Next-generation sequencing of human mitochondrial reference genomes uncovers high heteroplasmy frequency

Abstract

We describe methods for rapid sequencing of the entire human mitochondrial genome (mtgenome), which involve long-range PCR for specific amplification of the mtgenome, pyrosequencing, quantitative mapping of sequence reads to identify sequence variants and heteroplasmy, as well as de novo sequence assembly. These methods have been used to study 40 publicly available HapMap samples of European (CEU) and African (YRI) ancestry to demonstrate a sequencing error rate <5.63×10(-4), nucleotide diversity of 1.6×10(-3) for CEU and 3.7×10(-3) for YRI, patterns of sequence variation consistent with earlier studies, but a higher rate of heteroplasmy varying between 10% and 50%. These results demonstrate that next-generation sequencing technologies allow interrogation of the mitochondrial genome in greater depth than previously possible which may be of value in biology and medicine.

Figure 1. Coverage across the mitochondrial genome.

The top portion of the figure shows where the three amplicons lie and overlap across the mtgenome. A) Coverage for all samples per population. For each sample the coverage at a particular position was normalized by dividing the total number of reads obtained for that sample by 1,000. B) The forward to reverse read ratio for the modal base was centered to 0 using the following statistic: [(forward/reverse)−1]/[(forward/reverse)+1]. C) GC content across the mtgenome was calculated using a sliding window of 51 bp centered on the position in question.

Figure 2. A. Frequency of polymorphic sites by population across the mitochondrial genome.

The top portion of the figure shows the physical locations of the amplicons with overlap across the mtgenome; red and blue dots represent YRI and CEU samples, respectively. Frequency of known polymorphic sites for all samples by population showed on top, and frequency of novel polymorphic sites by population on bottom. The term ‘known’ refers to sites that are listed in mtDNA databases and the term ‘novel’ refers to sites that have not been previously described. B. Frequency of Polymorphic sites per mitochondrial genomic region. Red and blue bars represent YRI and CEU, respectively. Frequency was calculated by dividing the number of polymorphic sites found in each region by the length of that region and then dividing again by the sample size. Each site in a region was counted only once.

Figure 3. A. Heteroplasmic sites per sample.

Total number of heteroplasmic sites found per sample for each YRI and CEU sample. B. Level of heteroplasmy per position across the mitochondrial genome. Top and bottom displays the level of heteroplasmy for all sites found in YRI and CEU samples, respectively, across the mitochondrial genome.