Interrogating the "unsequenceable" genomic trinucleotide repeat disorders by long-read sequencing. Homo sapiens

Rationale: Microsatellite repeats are abundant in human genomes, and excessive expansion of microsatellites have been found to cause more than 40 neurological, neurodegenerative and neuromuscular disorders. For example, normal ATXN3 has less than 41 CAG repeats, but pathogenic ATXN3 has more than 55 repeats leading to spinocerebellar ataxia type 3 (SCA3), a condition characterized by progressive problems with movement. SCA10 even contains up to 4,500 polymorphic (22.5 kb) repeats. Such kind of much long repeats resulting in most serious diseases are too long to be sequenced using short reads or Sanger sequencing techniques.Goal: We use Pacbio long reads sequencing techniques to detect long reads data of those disease genes, that is, ATXN3 gene here. But long reads data contains higher error rates, and it is inaccurate to infer repeat counts directly. We thus develop RepeatHMM for estimating repeat size from long reads. RepeatHMM can be used to analyze any type of microsatellite with long repeats for helping the analysis of human disorders caused by microsatellite expansion.