Genetic architecture of routinely acquired blood tests in a UK cohort of South Asian ancestry reveals ancestry-specific causal variants

Abstract Understanding the genetic basis of blood tests routinely acquired in clinical care can provide insights into several aspects of human physiology and disease. Previous efforts have almost exclusively focussed on individuals of European descent, severely limiting generalizability to ancestrally-diverse populations. Here we describe the largest genome-wide association study of quantitative traits in a cohort of South Asian ancestry. We tested the association of >7 million autosomal, common (minor allele frequency [MAF] > 1%) Single Nucleotide Polymorphisms (SNPs) genotyped using the Illumina Global Screening Array version 3 and imputed from the TOPMED-r2 panel. We defined 39 quantitative blood test traits from Electronic Healthcare Records (EHRs) – both primary care and hospital records - of ~40,000 British Bangladeshi and British Pakistani adults. We discover eight novel locus-trait associations which were specific to Genes & Health, including CUX2 (serum vitamin D), BRD3OS (glycated haemoglobin) and STMN2 (lymphocyte count). By performing multi-ancestry meta-analysis and fine mapping using data from UK Biobank, we demonstrate ancestry-specific causal variants within the PIEZO1 and HBB loci which were associated with alterations in red cell traits and glycated haemoglobin. The South Asian ancestry-specific causal signal for glycated haemoglobin, rs563555492(MAF 3.2% in 1,000 Genomes South Asian ancestry samples), is a missense variant in PIEZO1 which is monomorphic in other ancestries, and illustrates how ancestry-specific genetic variants may be of relevance for the design of clinical reference ranges. Our results shed light on the genetic basis of clinically-relevant quantitative traits in an under-represented population, and emphasise the importance of ancestral diversity in genetic studies.