For the final credible sets datasets, we need to be able to decide which credible sets are included in the outputs and which ones are discarded. For example, the same locus/region might have been fine-mapped through different methods, so we need an algorithm that decides which credible set will be included in the final set.

Because inclusion and validation both involve filtering rows in credible sets, we want to reuse this step to implement this logic. Also, parts of the validation (e.g., duplicate studyLocusId) would only make sense after the inclusion filter.

This ticket only covers one of the required QCs. PICS disambiguation would require a different QC flag.

QC - "Explained by Susie credible set in the region"

• For every credible set
  • if there is another credible set in the same STUDY and REGION as the credible set LEADVARIANT
    • if the lead has NOT a flag reporting the variant is not in the LD matrix
      • flag credible set as "Explained by SuSie credible set in the region"

QC method needs to be added to StudyLocus and QC reason to the enum.

The method must be called from the study_locus_validation step before checking the duplicated studyLocusIds.
The orchestration repo will then need an additional flag to run this validation.