Due to the unforeseen popularity of mitochondrial DNA analysis, in 2013 forensic science is bumping up against the few remaining technical challenges in mtDNA analysis and ready to embrace some new tools for dealing with those challenges. Next Generation In 1996, the forensic community could not have envisioned the eventual widespread and successful usage of mitochondrial DNA (mtDNA) as a forensic tool throughout the world. The first recorded court case, Tennessee v. William Ware, heralded in the wider press the ability of forensic scientists to connect the smallest hair shafts and most degraded skeletal remains to individuals and their families for the resolution of criminal cases and missing persons identifications. Indeed, analysts at the Armed Forces DNA Identification Laboratory had, by 1996, quietly been using mtDNA and Sanger sequencing for several years to identify missing war dead from military conflicts as far back as the Civil War.

Due to this unforeseen popularity of mitochondrial DNA analysis, in 2013 forensic science is bumping up against the few remaining technical challenges in mtDNA analysis and ready to embrace some new tools for dealing with those challenges. Next Generation Sequencing (NGS) is one of those new tools, poised to become a big player in forensic testing and equal to the challenges experienced by mtDNA practitioners. Throughput needs for mtDNA testing have increased, due to large numbers of missing persons cases, human rights investigations, and natural disasters. However, the old system of Sanger sequencing is notoriously low-throughput; large projects can easily take months to process. Research has shown that whole mitochondrial genomes—which take even longer to process than the two traditionally analyzed hypervariable regions—can resolve some cases where common mtDNA types cannot distinguish between individuals. Dr. Walther Parson, of the Institute of Legal Medicine in Innsbruck, Austria, says, “Whole genomes not only increase discrimination power, but form the basis for nomenclature of mitochondrial DNA types as well as serve as the basis for quality control checks of our data.” Dr. Mitchell Holland, director of the forensic science program at Penn State University, is focused on the way heteroplasmy, the state where more than a single mtDNA type is present within one person, could inform and enhance mtDNA forensic applications. Heteroplasmy has been found to be ubiquitous in all tissues at low levels, but is difficult to detect and quantitate with Sanger sequencing, which displays nucleotide sequence as pooled strands of amplified product. He comments that “the Next Generation approach will allow us to detect mtDNA heteroplasmy at a much lower level and report the heteroplasmy as a finding.” Finally, mixtures containing biological material from more than one person cannot be deconvoluted with Sanger sequencing. Next Generation Sequencing could open up the range of samples eligible for mtDNA analysis to include forensic stains, because mixture interpretation is possible with this technology.

In October 2013 a working group convened at the International Symposium on Human Identification in Atlanta, Georgia, to discuss progress in Next Generation Sequencing methods for mitochondrial DNA analysis. These academic and government researchers and practicing forensic scientists are moving rapidly toward validation of these methods with the expectation that NGS will replace Sanger sequencing for all but a few forensic mtDNA cases within the next five years. A series of papers and presentations at the workshop delineated the processes, advantages, and challenges that are being encountered in bringing a wholly new methodology online in forensic laboratories. The group routinely shares data about instrumentation, chemistries, and software to expand understanding of the technology within the forensic community.

Simply described, Next Generation Sequencing (sometimes called “Current Generation Sequencing”) reveals the DNA sequences of the thousands of individual molecules inherent to a forensic sample. Instrumentation with a small benchtop footprint, such as the Illumina MiSeq/HiSeq, Life Technologies Ion Torrent PGM, or Roche GS Junior/454 carries out massively parallel sequencing on thousands of DNA sequence clusters that have been generated from DNA libraries prepared from extracted and amplified DNA. Chemical bar-coding of individual samples allows multiple samples to be analyzed within the same run, increasing throughput. The bar-coding also permits re-association of fragments from within a sample that may display mixtures or heteroplasmy and permits linkage associations between fragments, giving a much deeper glimpse into the variation that characterizes an individual sample. The additional variation that is observed above and beyond that revealed by Sanger sequencing illuminates mitochondrial DNA haplotypes at the individual nucleotide level, whether they are from a pristine single source sample, a heteroplasmic mixture innate to the sample, or a sample comprised of DNAs from more than one person. The process can target poor quality degraded DNA as well as high quality abundant DNA by adjusting pre-analysis preparation steps for difficult samples.

The throughput from this system of analysis is astonishing. In an automation compatible, 96 well format, 15 to 600 gigabytes of data are generated per run depending on the instrument, simplifying, accelerating, and streamlining the production of mtDNA sequence data. This approach revolutionizes databasing projects, where good quality samples require only limited handling. While processing of casework samples is more time consuming, the amount and depth of captured data exceeds that generated by Sanger sequencing by many orders of magnitude. Although mitochondrial DNA control region or hypervariable region data analyses are the current primary focus of some laboratories’ research, whole genome data recovery is occurring in other laboratories for both high quality samples or by using whole genome amplification (WGA) on low template samples prior to the NGS process. Whole genomes are a logical extension for mitochondrial DNA testing beyond control or hypervariable regions, now eased by the phenomenal throughput and shortened timeframe allowed by NGS.

For forensic samples where quantity and quality of the DNA template is a concern, parallel research is being carried out to improve DNA yields at the time of sample extraction. Dr. Mark Wilson, director of the forensic science program at Western Carolina University, is actively developing a method for optimizing DNA yield from hair shaft extractions in order to increase the quantity of template that may be input into NGS. Post-extraction quantification methods indicate that >105/2 ul copies of mtDNA are routinely available using this new method, which will eventually be applied to difficult bone samples as well. Dr. Wilson is now routinely applying NGS methods to human hair shafts with or without WGA to target the whole mitochondrial genome. He comments, “our preliminary data support that the improved extraction method, coupled to multiplex amplification around the genome, may be sufficient to generate whole genome mtDNA from hair shaft samples.”

Related work is active in the analysis of difficult skeletal remains as well. Dr. Odile Loreille, a senior research scientist of the Armed Forces DNA Identification Laboratory (AFDIL) on Dover Air Force Base, Delaware, says, “for specific applications at AFDIL, NGS is currently the only way to get sequencing information from DNA that is degraded to less than 100 base pairs, since Sanger sequencing is not easily amenable to such small fragments. Additionally, the ability to immortalize an extract during sample preparation ensures an unlimited amount of extract volume.”

While NGS is driving the need to improve concomitant and complementary methods, NGS challenges themselves remain. Contamination identification, tracking, and control have always been of concern in mtDNA analysis, due to the natural abundance of this molecule. NGS is extremely sensitive, and while this bodes well for the detection and quantification of informative heteroplasmy as well as the ability to deconvolute mixtures, interpretation guidelines will need to be refined to account for sometimes inevitable low-level contamination when performing casework. When working with skeletal remains, Dr. Loreille has observed that microbial sequences can erroneously map to the human mtDNA reference; parameters within current software packages should help to address this challenge, but bioinformatic tools specific to the needs of the forensic community could ease the transition to this new technology. Indeed, the quantity of raw data generated by NGS is sufficiently vast to require tools to manage terabytes of mitochondrial DNA data. Dr. Loreille adds that current software packages are not designed to produce alignments according to mtDNA forensic guidelines in regions of length heteroplasmy (such as the hypervariable regions) or repetition (such as the AC stretch), and custom solutions may be needed. NGS plus software enhancements designed to aid forensic mtDNA analysts will also aid in researching other fascinating features of mitochondrial DNA inheritance, such as transmission of heteroplasmy between generations of maternal relatives and its tissue-specific characteristics, which may inform disease-related studies.

Ultimately the power of mtDNA analysis resides in the frequency estimate of a matching type derived from regional, national, and other population databases containing DNA sequence data from thousands of samples. To date, hypervariable region or control region databases have been the norm, due to the considerable investment of time required to create them. With the ease of generating data using NGS, whole mtDNA genome databasing will become the norm, with 16kb genomes from thousands of individuals around the world comprising open-access databases such as EMPOP (, a database managed by the Institute of Legal Medicine in Innsbruck. Increasing database sizes allows for refinement of haplotype frequency estimates, which in turn strengthens mtDNA as a court tool or missing persons identification method. Dr. Parson states, “getting a mito genome from a sample just within two days is exciting. We now do much more of this work for population genetic studies…it is fascinating.”

Another question remaining is the cost of this form of mtDNA analysis. On the positive side, the ability to run many samples simultaneously means that overall, instrument runs are less expensive, and in particular, databasing costs will drop substantially. However, for forensic casework, when positive and negative controls that are mandated by forensic testing standards are added, along with increased bioinformatics requirements for interpretation, it remains to be seen whether NGS for mtDNA can be cost effective when adopted outside of specialized forensic testing centers. Nevertheless, the enhanced ability that is offered by NGS to look deep within the mtDNA genome in every tissue and diverse forensic samples will increase speedy matches for skeletal remains, improve investigations of crime scene samples and their relationships to victims and suspects, and improve diagnostic and treatment methods for mitochondrial diseases.

Disclaimer: The opinions or assertions presented are private and should not be construed as official or as reflecting the views of the Department of Defense, its branches, the US Army Medical Research or Materiel Command or the Armed Forces Medical Examiner System.

Terry Melton, Ph.D. is Laboratory Director for Mitotyping Technologies, a Division of American International Biotechnology. She has over 20 years experience in the practice of forensic mitochondrial DNA analysis. Mitotyping Technologies, 2565 Park Center Boulevard, Suite 200, State College, PA 16801; PH (814) 861-0676;;