Understanding HiChIP data and interactions

Question

[User 7754]

Hi,

I am new to HiChIP data and I was wondering what is the best way to call interactions, using only reads within peaks, or using reads where at least one segment is a peak?

I have seen both approaches for HiChIP so I am not sure which is the way to go. Any thoughts are very much appreciated. Thank you!

Answer 1

[Devon Ryan]

Realistically speaking this will be dependent on how good the ChIP part was. If you got good enrichment then you can be more restrictive and require both reads to be in peaks. This will tend to cut down on lower probability contacts. If you got poorer enrichment then you'll probably need to be less restrictive just to have enough data.

[User 7754]

Thanks very much Devon! By poor/good enrichment do you refer to enrichment of your protein of interest within the anchors, so for example if 80% of the anchors overlap your protein of interest?

[Devon Ryan]

Yup, exactly.