Probabilistic Genotyping in the Forensic Lab

Probabilistic Genotyping in the Forensic Lab

PG software has potential to be “game changer” in resolving criminal cases

Michael Coble, PhD

Since its introduction more than a decade ago, probabilistic genotyping (PG) software has been regarded as a game changing method for interpreting DNA profiling evidence. Relying largely on statistical methods, mathematical algorithms, and probability distributions, PG software has enabled forensic analysts to interpret low-level, degraded, or mixed DNA samples that previously would have been determined as uninterpretable or inconclusive.

This has led to successful use of PG software in more than 220,000 cases worldwide, including production of interpretable and admissible DNA results in a wide range of violent crime, sexual assault, and other criminal cases. (See PG software has also proven to be highly effective in excluding persons wrongly associated as a contributor to crime scene evidence and exonerating individuals wrongly convicted via post-conviction cases. It has helped to crack cold cases in which evidence originally dismissed as inconclusive was reexamined and used to develop new investigative leads. 

But despite its success in resolving complex DNA profiles, many crime labs have been slow to adopt PG software, with lab managers citing everything from concerns about how to communicate the likelihood ratios, to the statistical output of PG software, to the challenges of testifying about it in court. At present, about one-half of the more than 200 forensic labs in the US have some form of PG software at their disposal. Another one-fourth are in the process of purchasing, validating, or implementing a PG system. That means about one-fourth of the US are in a holding pattern for the moment.

Given PG software’s success in resolving complex DNA profiles, why aren’t more labs adopting these methods of interpretation? One obvious reason is the very nature of crime labs, which by and large need to take a slow, meticulous approach to implementing any new technology. This certainly applies to PG software, which requires proper training in the interpretation of DNA evidence produced by PG software; the assignment of meaningful propositions; the principles and practices of the PG software being used; and interpretation of the data generated by the software. The last point is an important one to remember, as PG software is a tool for the analyst to render the final decision, not simply a black box that provides an answer.  

Validating PG software can also appear to be both daunting and cumbersome. To have a better understanding of the data being produced by PG software, lab managers, technical leaders, and analysts must regularly review validation studies that define the limitations of PG software and properly validate their own PG software with in-house studies.

When you couple these concerns with the heavy caseload most crime labs have to handle, it’s easy to see why PG adoption might be slow. Based on the experience of the labs that have already adopted PG software, though, the chances for successful execution can be substantially improved by treating implementation like a project and using a team approach (instead of placing the task on one person). By assigning proper resources, setting milestones, allowing analysts sufficient time to complete validation work, and being realistic about timeframes, labs can and will succeed. This is also beneficial for implementation as the team has had a hand in all aspects of the software from the beginning. 

Implementing PG software in labs

Lab managers grappling with PG adoption should start the process by taking advantage of the forensic labs throughout the country and internationally that already have implemented PG software and are now using it for casework. Reach out to these labs to find out what worked for them, what didn’t, and what lessons they learned along the way. PG providers generally are happy to provide recommendations for adopters of their software which, in turn, are open to providing guidance and best practices for successful implementations. As the old saying goes, “the early bird may get the worm, but it’s the second mouse that gets the cheese.” 

While each PG software package is unique, laboratories generally need to test a series of single source profiles amplified with a range of DNA quantities to model the peak height variability inherent to the lab system. Validation studies will use data from the laboratory DNA profiling kit following their specific protocol, and data generated on the capillary electrophoresis instrument.

Beyond single source samples, crime labs must also run an appropriate number of mixtures to validate their software. Organizations like the Scientific Working Group on DNA Analysis Methods (SWGDAM), the International Society for Forensic Genetics (ISFG), and the UK Forensic Science Regulator, have published guidance documents that labs will find particularly useful in validating their PG software. Several recent papers in the peer-reviewed literature also provide helpful validation suggestions, which labs can benefit from reviewing.

PG software in criminal cases

As analysts become increasingly comfortable switching from their former mixture interpretation methods to PG software, they will likely face another hurdle: concerns surrounding their ability to effectively communicate how PG software works in court proceedings. Many forensic analysts working in the field of DNA testing have never had to prepare for or endure an admissibility hearing or court proceeding, let alone try to explain the intricacies of mixture interpretation, likelihood ratios, or number of contributor assignment in terms that a judge and jury will readily understand.   

While PG software provides substantial benefits over older methodologies, it can only be effective if analysts are able to clearly explain what PG software is, how it works, and why its findings are scientifically reliable in court. Here again, numerous resources are available to help. Transcripts of other court proceedings are readily available online. PG software developers and other forensic experts can provide counsel and training. With adequate preparation, analysts will be able to identify resources related to the methodology and software used in probabilistic genotyping, demonstrate how they have been validated and subjected to peer review and publication, describe how PG software and its underlying scientific principles are founded and generally accepted in relevant communities, and address their state’s legal standard for the admissibility of scientific evidence. 

Bottom line, lab managers or analysts shouldn’t need to feel they are alone in this process. While adapting PG software doesn’t happen overnight and is likely to face a few hiccups along the way, it can happen with advance preparation and proper training. Once it does, forensic labs will find themselves in a substantially better position to provide reliable data from a broader range of DNA evidence and contribute results that can assist both investigators and the judicial system.