Predicting Appearance, Ancestry, and Age from DNA

Unknown perpetrators of crime cannot be identified with the current forensic use of DNA. The European Horizon 2020 Project VISAGE (Visual Attributes Through Genomics) aims to overcome this limitation by developing, validating, and implementing a set of molecular genetic tools for predicting appearance, age, and ancestry from unknown trace donors directly from their traces left at crime scenes.


In this interview, Walther Parson, Associate Professor at the Institute of Legal Medicine at the Medical University of Innsbruck, shares what traits can currently be predicted, how these predictions are generated, and what is being done to ensure that these tools are not being misused.


For more on DNA phenotyping, check out our playlist on YouTube.






Laura: Hello and thank you for joining us for the annual video series from the International Symposium on Human Identification. Today, we have Walther Parson with us. How are you doing today Walther?


Walther: I’m fine, how are you?


Laura: I’m doing really well, thank you. Thank you for doing this over video this year. We had our 31st conference, but our first digital version with the pandemic – we wanted to be safe. So, why don’t we kick this off and why don’t you tell us a little bit about yourself?


Walther: Well my name is Walther Parson. I’m a molecular geneticist working in the forensic field for I think it’s now 25 years. I’m located at the University of Innsbruck; the Medical University of Innsbruck. I do teaching and research there primarily. My research interest focus is about all kinds of different markers that are useful for law enforcement purposes. Most, and foremost, mitochondrial DNA, because it’s the best and most beautiful marker, but we also have interest in DNA phenotyping lately, and that’s the topic that I’ve been talking about during the ISHI meeting and massively parallel sequencing in general.


Laura: Alright, yes you presented about VISAGE during the meeting. For some of our viewers who aren’t familiar with that project, could you tell us a little bit about it?


Walther: Of course! VISAGE is a French word for face, and it is the acronym that we found for the European Consortium, which stands for visual attributes for genomics. And by the way, there was a pop band in the eighties with the name Visage, which we found to be a very nice energy. So, when we formed this consortium a few years ago, we found it a nice energy to use this name.


Laura: I think I remember them about that age.


Walther: VISAGE is about DNA phenotyping. We would like to make progress and research in that area. And with phenotyping we mean the prediction about particular traits from the DNA. We are focusing on appearance, ancestry, and age prediction. So, we are targeting the forensic crime scene sample; the unknown sample that doesn’t have a match to the perpetrator or the suspect, or doesn’t have a match to DNA database searches and we would like to provide investigative leads to support the investigations by predicting those traits.


Laura: Ok, now for my understanding, the framework for VISAGE lies on the predictive tools; things like ancestry, appearance, age. What appearance traits can you detect? How does that work?


Walther: We were focusing in the appearance traits on those markers that are currently en vogue, if you will. So those that were investigated in previous studies and where we understand from large studies that they would be beneficial for forensic purposes. In detail, those are the markers that we use to predict eye, hair, and skin color. This combination of markers is also know as the Rasplex Tool that was developed by our colleagues at the Erasmus University in Rotterdam, The Netherlands. The team around Manfred Kayser. But we were adding additional traits to this. So we were adding SNPs that target freckles and the eyebrow color, the hair loss in men, and the hair morphology. And then those markers go together in a single panel together with ancestry markers that we use to get this information from a single sample in a single assay.


Laura: Wow, there’s just so much that you can do with that. What tissues can you use to predict some of those traits?


Walther: For the appearance and ancestry markers, we can target any tissue that has DNA. So, it’s similar to the STRs that we use. As soon as we extract genomic DNA from a tissue, we are also able to type those ancestry and appearance markers. The sensitivity of the assay is very similar to STRs. We are very happy to be able to achieve this goal and get full profiles done to something that equals 15-20 cells and is about 100-150 picograms of DNA.


Laura: Wow, that’s incredible. What instruments have you been using to analyze the DNA and perform the predictions?


Walther: We have been using massively parallel sequencing instruments, and that was quite a change, because many of those markers are typed with different technologies before. For example HIrisplex X was typed with Snapshot, which is different technology that uses capillary. Other markers or other traits are typed with different technology. So, we wanted to make sure we harmonized this to a level that forensic laboratories can benefit from the experience of each other. Forensic labs continue to install MPS instrumentation, and we wanted to make sure that we can use this instrumentation, these platforms, for performing these analyses.


Laura: Are these tools ready for use? When and how can industry professionals get access to them?


Walther: The scope of the project was a research-oriented project with validation and implementation tools. That means… So the European Union was funding us for this particular purpose. That means that we cannot provide final products, but we can provide prototypes that are pretty well validated and that some laboratories can start to implement. We are currently in that part of the process. We are in year 3 of the VISAGE project, and in that timeframe, most of the tools have been completed. They have been tested in a different laboratory. Currently we need to look at a couple of EV tools that come with the development of such tools and development of the foreground. Then, we are more than happy to have contacts to companies and a handshake to companies to get this technology further developed and out to the forensic community.


Laura: That sounds wonderful. So, it’s been interesting. We’ve talked a few different times in the past, so it’s been fun to watch the progression that things have been going over the next few years. What’s next for you? What’s on the horizon until we see you in 2021?


Walther: Well, the tool development is finished in VISAGE, but we still have a couple of major things to do. First of all, this is embedded in an ethical dilemma package, where we need to clearly understand the limitations and the ramifications of what we do here. We don’t want to misuse the tools. We don’t want to overestimate the importance of those tools. We don’t want to overestimate the data. So, the ethical working package is providing a framework in which we are trying to target the right audience. Then there is software development. The consortium partner in the group developed the software that is key in addressing the right person. The software speaks to the expert. It’s not a software that provides a phenotype, it’s a software that provides scientific data. It’s the probability values of the age estimates, and it’s the expert that translates it. And this is how we want to make sure the data are not misinterpreted and are correctly interpreted to the stake holders in the law enforcement chain that further work on these cases. And finally, we are looking at a teaching and education package that is happening in 21, where we want to train those experts with the software to make sure that the tools that we developed now are used in an appropriate way and are interpreted in a way that is useful for our job.


Laura: Absolutely. I think, just watching this industry from outside of it for a couple of years, that’s always the challenge with new tools. Definitely. So, I know you presented at ISHI, but did you have a chance to attend any of the virtual sessions as well?


Walther: I saw some of the sessions, but with the delay of 9 hours that I experience in the time where it happens to Innsbruck, I only saw the morning sessions. But, I enjoyed very much some of those talks.


Laura: Yeah, how did you like it this year. I love to get people’s feedback on it, since it was the first time that we were all virtual.


Walther: From a technical standpoint, it was fantastic. I had doubts if that lively moment would come across that we usually have at meetings, and of course, a virtual meeting will never really be able to replace a real, in-person meeting, but the way how this was structured, and the seamless workflow between the presenters and the chairpersons, it was really very entertaining. A good part about virtual meetings, I think, is that people start asking more questions. At least, I have the feeling that is a positive moment.


Laura: I agree. I think it’s a little easier to type that question then to hold up the group sometimes. Boy, we did miss seeing everyone’s faces in person, so I hope by next year we can see you in Orlando.


Walther: That would be great!


Laura: Thank you for taking time today, we really appreciate it.