We also investigated the benefits of using Eagle for pre-phasing5 within an existing imputation pipeline: the Sanger Imputation Service, which currently supports imputation using up to N≈32,000 sequenced reference individuals from the Haplotype Reference Consortium (HRC; see URLs). (We note that the HRC is predominantly European and contains a substantial fraction of UK samples but also contains samples of other ancestries; see URLs.) We considered two fast pre-phasing procedures: Eagle pre-phasing of all N≈150,000 UK Biobank samples and SHAPEIT2 10×15K pre-phasing of N≈150,000 samples. To benchmark imputation accuracy, we completely masked 700 SNPs (100 in each of seven MAF bins) in each of three chromosomes, pre-phased the remaining SNPs with Eagle and SHAPEIT2, imputed the same subset of N≈15,000 pre-phased samples using the Sanger Imputation Service, and computed R2 between the masked SNPs and their imputed genotype dosages across curated British samples (Online Methods; see URLs). This benchmarking procedure is commonly used to assess the accuracy of phasing and imputation pipelines5,9. We observed that when imputation was performed using the largest reference panel available (the N≈32,000 HRC), Eagle pre-phasing using
