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

Abstract

Microsatellite expansion, such as trinucleotide repeat expansion (TRE), is known to cause a number of genetic diseases. Sanger sequencing and next-generation short-read sequencing are unable to interrogate TRE reliably. We developed a novel algorithm called RepeatHMM to estimate repeat counts from long-read sequencing data. Evaluation on simulation data, real amplicon sequencing data on two repeat expansion disorders, and whole-genome sequencing data generated by PacBio and Oxford Nanopore technologies showed superior performance over competing approaches. We concluded that long-read sequencing coupled with RepeatHMM can estimate repeat counts on microsatellites and can interrogate the "unsequenceable" genomic trinucleotide repeat disorders.

Keywords: Long-read sequencing; Microsatellites; Nanopore; PacBio; RepeatHMM; Trinucleotide repeat disorders; Trinucleotide repeats.

Fig. 1

A flowchart of the procedure to infer repeat counts using RepeatHMM

Fig. 2

Analysis on simulation data to infer repeat counts for ATN1. a Performance on simulated long reads with random start and end sites that cover repeats. b Performance on simulated long reads with fixed start and end sites that cover repeats. c, d The distribution of the prediction errors (estimated repeat counts minus simulated counts) on random simulation data and PCR-based simulation data, respectively. RMSE root mean square error between simulated repeat counts and estimated counts for 100 participants

Fig. 3

Performance of RepeatHMM, RepeatCCS, BAMself, and TRhist on estimating the repeat counts in ATXN3 for 20 patients with SCA3 and five controls. The gold standards (x-axis) were determined by capillary electrophoresis for 20 patients or by Sanger sequencing for five controls. a Scatterplot of estimated repeat counts and true counts. b, c The difference of estimated repeat counts and true counts by RepeatHMM, RepeatCCS, BAMself, and TRhist. RepeatCCS refers to the use of RepeatHMM on error-corrected reads generated by the circular consensus sequencing protocol

Fig. 4

The distribution of repeat counts estimated by RepeatHMM for three patients with SCA10. The estimation of the pathogenic alleles by RepeatHMM for the three subjects A, B and C were 830 (a), 825 (b) and 488 (c), and the estimation by gel electrophoresis were ~840, ~820 and ~530, respectively

Fig. 5

Comparison of the estimation of repeat counts on NA12878 using three sequencing platforms. The sequencing platforms include Illumina short-read sequencing, PacBio long-read sequencing (a), and Nanopore long-read sequencing (b). We examined 40 microsatellites with repeat units in the range of 2–5 bp, which are short enough to be confidently called by the Illumina data

Fig. 6

The analysis of ATXN3 in HX1 using three different sequencing techniques. a Whole-genome long-read sequencing with ~100X coverage. b PCR-based long-read sequencing with three randomly down-sampled datasets, each with ~100X coverage. c Sanger sequencing. All methods concordantly predicted that there were 14 CAG repeats in ATXN3