Hello, I would like to use DESeq2 to identify peaks in ChIP-seq or CLIP-seq for given regions. As I have specific regions, I want to use DESeq2 instead of peak caller such as MACS2. Specifically, I have a read counts table for input and IP(pulldown) and want to compare them to identify enriched regions. In the case, can I use default DESeq2 pipeline similar as DiffBind ? Or Do I need to use one-sided test for enriched regions ? As I understand, if I want to check both enriched or depleted regions in IP relative input, I don't have to use one-sided test. But if I want to check enriched regions only, I need to use one-sided test for higher power.

Does choice between one-sided and two-sided test affect distribution of p-values, resulting in multiple test correction ?

I would just use a two sided by default. You will get a small increase in power with one sided but for the clear rejections of null it wont make a big difference.