Entering edit mode
2.2 years ago
gkunz
▴
30
I am looking for an R package that supports simoustanious IDR analysis on more than two ChIP-seq replicates replicates at a time? I have done some searching and was not able to find something that could take process more than input files at a time. If you are aware of a package that can accomplish this in R and are willing to share that info that would be great!
Thanks
By best knowledge, it does not exist. I usually do the IDR analysis on the two samples with the best signal to noise ratio.
Alright, Thank you. I was looking at ChIP-R (which although it sounds like an R program, apparently is not) and trying to find something functionally similar to that in R.
I saw the preprint on Twitter one or two years ago, sounded useful, but in my testing it was a mess, results (for me) made no sense. But take that with a grain of salt, maybe they fixed things, maybe I did it wrong (in that case at least the manual would be to blame), I would probably give it a try if I was you, checking how it goes, also checking on a genome browser what it retains and discards as peaks. IIRC in my case it discarded a lot of quality peaks so I abandoned it. Again, this was my experience, can well be that I did it wrong.
Then there are peak callers such as Genrich and PePr which can take replicates. I used Genrich for ATAC-seq and it worked very well, but it heavily depended on data quality in my hands, and with bad quality you often got no peaks at all while macs still returned many. The issue of getting reproducible peaks in ChIP-seq is still imho an unsolved problem, but usually IDR on the two best samples is ok.