FYI: VEP annotation inconsistencies gnomad.utils

Science
1

FYI. I ran VEP on my Matrix table (hl.vep) and also ran gnomad.utils.process_consequences, comparing the following terms:

  • rows.vep.most_severe_consequence (from VEP itself)
  • rows.vep.worst_consequence_term (from gnomad utils)
  • rows.vep.worst_csq_for_variant.most_severe_consequence (from gnomad utils)

The latter two match up, but in some edge cases the first does not match the latter two:


 1:50 PM
V   most_severe_consequence worst_consequence_term  worst_csq_for_variant  maf gene_symbol ld_indep
1:19325533:G:A  synonymous_variant  stop_retained_variant   stop_retained_variant   3.1069e-04  PQLC2   false
7:73596424:C:T  synonymous_variant  stop_retained_variant   stop_retained_variant   4.7609e-02  MLXIPL  false
11:60178272:T:C synonymous_variant  stop_retained_variant   stop_retained_variant   4.0228e-01  MS4A6A  false
12:47784123:C:T synonymous_variant  stop_retained_variant   stop_retained_variant   1.5470e-01  HDAC7   true
14:90301263:C:T synonymous_variant  stop_retained_variant   stop_retained_variant   3.3496e-01  NRDE2   false
16:11850392:C:T synonymous_variant  stop_retained_variant   stop_retained_variant   1.1161e-04  RSL1D1  true
19:11977067:G:A synonymous_variant  stop_retained_variant   stop_retained_variant   6.0644e-03  AC008770.2  false
19:51016003:C:T synonymous_variant  stop_retained_variant   stop_retained_variant   4.7594e-01  KLK10   false
20:44939185:A:G synonymous_variant  stop_retained_variant   stop_retained_variant   1.5596e-03  PABPC1L true

Just be wary and be consistent with your VEP annotation schema!