search

clingen-data-model / clingen-hail-reports

Performs filtering against gnomAD and ClinVar datasets. Uses Hail to report records with a population FAF above certain thresholds by gene.
1 stars 0 forks source link

Enable overlapping and non-same directional thresholds in FAF notebook #6

Open theferrit32 opened 3 years ago

theferrit32 commented 3 years ago

Right now it returns variants that are over certain specified frequency thresholds in gnomad. But for pathogenic thresholds we want to do the reverse, returning those less than the threshold. So we need to specify the directionality condition on each input criteria.

We also need to enable returning all of the thresholds that are met by a variant, as the criteria will now be overlapping in an inconsistent way and returning the highest threshold isn't necessarily the best.

theferrit32 commented 3 years ago

Example of input criteria:

Gene Criteria Threshold Condition Transcript
MYH7 BA1 0.10% >= NM_000257.4
MYH7 BS1 0.02% >= NM_000257.4
PKD1 PM2 0.0005% < NM_001009944.3
theferrit32 commented 3 years ago

Related to a data request, there should also be support for a catch-all threshold like that applies to variants not already included in another threshold:

Gene Criteria Threshold Condition Transcript
RUNX1 ANY 0.0% >=
KelseaChang5 commented 2 years ago

@theferrit32 Is this still a need/is something we will realistically do in the near future?

theferrit32 commented 2 years ago

@KelseaChang5 I think there is no reason to spend the time to implement this until there is need to run a report using these sorts of criteria. We can archive/icebox for now.