Features Technology
Improving DNA typing

Over the past 35 years, the field of forensic genetics and its accompanying analytical tools — often termed the “gold standard” — have become the premier forensic discipline. DNA can be derived from minute biological samples and used to differentiate most individuals in the population and, more importantly, exclude those who have been wrongly associated with forensic biological evidence. DNA typing is an extremely powerful tool for supporting investigations.

As with any analytical tool, DNA typing has its limitations. Poor quality samples, such as those of limited quantity or highly degraded ones, can challenge current capabilities. DNA mixtures (i.e. those samples containing genetic data from two or more individuals and particularly those with three or more contributors) have been quite vexing. It is important to understand those limitations so the results of an analysis can be deemed reliable and the significance of the findings not overstated. The fundamentals of the science and validation studies help to define the limitations so proper protocols can be implemented.

There is also a human component to the generation and interpretation of DNA results. While humans can and do make mistakes, the human interaction is an important part of a quality program and, in some situations, bias can have a positive impact.

As in most fields, technology advances have improved DNA typing. One particular advance has been the introduction of forensic DNA probabilistic genotyping (PG) software, which significantly improves forensic analysts’ abilities to interpret low-level, degraded and/or mixture profiles with greater efficacy and at a much higher degree of accuracy than any interpretational approach used previously.

PG software employs methods that are routinely employed in computational biology, weather prediction, physics, engineering and the stock market. Doing so allows for the reliable use of more of the information contained in a DNA profile — information that used to be discarded. Based on scientific principles of molecular biology, simulated data of millions of profiles and empirical data, such software can assist analysts in deriving sound interpretation on a broader range of DNA profiles. The software can grade profiles on how closely they resemble the observed DNA mixture profile.

Successful conceptual profiles are examined to identify individual DNA profiles that best describe the observed mixture profile. Each of these individual profiles then can be compared to a suspect’s (or victim’s) DNA profile to calculate the strength of the evidence. The probability of the data can be calculated under the proposition that a person of interest is the donor compared to the proposition of an unknown individual is the donor. These results are presented as likelihood ratios, hardly a new manner to present statistical results.

The most important outcome of such software is that previously deemed unresolvable and highly complex DNA mixtures are now providing meaningful and reliable data. To date, sophisticated forensic software has been used in thousands of cases around the world. For example, one such software, STRmix, has been used to interpret DNA evidence in more than 120,000 cases worldwide since its introduction in 2012.

Laboratories using such forensic software are experiencing an increase of interpretable DNA results from guns, other touch evidence and sexual assault evidence. PG tools are proving effective in helping to solve cold cases in which evidence originally dismissed as inconclusive has been reprocessed. PG tools have also been used to support exonerations of the wrongly convicted by re-examining inconclusive results in post-conviction cases.

The prominence and reliance of PG software is demonstrated by the codification of practices to guide its use. The International Society for Forensic Genetics has published guidelines for validating software and the Scientific Working Group on DNA Analysis Methods (SWGDAM) has produced guidelines for validating PG tools. In addition, the Organization of Scientific Area Committees (DNA Analysis 2 Sub-Committee) is working on standards for assessing PG software tools.

Despite those successes and the obvious power of PG software, it is perhaps not unexpected that some have criticized this technology, especially in countries with an adversarial system. Given that, it is incumbent on forensic scientists to:

1. receive proper training on the principles and practices of PG software and the data generated by the software
2. properly validate their PG software, which includes in-house studies and reliance on peer-reviewed literature
3. implement effective protocols to not overstate the strength of PG results
4.  address criticisms that are raised and have merit

For example, lack of peer-review and the potential existence of miscodes have been raised as criticisms of PG software. Peer-review is an important part of science. Science should consistently embrace constructive criticism to improve. Nothing in science is perfect and all that we use today will likely be replaced by better or novel methods. We should not be misdirected with the ultimate goal of perfection as the threshold for reliability. It is unobtainable and not necessary, though, in order to have reliable and valid systems.

As the majority of peer-reviewed papers on PG software are authored by the developers, some argue that these publications may be biased. It is important to recognize that publication of a paper is just the beginning of peer-review. The more effective part of peer-review comes once papers are published. Summaries of findings are presented to the greater scientific community who can read, assess and comment on the papers. If the data and interpretations seem suspect or improvements can be identified, the greater scientific community can comment, critique and perform studies to demonstrate flaws or alternatives to enhance the approach(es). To date, PG software appears to be generally accepted by the overall peer-review process.

Another criticism is that there are miscodes that may impact reliability. The only software that can claim to have no miscodes is the software that has not been written. To be sure, mechanisms should be applied to identify miscodes that can impact the type of data that should be routinely encountered. Some may suggest reading code as an effective method to identify miscodes, but doing so is neither cost-effective nor efficient.

Suppose a software has 100,000 lines of code and to manually scour the code takes about 30 seconds per line. That would equate to 833 hours of effort, or approximately 104 workdays — hardly a trivial exercise. Now consider that most miscodes that are not found during routine testing are those on the fringe of DNA typing results. Being on the fringe means they are not typically encountered and difficult to define. So, a person dedicated to the hypothetical 104 workdays of manual review is unlikely to identify such rare events by scouring code. Proper and comprehensive testing, accompanied by education and training, is a better avenue to identify miscodes and other problems that may arise.

Despite such criticisms, DNA analysis is one of the most effective tools available in modern law enforcement. The use of sophisticated forensic PG software is an advancement that improves on the “gold standard” and undoubtedly will continue to have a profound impact on criminal and civil investigations by providing reliable data from a broader range of DNA evidence and in particular mixtures.

Bruce Budowle, PhD, is a professor at the University of North Texas (UNT) Health Science Center, where he is involved in the research and validation of biotechnology and molecular genetics methodologies. Dr. Budowle has 26 years of experience in forensic science with the Federal Bureau of Investigation.