As a developer I want to give support to COLOC colocalisation as one of the steps in the ETL because we now have credible sets finemapped with SuSIE that we can colocalise using this method.
It will be useful to compare results between COLOC and eCAVIAR.
Background
The current colocalisation step only outputs results using eCAVIAR, which uses PIPs to calculate colocalisation PIPs.
Now that we have credible sets with Bayes Factors (finemapped with SuSIE) from Finngen and the eQTL Catalogue, we are in a position of running COLOC and presumably get more accurate colocalisation metrics.
Tasks
[ ] Update colocalisation.py to generate COLOC results for those credible sets where finemappingMethod == SuSie
I am thinking of writing both datasets independently. No schema changes are required.
Acceptance tests
How do we know the task is complete?
When I check the output of this step, colocalisationMethod has 2 values: COLOC and eCAVIAR.
When I check the output of eCAVIAR, there are no missing colocalising credible sets in comparison to the 24.01 data.
As a developer I want to give support to COLOC colocalisation as one of the steps in the ETL because we now have credible sets finemapped with SuSIE that we can colocalise using this method. It will be useful to compare results between COLOC and eCAVIAR.
Background
The current colocalisation step only outputs results using eCAVIAR, which uses PIPs to calculate colocalisation PIPs. Now that we have credible sets with Bayes Factors (finemapped with SuSIE) from Finngen and the eQTL Catalogue, we are in a position of running COLOC and presumably get more accurate colocalisation metrics.
Tasks
colocalisation.pyto generate COLOC results for those credible sets wherefinemappingMethod == SuSieI am thinking of writing both datasets independently. No schema changes are required.
Acceptance tests
How do we know the task is complete?
colocalisationMethodhas 2 values: COLOC and eCAVIAR.