View Single Post
Old 07-18-2012, 12:12 AM   #2
Senior Member
Location: Cambridge, UK

Join Date: May 2010
Posts: 311

Originally Posted by beefeng123 View Post
I have some RNA-seq data for statistic analysis. I do not have biological replicate for the data, so could anyone can give me some suggestion how to compare the significant expressed genes between two samples using Fisher's exact test in R? Thank you very much.
Fisher's test tests for independency between rows and columns in a table so I don't think you can use it. In fact, with only two samples there is not much that can be done (there has been a lot of discussion on SEQanswers about detecting DE genes from RNAseq in the absence of replication). You can look at the DESeq and edgeR packages from Bioconductor, which can estimate dispersion in absence of replication, but again, the results shouldn't be taken as conclusive.
dariober is offline   Reply With Quote