Open DrZhaoJie opened 3 years ago
Hi, did you download the FPKM values from GDC? If yes this might explain it, because in TCIA we used the expression data provided by CGHub hosted by UCSC back in 2015 (they used a different RNAseq pipeline than GDC is now using). If you download the expression data from TCIA and run IPS.R then you should get matching results.
Hi, I really enjoy your package. I downloaded TCGA RNAseq fpkm value and try to produce the IPS as follows
FPKMtoTPM <- function(fpkm){ exp(log(fpkm) - log(sum(fpkm)) + log(1e6)) } RNA_TPM <- as.data.frame(apply(RNA_FPKM,2,FPKMtoTPM)) gene_expression <- RNA_TPM
And then I run the code as you provided in IPS.R. But it had different IPS score as you presented in the TCIA web. Where was I wrong? Thank you very much.