I've to transform some RNA-seq read counts into linear regressions algorithms that were originally written for microarray data.
Not that I don't know how to transform read counts for linear regression; the problem is there're a few methods available--what I can think of right now include log(FPKM), rlog from DESeq, and voom--and I haven't noticed anyone comparing different transformation methods.
So, personally, which one do you prefer? If the algorithm used limma, I'd have used voom, but this is not the case.
Not that I don't know how to transform read counts for linear regression; the problem is there're a few methods available--what I can think of right now include log(FPKM), rlog from DESeq, and voom--and I haven't noticed anyone comparing different transformation methods.
So, personally, which one do you prefer? If the algorithm used limma, I'd have used voom, but this is not the case.