Accurate mitochondrial DNA sequencing using off-target reads provides a single test to identify pathogenic point mutations

Abstract

Purpose: Mitochondrial disorders are a common cause of inherited metabolic disease and can be due to mutations affecting mitochondrial DNA or nuclear DNA. The current diagnostic approach involves the targeted resequencing of mitochondrial DNA and candidate nuclear genes, usually proceeds step by step, and is time consuming and costly. Recent evidence suggests that variations in mitochondrial DNA sequence can be obtained from whole-exome sequence data, raising the possibility of a comprehensive single diagnostic test to detect pathogenic point mutations.

Methods: We compared the mitochondrial DNA sequence derived from off-target exome reads with conventional mitochondrial DNA Sanger sequencing in 46 subjects.

Results: Mitochondrial DNA sequences can be reliably obtained using three different whole-exome sequence capture kits. Coverage correlates with the relative amount of mitochondrial DNA in the original genomic DNA sample, heteroplasmy levels can be determined using variant and total read depths, and-providing there is a minimum read depth of 20-fold-rare sequencing errors occur at a rate similar to that observed with conventional Sanger sequencing.

Conclusion: This offers the prospect of using whole-exome sequence in a diagnostic setting to screen not only all protein coding nuclear genes but also all mitochondrial DNA genes for pathogenic mutations. Off-target mitochondrial DNA reads can also be used to assess quality control and maternal ancestry, inform on ethnic origin, and allow genetic disease association studies not previously anticipated with existing whole-exome data sets.

Figure 1

Average mitochondrial DNA (mtDNA) base read depth of 36 Illumina Truseq exomes and mtDNA base read depth for individual patient samples sequenced using three different whole-exome capture target kits.

Figure 2

Correlation between mean mitochondrial DNA (mtDNA) per base depth from whole-exome sequence and mtDNA copy number from quantitative polymerase chain reaction (qPCR). qPCR measurements were performed in triplicate.

Figure 3

Binomial 95% confidence intervals for the detection of heteroplasmies at differing whole-exome sequence coverage depths. Intervals are shown for total read coverage depths (n) of 10-, 20-, 30-, and 50-fold, with the total range of variant allele counts (v) from 1 to n − 1, and the resulting estimated heteroplasmies (v/n) of between 0.0 and 1.0 (0–100%). The plots demonstrate a reduction in the size of the confidence intervals at increasing exome read depths, indicating that the higher the total depth, the closer the estimated heteroplasmy level is likely to be to the true heteroplasmy level.