fail to run AUCell_direct due to memory.

I'm working on 160k cell multiome dataset. I started from more than 200k but I downsampled it.
Ive been running one rule at a time using the snakemake pipline.
I'm using 1Tb memory, and 20-70 cores, it keeps crashing during the enrichment score calculation in AUcell_direct step.
Do you have any suggestion on this issue? I cannot request more resources than 1Tb and 70 cores.

Thanks!
