Question regarding fig. 2 in your paper

[zhyanlin]

Hello,

I have a question regarding figure 2 in your paper. The figure said the recalls are 60%~80% when using ChIA-PET or HiChIP data as target. But peakachu's prediction is around 15K, and the number of peaks in CTCF ChIA-PET is almost 100K. Did you pool ChIA-PET as well?

[tariks]

Yes, we pooled ChIA-PET coordinates and allowed for a +/- 1 bin tolerance when determining matches.

[zhyanlin]

Thanks. It makes more sense!

How did you pool the ChIA-PET data? When I tried to use the peakachu's pooling tool to pool ChIP-PET, I got 0 loops.

I covert ChIA-PET data to the input of pool tool as follows:
chr9 99230000 99235000 chr9 100010000 100015000
-->
chr9 99230000 99235000 chr9 100010000 100015000 1 IF

Thanks!