How should I run LDSC on a small annotation?

[JeGrundman]

Hi,

I hope all is well. I have cell-specific open chromatin regions that I'm trying to run S-LDSC with. The problem is that the annotations resulting from these files are very small and don't cover 1% of the genome, which I've read is the heuristic used to determine whether the annotation is large enough to run S-LDSC with.

I also saw a medRxiv paper, "Significance testing for small annotations in stratified LD-Score regression," which talks about how another threshold for whether you should run S-LDSC with your annotation is if it overlaps 1.7% of 0.01 centimorgan windows.

My questions are as follows:

1. If my annotation is quite small and doesn't meet the 1% genome coverage threshold, what is an acceptable way forward to work with S-LDSC and the annotations? To meet that 1% coverage threshold, I'd have to add rather large (100 kb) windows to my peaks, which seems excessive.
2. What did the authors of the paper I mentioned use for that genome that had 0.01 cm bins? Is there a more specific set of instructions for how they carried out this test? (I did read the methods of the paper, but it's not clear to me how to go about testing this criteria for my own annotations).

Thanks in advance!

[roxyisat-rex]

I have the exact same question. Does anyone have an answer to this? Thank you.