Comparative marker selection on rpm-plus1log2 Ion Torrent RNAseq data

[Josephine Klitgaard]

Hi,

I'm comparing a cohort of patient samples, treated vs CTR using the comparative marker selection. The data are generated on the Ion Torrent Analyzer and are rpm plus1 log2 transformed. I choose a pairwise T test and say the data are log transformed, do I need to do antilog or transform the fold changes from comparative marker selection afterwards?

Also, if I need to identify markers which are downregulated with the treatment, are those markers then the ones that are upregulated in CTRs?

Many thanks for your help in advance,

Josephine

[mmr...@cloud.ucsd.edu]

Hi Josephine,

Apologies for the delayed response. Because of the distribution assumptions of the T-test, that is not recommended for read-based data, and RPM is not generally recommended for between-sample comparisons for RNA-seq data. Is the data available as a counts matrix? If so, it is possible to use DESeq2 to calculate differential expression.

Best,
Michael