Negative p-values from SKAT-O

[jbgaither]

I have SeqSpark working, but some of my SKAT-O runs have resulted in negative p-values (e.g. -3e-7).

Obviously this is incorrect. But wondering if this is is a known issue in SeqSpark/SKAT-O in general? Also, do the authors have any idea of the cause? Thanks!