Dear Bio Communities,
Now I'm try using RSEM expected count data to perform GSVA analysis. But this kind of count data doesn't the integer count, which is usually calculated by HT-Seq et al. As the author of GSVA suggested (Why Negative value from GSVA), the row count data of RNA-Seq calculated by HT-Seq or STAR should be normalized by logCPM, which is then fed into GSVA function. Should I do the same procedures for the expected count data? Any suggestions would be greatly appreciation!
dge <- DGEList(expected_count) dge.norm <- calcNormFactors(dge) normLogCPMs <- cpm(dge.norm, log=TRUE, normalized.lib.sizes=TRUE) gsva_es <- gsva(normLogCPMs, gset.idx.list=hallmark)