New milestone for comparisons with DNA database: “This is going to lead to breakthroughs”

The Netherlands Forensic Institute (NFI) is able to do more and more with minimal and complex trace evidence, and the technological possibilities in the area of DNA continue to grow. The NFI has now developed new software that makes comparisons with the DNA database even faster and more detailed. “This is a new milestone”, says scientific researcher Jerry Hoogenboom. “The program is going to produce results in even more cases. The new search method also has potential for cold cases in which DNA has been preserved.”

Analyseresultaten (foto ter illustratie)
Analysis results (photograph for illustration purposes)

The new program is called ProbRank. It uses a new way of calculating that is even more accurate than previous programs, in part because it includes more data in its computations. A DNA profile obtained from a forensic trace is represented by a series of peaks, with each peak corresponding to a characteristic of the DNA. ProbRank goes beyond previous methods by including not only the position of the peaks but also the height of the peaks in its calculations. “This will help when analysing and comparing complex trace evidence that is lacking certain characteristics, or trace evidence containing the DNA of multiple individuals”, says Hoogenboom.

Peak heights also included

“If a person has DNA characteristics that are found in the DNA profile of the trace, the program also analyses the heights of the peaks”, explains Hoogenboom. If one person has contributed more DNA than another to a mixed DNA profile, their peaks will differ in height. “In principle, the more DNA, the higher the peaks. In the calculation, ProbRank also considers whether the peaks that belong to the same person correspond in terms of height. This allows the new method to detect similarities even more accurately.”

Currently, it is sometimes not possible to identify a single individual when comparing a complex trace with the DNA database for criminal cases because there are too many possible combinations between the different peaks. This is because the trace consists of DNA from many different people. “In those cases, there is now a better chance of finding a person who actually contributed DNA, because ProbRank also considers the heights and whether they correspond. That is extremely informative.”

Less time, fewer actions

The new method also reduces the time and actions required, says software developer Martin Slagter. He and his colleagues helped Hoogenboom program the new computational model. ProbRank is replacing two programs which the NFI currently uses to perform comparisons and calculations, SmartRank and DNAStatistX.

The NFI has been using SmartRank to search the DNA database since 2017. The software compares all the profiles in the DNA database with characteristics of the DNA profile of a trace found at the crime scene or, for instance, on a piece of clothing. DNAStatistX then calculates the probability of the DNA profile of the trace being obtained if the person in the DNA database had actually contributed to a trace, compared to another random person. “We call that probability the ‘evidentiary value’, and only if it is high enough will the expert name the person in question in the report. ProbRank brings together what we do with the SmartRank and DNAStatistX programs”, explains Slagter.

Apparatuur op het DNA-laboratorium
Equipment in the DNA laboratory

Future

This summer, ProbRank will be incorporated into the Fast ID Line, in which the NFI has automated the entire DNA process from A to Z. The police and the Public Prosecution Service in Rotterdam make use of the Fast ID Line and normally receive an answer as to whether or not the DNA found at the crime scene matches the DNA database after three days. “The Fast ID Line currently only reports the provisional results of the person who has contributed the largest amount to a trace, the primary donor. Thanks to ProbRank, the Fast ID Line will in future also compare the DNA of individuals who have contributed less DNA to the trace with the DNA database and indicate whether a person has been identified, along with the corresponding evidentiary value”, says Slagter. The NFI is currently working on the validation phase.

Friday afternoon

How did the new calculation method come about? Hoogenboom laughs: “In 2021, we made some modifications to DNAStatistX that made the program much faster. This prompted the question – initially light-heartedly – whether it was now fast enough to calculate the evidentiary value for a mixed DNA profile from a trace for all 400,000 individuals in the DNA database... I started doing some coding on a Friday afternoon. It turned out that it worked. That meant I had made an initial ‘proof of concept’, but it did still need to be programmed more robustly. The project was worth taking to the next stage.”

Slagter and his team got involved, and the rest is history. Slagter beams: “The Swedes have already let us know they are keen to use it. That means ProbRank will also be used outside the Netherlands. That makes me incredibly proud.”