mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>
> 
> _______________________________________________
> Bioc-devel at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/bioc-devel
> 

-- 
Robert Gentleman, PhD
Program in Computational Biology
Division of Public Health Sciences
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N, M2-B876
PO Box 19024
Seattle, Washington 98109-1024
206-667-7700
rgentlem at fhcrc.org