Unrecognized high prevalence of expanded composite repeats in Friedreich ataxia

Many diseases are caused by pathogenic expansion of microsatellite repeats. Longread sequencing allows evaluation of the content of such expanded repeats. Friedreich ataxia patients are typically homozygous for an expanded GAA repeat in intron 1 of the FXN gene. Longread whole genome sequencing identified expanded composite alleles, consisting of substantial tracks of tandem GGA triplets within the expanded GAA repeat. In a prospective series of 112 unrelated patients, we found that approximately 20% of people with Friedreich ataxia have at least one such expanded composite allele. Other minor sequence interruptions in the expanded GAA repeat were detected in a further 10% of patients. Most expanded composite alleles revealed by longread genome sequencing are not detectable by standard PCR-based testing, and have therefore remained hidden despite their relatively high prevalence. This results in erroneous genotyping of patients and heterozygous carriers. We describe an optimized workflow to detect these expanded composite alleles, which permitted accurate genotyping and heterozygous carrier identification. A recurrent proximal FXN gene deletion caused by Alu-mediated non-homologous recombination was identified in an additional 2% of patients. These findings redefine the spectrum of pathogenic alleles in Friedreich ataxia, and demonstrate that expanded alleles containing substantial non-GAA interruptions are prevalent and pathogenic.