damiansm
commented
4 years ago Still same issues for 2003 build but all other updates fine. Decided after consultation with mitochondrial genome experts to just blacklist these two Mt variants from our whitelist
For the 1909 variant builds we attempted to regenerate the clinvar whitelist files but ran into issues when testing and reverted back to using the old 1902 files. Variants drop in and out of the clinvar weekly builds (or change status) and needs further investigation whether this is correct and how best to handle.
In particular, two very common, synonymous mitochondrial variants (rs2853493 and rs2853499) are now considered likely pathogenic and become the top hit in several solved 100KGP cases rather than the correct diagnosis that was previously top:
tabix 1909_hg19/1909_hg19_clinvar_whitelist.tsv.gz 25:11467-11467 25 11467 A G ALLELEID=513402;CLNSIG=LIKELY_PATHOGENIC;CLNREVSTAT=criteria_provided,_single_submitter
Also an expert reviewed pathogenic variant (https://www.ncbi.nlm.nih.gov/clinvar/variation/17004/) was dropped in the weekly release we used but back the week after