Table 1.
Summary of mtDNA mapping and sequencing results for 20 cases.
Figure 1.
Variability in 454 sequence coverage.
Total read coverage (redundant and non-redundant) is plotted at each mtDNA position (1 to 16569) and graphed as a continuous thin gray trace for each case (different shades for each case). The mean coverage for all 20 cases is represented by a thick dark black trace. The shape of the traces shows coverage variability both between cases and along the same mtDNA. The black horizontal lines (A, B & C) above the graph represent the three mtDNA PCR fragments used for 454 sequencing. Greater coverage was noted in the regions in which the PCR fragments overlap compared to coverage in non-overlapping regions.
Table 2.
Discordant and/or heteroplasmic substitution mtDNA variants.
Figure 2.
Comparison of 454 and Sanger sequencing results.
Histograms show the distribution of variants by type: single nucleotide substitutions (subst), repeats and insertion/deletions (indel). A. Sequencing method. Results are separated by detection using 454 (left panel, light gray) or Sanger (right panel, dark gray). B. Concordance. Results are separated by concordant detection using 454 and Sanger sequencing (left panel, dark gray) or by discordant detection, 454 or Sanger sequencing only (right panel, light gray). We observed high concordance between the two methods for mtDNA substitutions and significant discordance for indels.
Figure 3.
Novel, potential mtDNA mutation 7501T>C detected by 454 sequencing.
A. Case 5 pedigree. Circles represent females and squares represent males; solid shapes are affected individuals; an arrow indicates the proband that died from sudden cardiac death (SCD). B. 454 sequencing. Shown are representative 454 data flowgrams. We obtained 298 total unique reads, 54 reads had the 7501T>C variant (top) and 244 reads had the reference sequence (bottom). C. Sanger sequencing. Retrospective analysis of Sanger data revealed the missed variant. As shown in the chromatogram, two peaks are found at position 7501 consistent with a low level of heteroplasmy.
Figure 4.
Coverage simulations and variant detection rate for mtDNA variants.
By subsampling all mapped 454 reads, we simulated nine levels of coverage (2× to 500×) for 419 mtDNA substitution variants. Graph shows the variant detection rate by the log10 (coverage) for homoplasmy (black solid line, n = 413) and heteroplasmy (black dashed line, n = 6). We estimated that the minimum coverages to detect 95% of the variants (gray dashed horizontal line) are >20× for homoplasmy and >200× for heteroplasmy.
Figure 5.
Mitochondrial DNA sequencing approach.
Schematic depicts steps in ABI Sanger and Roche 454 sequencing [23]. We isolated DNA from 20 cases and amplified mitochondrial DNA (mtDNA) by long range PCR (LR-PCR). We sequenced the PCR fragments using standard Sanger and 454 sequencing methods. For 454, the fragments were sheared and then ligated to Multiplex Identifier (MID) adaptors. Next, we made two pooled samples by combining the tagged fragments for cases 1–10 and cases 11–20. Single-stranded DNA (ssDNA) libraries were made from each pooled sample and clonally amplified by emulsion PCR. Parallel sequencing was done on a single picotiter plate divide into two regions. The final data, represented as a Sanger chromatogram and 454 flowgrams, are shown.