"Duplicate mutation events" import error

[ao508]

Notes:

Studies known to be affected:

• mskimpact
• lymphoma super cohort
• mskimpact_ped
• mixed_cbe_rudinc_CR_pdx2 (yichao's study)
• glioma_dfci_2020 (ramya's study)

These studies import without issues into cgds_triage. Error is only thrown during import into cgds_gdac.

Completed Items:

• Confirmed that duplicates are not present based on key fields (entrez id, chrom, start, end, ref allele, tum allele, variant type, sample id, hgvsp_short)

  cut -f2,5,6,7,11,14,40 private/glioma_dfci_2020/data_mutations_extended.txt

• Ran the check for duplicates manually on both cgds_triage and cgds_gdac - both databases pass the check. Maybe check with Java to determine if there's a difference between the mysql check and java check?

  SELECT 
  (SELECT COUNT(*) FROM mutation_event) = 
  (SELECT COUNT(DISTINCT ENTREZ_GENE_ID, CHR, START_POSITION, END_POSITION, TUMOR_SEQ_ALLELE, PROTEIN_CHANGE, MUTATION_TYPE) FROM mutation_event);

• built new importer with changes to cbioportal core with additional logging for debugging

To-do:

• import with updated importer and attach debugger to importer to evaluate the state of the database as one of the affected studies is attempting to import

[ao508]

cbioportal PR: https://github.com/cBioPortal/cbioportal/pull/7835

There will be 3 follow up PRs to these repos as well to update the commit hashes for jitpack:

• genome-nexus-annotation-pipeline https://github.com/genome-nexus/genome-nexus-annotation-pipeline/pull/143
• cmo-pipelines
• pipelines

[ao508]

TO-DO:

• Script is running to generate a set of affected mutation_event_ids and affected cancer_studies
• Once these are finished the cancer studies should be removed and reimported
• We should also purge the affected mutation_event_ids ** this is done automatically as a cleanup step that executes after a cancer study has been deleted