is it necessary to filter zero-count gene and constraint regulator number

[shangguandong1996]

Hi, I have two question about GENIE3.
First, I am wonering wether I should filter zero-count gene before calculating weights. I found some paper may use the TPM > 5 as cutoff.

And I also found if I set all TF I know (about 3000+) as regulators, the weights is so small while I only input 4 TF, the weight is big. So I am wondering whether I should constraint regulator number.

Best wishes
Guandong Shang