Hi everyone, A colleague of mine has been using BioC to do his analyses on his custom stickleback array, and asked me to check to make sure he was doing everything correctly. I'm not sure about one part, so I'm asking the list. The stickleback genome is not well annotated in regards to GO terms, but he has mapped all the stickleback microarray probes to zebrafish EntrezGene IDs. Instead of making his own chip-specific AnnotationDBI package, he's doing GO testing using the zebrafish.db packages from BioC. Everything seems to be running fine and there are no error messages, but I'm not sure if there aren't some hidden problems. For example, what if some of his zebrafish EntrezGene IDs are not on the Affy zebrafish array? Would it be fine as long as they were in the org.Dr.eg.db based data package? According to the GOstats vignette on Hypergeometric Tests, when you set up your GOHyperGParams object, you have to specify a chip-specific AnnotationDBI package, but you input your selected genes and universe as (usually) EntrezGene IDs, not probe set IDs. This makes me wonder if hyperGTest is doing everything through the zebrafish.db package, or going directly to the org.Dr.eg.db package? He hasn't checked yet to see if all of his EntrezGene IDs are or are not on the Affy zebrafish array, and if they are then I guess there is no problem in simply using the zebrafish.db package. Mainly my question is why does the GOstats hyperGTest() require a chip-specific annotation package when you have to input EntrezGene IDs (the mapping key for the org.Xx.eg.db packages), not probe set IDs? Could hyperGTest() be (easily) modified to directly accept the org.Xx.eg.db packages? Thanks, Jenny Jenny Drnevich, Ph.D. Functional Genomics Bioinformatics Specialist W.M. Keck Center for Comparative and Functional Genomics Roy J. Carver Biotechnology Center University of Illinois, Urbana-Champaign 330 ERML 1201 W. Gregory Dr. Urbana, IL 61801 USA ph: 217-244-7355 fax: 217-265-5066 e-mail: drnevich at illinois.edu

<snip> > Mainly my > question is why does the GOstats hyperGTest() require a chip- specific > annotation package when you have to input EntrezGene IDs (the > mapping key for the org.Xx.eg.db packages), not probe set > IDs? Could hyperGTest() be (easily) modified to directly accept the > org.Xx.eg.db packages? > It already does :) At least when I use hyperGTest() from the Category package for GO/KEGG/Chrlocation, it works fine with org.Mm.eg.db. I actually don't know whether there are any differences between hyperGTest() from GOstats and the one from Category. Cheers \Heidi > Thanks, > Jenny > > Jenny Drnevich, Ph.D. > > Functional Genomics Bioinformatics Specialist > W.M. Keck Center for Comparative and Functional Genomics > Roy J. Carver Biotechnology Center > University of Illinois, Urbana-Champaign > > 330 ERML > 1201 W. Gregory Dr. > Urbana, IL 61801 > USA > > ph: 217-244-7355 > fax: 217-265-5066 > e-mail: drnevich at illinois.edu > > _______________________________________________ > Bioconductor mailing list > 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 >