We analyzed data from the UK Biobank consisting of 152,729 samples typed at ≈800,000 SNPs. Using PLINK31 (see URLs), we removed 480 individuals marked for exclusion from genomic analyses based on missingness and heterozygosity filters and 1 individual who had withdrawn consent, leaving 152,248 samples (see URLs, UK Biobank Genotyping and QC). We restricted the SNP set to autosomal, biallelic SNPs with missingness ≤10% and we further excluded 65 autosomal SNPs found to have significantly different allele frequencies between the UK BiLEVE array and the UK Biobank array, leaving 707,524 SNPs (57,753 on chr1, 41,538 on chr5, 34,588 on chr10, 22,367 on chr15, and 18,349 on chr20). We identified 72 trios based on IBS0<0.001, sex of parents, and age of trio members (see URLs, Genotyping and QC). Of the 72 trio children, 69 self-reported British ethnicity, one self-reported Indian ethnicity, and one self-reported Caribbean ethnicity. The remaining trio child did not self-report any ethnicity, but her parents self-reported Irish and “Any other white background” as their ethnicities, so we included this trio child in the 70 European-ancestry trio children we used to benchmark phasing accuracy.