Whole Genome Sequencing Coverage & Quality at Prenatal Carrier Screening Loci: An Analysis of 15,496 gnomAD Genomes [23K

Abstract

INTRODUCTION: Current expanded prenatal carrier screening (EPCS) tests leverage several technologies to analyze genomic regions associated with Mendelian disorders. As whole genome sequencing (WGS) becomes more widely utilized some have speculated WGS could replace current EPCS methods. However, a large-scale technical analysis of WGS coverage and genotype dropout at EPCS loci is, to our knowledge, unpublished. METHODS: In order to assess the potential to use Illumina WGS methods for EPCS, we downloaded autosomal data for 15,496 genomes (gnomAD browser). We analyzed read coverage at ∼3,400 EPCS loci and frequency of missing/low-quality genotype calls at ∼1,300 EPCS loci. RESULTS: We found that WGS provided an average and median coverage of 32 and 31, respectively, across the ∼3,400 EPCS loci. Coverage predicted to yield <99% accuracy (≤10 reads) was found in an average of 17/15,496 genomes (median 11, range 0–728) across the ∼3,400 loci. Considering each of the ∼1,300 loci with gnomAD genotype calls, we found a missing genotype call in an average of 52/15,496 genomes (range 2–11,529). Simulations using a negative binomial distribution predict WGS coverage ≥65X is needed to achieve 99.99% adequate coverage (≥20X) across all genotypes. CONCLUSION: This investigation demonstrates that the methods used to generate gnomAD WGS data would provide missing genotype calls that fail to meet clinical standards for a maximum of ∼74% of analyzed genomes; 6 loci had missing genotype calls in more than 500 genomes. Targeted EPCS panels or Sanger confirmation would be required to supplement these WGS results.