Of these 1,325 individuals, 381 were one member of a monozygotic (MZ) twin pair. Zygosity has been validated in this sample through questionnaire, in-person review of the appearance of the twins by experts, anthropomorphic measurements, DNA concordances for all fraternal twins, and DNA concordances for many MZ twin pairs, resulting in a zygosity diagnosis error rate well under 1%. To maximize the yield of phenotypic data, co-twins of MZ twins who were sequenced were added to the sample. Genotypes were simply copied from the sequenced MZ twin to his/her unsequenced co-twin, under the assumption that the sequence was identical between MZ twins. This is a reasonable assumption here, especially for common and low-frequency SNPs, because differences between members of an MZ twin pair would be indistinguishable from sequencing errors at the sequencing depth obtained here (10×). For rare variants, specifically singletons, copying genotypes from one MZ twin to the other will introduce errors due to somatic mutations (Poduri, Evrony, Cai, & Walsh, 2013), which are present only in one MZ twin and not the other. If we assume that each
