Developmental Validation of a MPS Workflow with a PCR-Based Short Amplicon Whole Mitochondrial Genome Panel

Abstract

For the adoption of massively parallel sequencing (MPS) systems by forensic laboratories, validation studies on specific workflows are needed to support the feasibility of implementation and the reliability of the data they produce. As such, the whole mitochondrial genome sequencing methodology—Precision ID mtDNA Whole Genome Panel, Ion Chef, Ion S5, and Converge—has been subjected to a variety of developmental validation studies. These validation studies were completed in accordance with the Scientific Working Group on DNA Analysis Methods (SWGDAM) validation guidelines and assessed reproducibility, repeatability, accuracy, sensitivity, specificity to human DNA, and ability to analyze challenging (e.g., mixed, degraded, or low quantity) samples. Intra- and inter-run replicates produced an average maximum pairwise difference in variant frequency of 1.2%. Concordance with data generated with traditional Sanger sequencing and an orthogonal MPS platform methodology was used to assess accuracy, and generation of complete and concordant haplotypes at DNA input levels as low as 37.5 pg of nuclear DNA or 187.5 mitochondrial genome copies illustrated the sensitivity of the system. Overall, data presented herein demonstrate that highly accurate and reproducible results were generated for a variety of sample qualities and quantities, supporting the reliability of this specific whole genome mitochondrial DNA MPS system for analysis of forensic biological evidence.

Interpreting NUMTs in forensic genetics: Seeing the forest for the trees

“…The technology has meanwhile matured and provides a comparably cost-effective and less labor-intensive solution compared to Sanger-based techniques, particularly when many samples or full mitogenomes are analyzed [46]. The forensic application of mitogenome MPS, which expanded the targeted portion from the CR to the entire mitogenome, meanwhile increased the abundance of observed and reported NUMTs (e.g., [42,[47][48][49]. This is in part due to the enhanced sensitivity gained with MPS, which allows for the detection of low frequency variants in mtDNA sequence alignments, such as those originating from NUMT reads.…”

“…Yet, each of them has certain properties that help with its diagnosis when assessing threshold-based variant calls. NUMTs are likely to be observed at known NUMT locations [29,30,42,[47][48][49]56], whereas point heteroplasmyparticularly in the codR - [57][58][59][60][61] and stochastic error occur at random locations across the mitogenome [62][63][64]. Furthermore, point heteroplasmy is not observed in a majority of mitogenome haplotypes when applying variant detection thresholds realistic to forensic samples, i.e., a minimum of 5-10% [57][58][59][60][61].…”

“…This is why mtDNA analysis in general, and particularly in tissues of forensic interest (e.g., rootless hair shafts, degraded bone/teeth), does not yield NUMT signals above background noise. However, NUMTs may appear frequently in reference-type samples assayed with mini amplicon enrichment techniques [42,47,49]. One method to address NUMT co-amplification in reference samples and relatively high-quality forensic samples is to control (i.e., reduce) the quantity of genomic DNA prior to PCR [49,70,77].…”

“…This is why quantitation of the mtDNA and the nDNA in a sample to direct the input quantity for mitogenome PCR may be important in mini amplicon laboratory workflows (e.g., [79,80]), or those involving high-quality DNA followed by WGS or hybridization capture. Validation studies have demonstrated that mini amplicon strategies require mtDNA amounts as little as the equivalent of a single cell or even less as template to generate reliable sequences [81], and, as little as 23 copies of mtDNA to generate a full mitogenome sequence [47]. This indicates that smaller amounts of template mtDNA are sufficient for mitogenome sequencing.…”

Graph Algorithms for Mixture Interpretation

“…One major issue is that a mitochondrial DNA haplotype may be comprised of potentially 100 s of amplicons. These amplicons may overlap to varying degrees and may have varied PCR efficiencies [ 17 ]. In contrast, continuous methods based on massively parallel sequencing (MPS) data need to balance purely quantitative data (variant read-counts) against the linkage (and resultant linkage disequilibrium, LD) inherent to the mitochondrial genome (e.g., as per [ 18 ]).…”