RPKM and statistical test

[paolo.kunder]

Dear All,
I have an output with RPKM values for each transcript (using LifeScope ABI), each column represents a different sample ( I have both techincal and biological replicates).
What kind of test to asses statistical difference may I do?

I am using DESeq and DEXseq with rawcounts, but no clue ho to analyze directly RPKM values,

Thanks,
Paolo

[joxcargator73]

Same question

Hello Paolo,
I am using RPKM values too. I was wondering if there is a good test for analyzing these values. ?

[rfilbert]

I think that Partek's Gene Specific Analysis is most appropriate. Rather than assuming a distribution (lognormal, poisson, etc.) like DESeq and DEXSeq, they run multiple tests and choose the best model fit.

[chadn737]

Don't use RPKM values. stick with count based methods. In particular either DESeq or EdgeR.

http://bib.oxfordjournals.org/content/early/2012/09/15/bib.bbs046.long