Closed gtiao closed 10 months ago
I think maybe this was because when we have the gnomAD v4 variant QC, if RF looks good, we might want to see how well the UKB model would perform on the gnomAD v4 data and how well the a full training (or maybe even a training without the UKB data) performs on just the UKB data?
closing because of Julia's comment ("if RF looks good") -- we chose AS VQSR for v4
Is this task still relevant/needed?