Extended Data Fig. 5: Somatic instability of long TCF4 repeat alleles.

a, Mean estimated length (in repeat units) of long TCF4 alleles (≥ 45 repeat units) in UKB participants of different ages. Heterozygous carriers of long TCF4 alleles were first stratified into quintiles of imputed TCF4 allele length, a proxy for inherited allele length. b, Mean number of IRR pairs observed per UKB participant heterozygous for a long TCF4 allele, again stratified by imputed TCF4 allele length and by age. In both a and b, analyses were restricted to 38,558 individuals carrying no other long CAG repeat except possibly in CA10 (such that IRR pairs could be assumed to have originated from TCF4). Error bars, 95% CIs. c, TCF4 allele lengths directly measured from long-read sequencing of carriers of long alleles in AoU. Each horizontal line corresponds to a single AoU participant; black markers indicate repeat lengths observed in long reads that span the TCF4 repeat (dots) or partially overlap the repeat (pluses, which lower-bound allele lengths), while blue crosses indicate allele lengths estimated from short-read WGS. Long TCF4 alleles exhibit somatic mosaicism, with alleles sometimes varying in length by hundreds of repeat units within blood cells from the same individual, indicating high somatic instability. We have received an exception from the All of Us Resource Access Board to disseminate participants counts less than 20.