Genome-wide association studies (GWAS) have grown exponentially
over the past 15 years, rapidly increasing in statistical power to enable
the identification of hundreds of thousands of associations between
genetic variants and human traits 1 . While these discoveries have been
facilitated in part by precipitous drops in sequencing costs, microarrays
have been the primary technology used for GWAS to date because of
their lower costs. However, by design, they have biased ascertainment
of genetic variants; sites that are included on many GWAS arrays, such as
the widely used Illumina Global Screening Array (GSA) or Global Diver-
sity Array, are most common in European ancestry populations. Previ-
ous work has shown that low-coverage sequencing is a cost-effective
alternative that can more accurately capture genetic variants across the
allele frequency spectrum for variants present in imputation reference
panels 2,3. Low-coverage sequencing is especially useful in populations
underrepresented in genomics, even compared with GWAS arrays that
have been designed to reflect variation within those populations, such
as the H3Africa GWAS Array.