Seth Falcon wrote: > Hi Jim, > > Any insight on this one? > > "Na, Ren" <na at="" uthscsa.edu=""> writes: > >>Hi, I tried to find the genes in some of significant GO categories, >>and find out Counts of EntrezIDs returned from summary() are not >>same as the numbers of entrezIDs extracted from probeSetSummary() >>for some significant GO categories. My question is whether they >>should be same. Here is the code I used, They should be the same - you have uncovered a bug. This function works perfectly well if conditional = FALSE, but if you condition then the simple way we are mapping from Entrez Gene IDs ==> Probeset IDs no longer works, because we are ignoring the fact that the number of Entrez Gene IDs under consideration at any level of the GO DAG is conditional upon any significant child terms. Right now we are in the final stages of releasing BioC 2.0, so the GOstats package will almost certainly be released with this bug intact. However, I will be working to correct the problem and will push any fixes to the release repository. I will send a note to this list when the fix is in. The version will also increment, so a simple library(Biobase) update.packages(repos = biocReposList()) should get things updated when that happens. Best, Jim -- James W. MacDonald, M.S. Biostatistician Affymetrix and cDNA Microarray Core University of Michigan Cancer Center 1500 E. Medical Center Drive 7410 CCGC Ann Arbor MI 48109 734-647-5623 ********************************************************** Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.

I have just pushed a fix for this to the release repository, so it looks like GOstats will ship with a fixed version of probeSetSummary(). Note that this function still doesn't work exactly as I would like - the vector of 'selected' identifiers only counts one probeset per Entrez Gene ID because of the necessity to subset to unique Entrez Gene IDs before doing the hyperGTest(). The devel version now has an additional argument 'sigProbesets' that can be used to overcome this limitation, but we are well past the point that changes to the API can be introduced to release packages so this will have to wait for release 2.1 (and those who like to live on the bleeding edge with devel packages). Best, Jim James W. MacDonald wrote: > Seth Falcon wrote: > >>Hi Jim, >> >>Any insight on this one? >> >>"Na, Ren" <na at="" uthscsa.edu=""> writes: >> >> >>>Hi, I tried to find the genes in some of significant GO categories, >>>and find out Counts of EntrezIDs returned from summary() are not >>>same as the numbers of entrezIDs extracted from probeSetSummary() >>>for some significant GO categories. My question is whether they >>>should be same. Here is the code I used, > > > They should be the same - you have uncovered a bug. > > This function works perfectly well if conditional = FALSE, but if you > condition then the simple way we are mapping from Entrez Gene IDs ==> > Probeset IDs no longer works, because we are ignoring the fact that the > number of Entrez Gene IDs under consideration at any level of the GO DAG > is conditional upon any significant child terms. > > Right now we are in the final stages of releasing BioC 2.0, so the > GOstats package will almost certainly be released with this bug intact. > However, I will be working to correct the problem and will push any > fixes to the release repository. > > I will send a note to this list when the fix is in. The version will > also increment, so a simple > > library(Biobase) > update.packages(repos = biocReposList()) > > should get things updated when that happens. > > Best, > > Jim > > > > > -- James W. MacDonald, M.S. Biostatistician Affymetrix and cDNA Microarray Core University of Michigan Cancer Center 1500 E. Medical Center Drive 7410 CCGC Ann Arbor MI 48109 734-647-5623 ********************************************************** Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.