Unexpected frequency of the pathogenic AR CAG repeat expansion in the general population

Abstract

CAG repeat expansions in exon 1 of the AR gene on the X chromosome cause spinal and bulbar muscular atrophy, a male-specific progressive neuromuscular disorder associated with a variety of extra-neurological symptoms. The disease has a reported male prevalence of approximately 1:30 000 or less, but the AR repeat expansion frequency is unknown. We established a pipeline, which combines the use of the ExpansionHunter tool and visual validation, to detect AR CAG expansion on whole-genome sequencing data, benchmarked it to fragment PCR sizing, and applied it to 74 277 unrelated individuals from four large cohorts. Our pipeline showed sensitivity of 100% [95% confidence interval (CI) 90.8-100%], specificity of 99% (95% CI 94.2-99.7%), and a positive predictive value of 97.4% (95% CI 84.4-99.6%). We found the mutation frequency to be 1:3182 (95% CI 1:2309-1:4386, n = 117 734) X chromosomes-10 times more frequent than the reported disease prevalence. Modelling using the novel mutation frequency led to estimate disease prevalence of 1:6887 males, more than four times more frequent than the reported disease prevalence. This discrepancy is possibly due to underdiagnosis of this neuromuscular condition, reduced penetrance, and/or pleomorphic clinical manifestations.

Keywords: androgen receptor; bioinformatics; population genetics; spinal and bulbar muscular atrophy; whole-genome sequencing.

Figure 1

Validation of the WGS pipeline. (A) Visualization of repeat expansion reads from ExpansionHunter shows reads revealing 23 and 41 CAG repeat alleles. (B) Comparison of repeat size estimation between WGS pipeline and PCR. n = 133 alleles. Dark points indicate length confirmed by reads spanning the whole repeat and both the flanking sides; light points indicate length confirmed by reads spanning part of the repeat and one flanking side.

Figure 2

WGS pipeline detects increased AR CAG expansion in four large cohorts. (A) Allele size distribution across 75 035 100k GP genomes; inset highlights the distribution of alleles containing ≥34 repeats. (B) Frequency estimation of AR CAG expansion. WGS pipeline detects 1:3182 AR expansion ≥38 repeats in the pooled 100k GP, gnomAD, NIH, and MinE cohorts. Error bars show 95% CI. (C) Top) Cumulative distribution of SBMA age of onset for n = 983 KD cases from the most recent KD meta-analysis; (middle) distribution of the English male population (n = 27 827 831); and (bottom) resulting estimated prevalence of SBMA by age group, considering the literature reported male prevalence of 1:30 303 or less (dark area), and the novel estimated prevalence according to our WGS result (1:6887 males, light area).