For eCLIP-seq and iCLIP-seq, should I use the left or center position of the peaks as the binding site?

Question

[li487]

I want to know the accurate (single nucleotide resolution) binding sites in eCLIP-seq experiment. I have many peaks, so the center position or the left position (upstream) in each peak are the binding sites?

Answer 1

[i.sudbery]

Are these "peaks" or reads?

For reads, its seems that the cross-linked base depends on the protein being studied. Although the original thought with iCLIP was that the cross-linked base would be the one upstream of the first base of the read, as this where the reverse transcriptase would terminal, recent evidence suggest that this isn't always true, and that for some proteins, the reverse transcriptase can skip the crosslink. You can use the tool iCLIPro to decide which type of data you have. If it turns out your experiment is centre enriched, then the experiment effectively become a CLIP experiment and you will not have base-pair resolution.

For peaks The binding could be anywhere within the peak, although I guess the centre is probably more likely than right at the edge. Ideally you'd want the summit of the peak - that is the base with the largest number of tags on it, but whether you have that or not probably depends on the software used to call the peaks.