CuffDiff_Library Normalization Options

[mpalegre]

Hi everybody. I have RNA-Seq single-end data (Ion-Torrent) and I want to study DE using Cuffdiff. Anyone knows which is the best "lbrary normalization" options? --quantile --classic-fpkm or --geometric. i don´t understand the difference between them.

Thanks in advance

Regards

[sdriscoll]

Probably --geometric. They go like this:

--quantile normalizes the samples to each sample's upper-quartile read count
--geometic normalizes the samples to each sample's median read count level
--classic-fpkm is a double normalization that normalizes each gene's read count by the gene length and by the total millions of mapped reads for the experiment.

[mpalegre]

Ok. Thanks. A problem I have now is I have many reads of one sample because I have 2 Runs of 318chip sample 1 and 3 Runs of another so the diferrence between number of reads is important.