Dear goseq developers：

      Hello ,my name is Xu,ZhengZheng ,a student of Beijing Institute of Gennomics ,Chinese Academics of Science. After I use edgeR to do difference Gene expression ,I happy to known that I can use R package(goseq) to do GO analysis, thanks for your good job for this package, But when I use it I come cross some problems,I'm confused about that,because I strictly follow the goseq manual;I don't known how to conquer them:

Question1:why pwf ojects come cross NA,but also still shows DE("1")?

http://pan.baidu.com/s/1ufoBW#path=%252FR_question&render-type=grid-view

Question2:when I use goseq(pwf,"hg38","refGene"),it pop-up code with errors:  

in names(tmp) = rep(names(map),times=as.numeric(summary(map)[,:attempt to set an attribute on NULL

http://pan.baidu.com/s/1ufoBW#path=%252FR_question&render-type=grid-view

the URL "http://pan.baidu.com/s/1ufoBW" has the problem picture and the rawdata and my analysis code and error;

thanks 

yours

sincerely!

Most of the geneids in your data file are not actual gene symbols. There is no way to do a GO analysis unless you provide valid gene symbols or identifiers.

Hi,

It looks like you are using gene symbols rather than refseq IDs. For questions 2, try running goseq(pwf,"hg38","geneSymbol") and see if that works. You should all pass "geneSymbol" to the nullp function rather than "refGene".

Cheers, Nadia.