Questions on simulating non-TF knockout

[lh12565]

Hi, I'm trying to use CellOracle to simulate the knockout of CSMD1, a non-TF gene evident in my ATAC-seq data (high degree and eigenvector centrality). According to the documentation, the knockout simulation seems designed for TFs. Is there a recommended approach or workaround to investigate the role of such non-TF genes using CellOracle?
I have run both SCENIC and CellOracle on my dataset. The list of cell-type-specific TFs identified by each tool shows significant discrepancies. Could you provide some guidance on how to interpret these differences and whether there are parameters in the Base GRN construction process that I could adjust to make the results more biologically plausible?
Thanks!