Replication of SKAT and Monti p-values

[OmerRonen]

Dear deeprvat authors,

Thank you for your great contribution and well organized code base.
We are trying to replicate your empirical results for the SKAT-O and Monti methods, and we would like to ask some clarifying questions.

1. For the calculation of the p-values, did you use a permutation based approach as in Monti et al, or simply used the p-value obtained by SEAK?
2. For the SKAT-O pvalue did you consider only variants with impact score higher than 0.8 or the full set of variants? How would the current code-base behave under var_weight_function == "beta_maf" in comparison_methods/monti/monti_association.py?

Thanks again for your great contribution!