MACS3 callpeak vs HMMRATAC

[DineshRavindraRaju]

Hello All,

With the new Cell Ranger ARC v2, the built-in peak caller has improved compared to earlier versions. I’m wondering about best practices for downstream analysis with Seurat/Signac + MultiVI:

If ARC v2 already produces a higher-quality peak set, is it still recommended to re-call peaks with MACS3?

I’ve seen that HMMRATAC (implemented in MACS3) can give more biologically meaningful peaks . In practice, how does this compare to the ARC v2 peak sets or standard MACS3 in particular to 10xMultiome?

For downstream tasks like integration, clustering, differential accessibility, and peak-to-gene linking — which peak caller tends to give the most reliable results?

I’d really appreciate advice on whether it’s worth re-quantifying with external peak callers after ARC v2, or if ARC peaks are sufficient for most workflows.

Thank you!