Haplotype information of large neuromuscular disease genes provided by linked-read sequencing has a potential to increase diagnostic yield

Abstract

Rare or novel missense variants in large genes such as TTN and NEB are frequent in the general population, which hampers the interpretation of putative disease-causing biallelic variants in patients with sporadic neuromuscular disorders. Often, when the first initial genetic analysis is performed, the reconstructed haplotype, i.e. phasing information of the variants is missing. Segregation analysis increases the diagnostic turnaround time and is not always possible if samples from family members are lacking. To overcome this difficulty, we investigated how well the linked-read technology succeeded to phase variants in these large genes, and whether it improved the identification of structural variants. Linked-read sequencing data of nemaline myopathy, distal myopathy, and proximal myopathy patients were analyzed for phasing, single nucleotide variants, and structural variants. Variant phasing was successful in the large muscle genes studied. The longest continuous phase blocks were gained using high-quality DNA samples with long DNA fragments. Homozygosity increased the number of phase blocks, especially in exome sequencing samples lacking intronic variation. In our cohort, linked-read sequencing added more information about the structural variation but did not lead to a molecular genetic diagnosis. The linked-read technology can support the clinical diagnosis of neuromuscular and other genetic disorders.

Figure 1

Variant phasing in large skeletal muscle genes. The figure illustrates that variant phasing is successful using the linked-read sequencing technique. On average, 97.5% of the exonic and splicing variants were phased in the ten largest skeletal muscle genes.

Figure 2

Fragmented or low-quality DNA, a high degree of homozygosity, and a selected sequencing method can increase a phase block number. (a) The plot shows the average state of homozygosity and the average number of phase blocks in the ten largest skeletal muscle genes for each sample. The size of the dot indicates the percent of long (> 100 kb) DNA molecules in the sample based on LongRanger calculation and the color of the dot expresses the used sequencing method. Our data show that fragmented or low-quality DNA or other sample processing related factors increased phase block number more than a high degree of homozygosity. (b) Despite long DNA fragments, NEB was divided into 13 phase blocks in two samples. Homozygosity can increase the number of phase blocks, especially in whole exome sequenced samples where intronic variation is lacking. Overall, whole genome sequenced (WGS) samples have fewer phase blocks than whole exome sequenced (WES) samples.

Figure 3

The linked-read IGV visualization shows that (a) intronic TPM3 variants of patient 1 are biallelic, and that (b) two ACTA1 variants of patient 2 are in the same allele. The haplotype view of Loupe software (10 × Genomics) also confirms that (c) two TPM3 variants are biallelic, and that (d) both identified ACTA1 variants are in the same allele.

Figure 4

The flow chart of the analyses conducted in the study. Both linked read whole exome sequencing (WES) and whole genome sequencing (WGS) were used. The study was targeted at neuromuscular disorder (NMD genes). Single nucleotide variants (SNVs) and structural variants (SVs) were studied.