explore and interpret SCENT results

[GGboy-Zzz]

Thanks for your previous reply, I successfully completed the SCENT pipeline in multiome datasets included multiple celltypes. yet I have some questions about exploring and interpreting SCENT prediction results. I inputed known E2G links inferred from SCENIC+ tools, and tried to verify it using SCENT algorithm. Finally, based on the boot_basic_p<0.0005, I filtered out 1.4K out of 13k results. And I plotted the boot_basic_p and beta distribution before and after filter, I noticed there are a group of particularly strong negatively correlated E2Gs.

Do they have biological significance? Is the SCENT algorithm biased toward increasing the true positive rate?
Best regards!