[BioC] Problem with invalid GO term in HyperGResult object

[Seth Falcon]

Hi Jenny,

Jenny Drnevich <drnevich at uiuc.edu> writes:
> I posted this message last week, but didn't get any responses. I've
> been trying everything I can think of to remove this one problematic
> GO term from my HyperGResult object so I can get the other 45
> overrepresented GO terms, but I'm not having any luck. I would
> appreciate any suggestions!!!

Sorry, things have been a bit busy around here :-)

> At 12:22 PM 9/5/2007, Jenny Drnevich wrote:
>>Hi everyone,
>>
>>I'm doing hypergeometric testing of GO terms using GOstats on some
>>data from ath1121501 Affy arrays. I had some trouble figuring out how
>>to work with these arrays at first because the annotation package
>>doesn't include EntrezIDs, but after searching the archives I found
>>all you had to do was this to get it to work:
>>
>>ath1121501ENTREZID <- ath1121501ACCNUM
>>
>>Anyway, everything was going fine, expect in one my HyperGResult
>>objects there is a significant GO term that throws up an error:
>>
>>  >SVC.up.BP <- new("GOHyperGParams", geneIds =
>>entrezUniverse.sortVctrl[EID.sortVctrl==1],
>>                  universeGeneIds = entrezUniverse.sortVctrl,
>>annotation = "ath1121501",
>>                  ontology = "BP", pvalueCutoff = hgCutoff,
>> conditional = FALSE,
>>                  testDirection = "over")
>>
>>  >SVC.up.BP.results <- hyperGTest(SVC.up.BP)
>>
>>  >summary(SVC.up.BP.results)
>>Error in .checkKeys(value, Lkeys(x), x at ifnotfound) :
>>    invalid key "GO:0002226"

First, can you try updating your packages.  We've very recently pushed
out new candidates for annotation as well as some bug fixes for
AnnotationDbi.

If the problem persists, can you send me (off-list) an rda file
containing your SVC.up.BP object?

+ seth

-- 
Seth Falcon | Computational Biology | Fred Hutchinson Cancer Research Center
BioC: http://bioconductor.org/
Blog: http://userprimary.net/user/