IDR on eCLIP

[Daniel Fulop]

Hi Clipper team,

We have run clipper on several eCLIP datasets and would now like to estimate IDR on the pairs of replicates of each of our conditions.

I know how to get merged eCLIP peaks by using e.g. bedops --merge, but after merging peaks the clipper p-value and fold-enrichment values that I could use to rank peaks for IDR are not associated with the merged peaks.

Hence my question: How do you run IDR on clipper output?  Do you rerun the clipper algorithm on the merged peak intervals instead?  Or do you do something else. 

I tried to do something analogous to what we would do with ChIP-seq data, but it didn't work out well.

Thanks in advance for help!  I'm looking forward to your feedback.

Daniel Fulop.