Reconcile discrepancies in how coverage data is stored/reference genomes used in different gnomad versions.

[jkunisak]

gnomad v3:

• reference genome: hg38
• coverage: tsv file for all chromosomes (WGS only)
• variants: individual chr VCFs or all chromosome hail table --> convert hail table to vcf with appropriate INFO fields and annotations

gnomad v2:

• reference genome: hg38 (liftover)
  • coverage: NA
  • variants: all chr VCF file (and hail table) --> split by WES and WGS
• reference genome: hg19
  • coverage: Split by WES and WGS
  • variants: all chr VCF file (and hail table) --> split by WES and WGS

Exac

• reference genome: hg19
• coverage: individual chr TSV (WES only)
• variants: all chr VCF file