Evaluating MAST model fit using AIC

Question

[mhale]

Hi all,

Reaching out to see if anyone has experience determining AIC values for models fit by MAST. I'm using MAST for differential expression analysis in scRNAseq data sets, and it would be really helpful to have a quantitative measure to evaluate model fit and guide inclusion/exclusion decisions for covariates. Other, non-AIC-based methods, would be equally welcomed.

Thanks.