I have sequenced human miRNA (total RNA enriched for 15-25nt) in different conditions. I mapped my reads with a minimum overlap of 15 + perfect match to the human genome, and extracted read counts using bedtools of miRBase miRNA annotations.
I now want to generally compare my different conditions and see which ones are more similar to each other at the small RNA level. I remember it is possible to do that with cummeRBound package in R, but it require cufflinks output, and it is my understanding that cufflinks is not necessarly a good option for miRNA.
Anyone have any idea how to do this?
Also, I read in some papers that people add a pseudocount of 1 to all read counts so as not to encounter division by 0 errors.
Bioconductor's RNAseq workflow may be of help. There are sections on PCA and MDS, and the data expected are just read counts, so your bedtools counts should suffice.
Nice, I see it. Thanks, this is the answer.