Who's who

Crime scenes are a mess of DNA, making it hard to spot the true criminal.
14 February 2018

Interview with 

Georgina Meakin, University College London


Bloody handprint on window pane


Forensic scientists gather and analyse DNA from crime scenes to help the police identify criminals. But it’s rare that a sample will be completely pure - DNA traces from victims, perpetrators and others can all be mixed up together. So how do they separate out the genetic fingerprints of different people? To find out, Kat Arney spoke to Georgina Meakin - a lecturer in crime and forensic science at University College London who is researching this very problem.

Georgina - At a crime scene there's the obvious body fluids that you think about - you think about the blood, maybe like some pool of tissue like brain matter, or maybe semen if it’s a rape case. They're the obvious things that we can detect easily and you may even just be able to see by the naked eye.

But nowadays, because our DNA technology has advanced and has become so sensitive, we can now sample a surface that you can't actually see anything on, but you might expect to find DNA. So a surface that you might expect someone has had contact with. So let’s say it’s a burglary. It looks like someone has broken in through the window. You might sample the window sill, maybe some glass samples, and you might find DNA from the person who’s been in contact with that.

Kat - But presumably, that criminal is not the only person to have ever touched that window. So, how do you analyse the sample? How do you get information out and then figure out who might have been there?

Georgina - That’s a very, very good point. As I mentioned, from a crime scene perspective, you're thinking about where to target. You have to think about who might have touched it. Obviously, you're thinking about the offender touching. But other people may well have had contact with the item. You know, a residential building or a room, people have been living there, they’ve got to be depositing their DNA in a variety of ways.

But we also can transfer DNA indirectly. So for example, you can shake someone’s hand, you can transfer their DNA onto a surface that they’ve never touched. You can perhaps touch a surface someone else has touched and transfer their DNA that way. So, when we sample a crime scene, quite often, you're going to get DNA from more than one person.

So you're quite right. It starts getting complicated and that’s where the kind of research that I do and other scientists around the world are trying to generate the kind of date to help inform the way we decide when you get a mixed profile – DNA from several people – is there a way of looking at that and say, “Okay, this profile is coming from the person who’s touched it most recently” for example.

At the moment, I don’t think we’re there yet with the data. Perhaps that’s not completely widely held view. I'm certainly challenging the status quo on this method with the research that I'm doing.

Kat - Let’s drill a little bit more into how you're analysing this DNA. With my genetic hat on, I'm, “Ohh! It’s all about DNA sequencing” but that’s not necessarily what you're doing. What are the ways in which you would analyse DNA and does it depend on what sort of sample it is and how much you’ve got?

Georgina - All of those are valid points. The way we routinely analyse DNA, we use profiling rather than sequencing, so we’re genotyping.

Kat - This is the genetic fingerprint rather than looking at all the letters of the DNA, right?

Georgina - Exactly, that. So, what we do is we target a number of different areas on the different chromosomes. It used to be that we looked at 10 areas that was routinely done within the UK but now, different jurisdictions in the UK use different numbers. So for example in England and Wales, we use 16 areas plus an area that looks at the sex, so male or female. And we will prefer to these technologies as DNA-17, so we’re looking at 17 areas.

Kat - You’ve got a profile from some DNA from a crime scene and it could be a bit mixed up. There could be several people in there. How do you then start trying to deconvolute that and work out whose DNA might be whose?

Georgina - So there's a number of ways of doing it. Firstly, you might know that someone’s DNA is definitely there. So for example, you’ve taken a sample from a victim, you expect to see their DNA present so you can essentially subtract their DNA and you get the unknown DNA that you then want to compare to someone else. Another way in which we do it is by looking at the respective proportions of the different DNA.

Kat - I always have this image of a forensic scientist from the telly, from the CSI where they're holding up this kind of gel or an x-ray with loads and loads of bands on. They’re like, “Yes! That’s the genetic fingerprint. That’s the killer.” Does it still look like that?

Georgina - No.

Kat - What does it look like now? You spoiled my dreams! What does it look like now?

Georgina - So literally, it looks like a series of peaks. So when we analyse the DNA, we essentially get what looks like a graph so we can look at the heights of the peaks that are present. The higher the peaks, the more DNA you have present. So it might be that all the peaks belonging to one profile are much higher, therefore you have more DNA from the other peaks so you can separate out a profile in that way.

Ultimately, if you can't do either of those things then it gets really difficult and then it kind of becomes out of the remit of a DNA analyst and more into the remit of a statistician. And there are a number of different statistical algorithms, programmes, probabilistic genotyping software that’s now available which it can assess the probability of matches between crime scene profiles and people of interest.

Kat - From your perspective and the research that you're doing, where are we going next? What are you excited about and what do you think will be kind of the next big thing or the big questions that people are really starting to get their teeth into?

Georgina - So in terms of applying DNA analysis to forensic science, we’re very good at getting a DNA profile. We can get DNA profiles in very, very small amounts of DNA, just a few cells. We’re getting better with the statistical analysis and being able to attribute that DNA to someone. So now the real question are less about who is it and how did it get there.

So when we find DNA at a crime scene, we want to know, is it coming from the person involved in the crime or was it already there, or did it get there through innocent means? This is the kind of research that I think is really critical and the forensic science regulator has raised this in her last few annual reports that we really need research to start assessing the transfer and persistence of trace evidence of which DNA is one.

Kat - That is kind of scary actually because we all walk through the world just shedding our DNA – you know, our skin, tissue, hair or saliva – all these things and that it could completely innocently end up being in some forensic examination. It’s quite scary.

Georgina - It is scary and there are documented cases where we’re seen it actually happened where someone has been accused of a crime that they didn’t commit because their DNA has been indirectly transferred to the crime scene. The question is, can we tell from the DNA that we recover from a scene, how it got there, and we really need more experimental data to be able to assess that or at least try and assess the probability of different explanations being the case.

Kat - What are the kind of ways that people are thinking of to approach this problem of how did this DNA get there? Is it just a trace that’s got there by accident?

Georgina - So essentially, what we want to do is when you're assessing that finding of DNA, you’ve got what the prosecution said happened and you’ve got what the defence say happened. And you want to try and see, can you distinguish between those two scenarios?

So we need data to be able to input into a statistical model to assess which version of event is more likely and that’s where the kind of research that universities and some forensic science providers are doing to try and generate that empirical data. So we’re looking at things like when you wear an item of clothing and someone else wears it, what DNA do you expect to find? Are you going to get more DNA from the first wearer or the second wearer?

Currently, research I'm doing is suggesting that actually, it tends to depend on the person because some people deposit more DNA than others. But these are the kind of questions we need to be investigating and generating data for.

Kat - Georgina Meakin from UCL.


Add a comment