Abstract
From 1901 to 1995, forensic investigation essentially constituted examining fingerprints. As recently as the 1990s, when I was first working at the Fingerprint Bureau at New Scotland Yard, the process of matching fingerprints was done largely by hand. By this time, forensic techniques also included marks and traces, so you could tell, for example, if a burglar was covered in a particular type of glass, and you could look for fibre transfer. You could do some examination of body fluids; if you had something the size of tuppence then you would be able to know whether that person was A, B or O, or Rhesus positive or negative, and you could do some presumptive tests. You could test the enzymatic reactions in blood or semen or saliva – processes still used a lot now – to tell where the source of a body fluid had actually originated.
The major breakthrough came when scientists working on Alec Jeffreys' discovery of short tandem repeats discovered that you can genetically ‘photocopy' DNA – in fact, make it copy itself. Once heated, the DNA will denature: the double helix structure will separate out. If we add other building blocks, as you cool it down, the DNA will latch onto its mirror image, effectively doubling it in size. Currently, a full DNA profile can be obtained from around 6 nanograms of original biological material. Roughly, this equates to 1 to the minus 10,000,000,000,000 of the size of a grain of salt.
Following these developments, changes to legislation were clearly needed. The big change, which came in 1995, slipping through both Houses of Parliament with unprecedented speed, was that the mouth is no longer an intimate orifice. Of course, forcing your hand into someone's mouth would technically be an assault, but if it is a police officer doing it with a cotton swab, he is absolutely permitted to do so. If the individual should be wriggling whilst he is doing it and he happens to hold on to their hair to keep them still, he may well pull out a few hairs with roots on them, which is just as helpful from a forensic perspective.
Having established the way in which we would be able to take samples, further changes happened fairly rapidly, meaning that the UK National DNA Database is the biggest such database in the world.
In 1997, further changes to legislation allowed police to take a DNA sample from somebody who was arrested for a recordable offence: essentially anything above parking offences. The Criminal Justice and Police Act 2001 extended this to allow DNA to be retained from people charged with an offence, even if they were subsequently acquitted. For example, you might go to a pub and witness a fight in the corner; perhaps you would be arrested as a result, because nobody was really sure who did what, and you would go to a police station and have a sample taken just because you happened to be there. Until very recently that sample stayed on the database.
An even more controversial change came in 2004, allowing police to take a sample from anybody over the age of 10 for a recordable offence. Even if, technically, being a juvenile, they were never going to be prosecuted as an adult, their profile would go on the adult DNA database.
These contentious rulings were in some way addressed by the Protection of Freedoms Act 2012 – for which we can thank Nick Clegg – which asked the question: if you haven’t been convicted of anything, should your DNA be kept on what is effectively a criminal DNA database? As a result of this Act, a million profiles have been removed, but there are still around six million profiles on there; that equates to about five million people.
Profiling technology: The science bit
Standard profiling: Enhancement techniques
Enhancement techniques are used when the DNA is old or degraded to an extent that a useful profile cannot be obtained. Probably the best known case in which this technique did not go well was the 1992 murder of Rachel Nickell, for which Colin Stagg was originally prosecuted. Rachel Nickell was found on Wimbledon Common and it was apparent that she had been posed. She hadn’t just fallen and been left there, she had actually been manoeuvred into a position, and therefore had exposed areas of skin which – it would seem – the murderer had to have touched.
As a Crime Scene Manager, the first action when approaching a body is to establish a forensic strategy. Where a CSM knows that the body has been deposited in a location other than the attack site, and they think there has been some sort of intimate contact, a major part of their protocol is to take tapings from exposed areas of skin, and indeed the rest of the clothing, ready for genetic ‘photocopying' as described above. In 1992, the optimum number of times a sample of DNA could be heated and cooled until it replicated itself was considered to be around 28. However, because less was known about DNA at the time of Rachel Nickell's murder, and because the sample was not from an identifiable body fluid that a presumptive test could be used on (it could have been just from somebody touching her leg or her arm, for example), it was decided that more PCR, or preliminary chain reaction, was required. The PCR process was carried out over 34 cycles in a technique known as ‘low copy number’: if the copy number of the DNA sample size is small, copy it more. No results were obtained from those initial tests – no DNA was found in the sample. This made no sense: at the very least, there should have been DNA from the victim herself. Clearly something was not working. Unfortunately by the time the Metropolitan Police and the Forensic Science Service had established this, Robert Napper had killed three other people. We have seen huge improvements in technology since then, but it is not going entirely our way.
Mitochondrial DNA
Mitochondria are organelles which act as a digestive system for the cell, taking in nutrients and breaking them down to create energy rich molecules. They also have their own DNA signature. The downside is that mitochondria are only maternally inherited, so in the general population an individual's distinctiveness is not that great. However, its huge advantage is that it has its own cell membrane, as well as lying within the cell membrane of the larger cell, and that there are lots more copies of it than there are of nuclear DNA sitting in the nucleus itself. This makes it far more robust: it is less prone to degradation, less prone to extremes of heat. In plane crashes and train crashes, where there have been substantial amounts of fire, mitochondrial DNA can still be found, quite often in the teeth of the jaw, and all that is needed is a maternal relative in order to compare it against the passenger list.
Y-STR
As everyone knows, women have XX and men have XY chromosomes. This is wonderful news in terms of DNA analysis: new kits can look at 23 sites just on the Y chromosome. In most rape and sexual assault cases most of the DNA samples will come from the (usually female) victim. Her profile can be taken out of the mix so that only the male suspect's DNA remains. By only examining the male elements of the profile, one man's DNA can be separated from that of another.
Animal/plant kits
One of my favourite subjects at the moment is animal/plant kits. Essentially, if you approach a biologist and say, “We are really worried about the importation of iguanas. How can we tell one iguana from another?” they can make a commercial kit to do exactly what we do with the human population: work out how individual that iguana is in a general population of them.
Example 1: a body has been found at a deposition site wrapped in a blanket that is covered in cats’ hairs. The investigators have a rough idea of where the attack site is; all they need to find out is where the cat lives and match the DNA in the hairs on the blanket to that particular cat.
Example 2: someone has crashed their vehicle into a tree in a forest or perhaps driven over some forested areas in order to deposit a body in the woodland. It is possible to tell if a sample is from a single tree, perhaps from a tiny piece of bark attached to the front radiator, rather than just “It’s an oak tree”, to be able to state: ‘It’s an oak tree and it’s
Familial searching
Familial searching is ridiculously expensive and therefore used as a last resort. However, if there is no match on the database from a crime scene sample it is possible to look for closeness: we can look to see if somebody that we think is closely related to that individual crops up somewhere else and then start to work out family trees.
It's all junk
In terms of legislation, and the improvements we already have and where we are going next, the most important thing about the DNA used for crime scene purposes is that it is junk. As far as we know, it doesn’t code for anything. But it does contain Short Tandem Repeats and DNA 2 strands. These are connected together, AT to TA, or A to C and G together, which are the ones that we try to break when we are doing our genetic photocopying. We see these tiny repeated sequences, and they are just nonsense, but the greater number of those repeated series of nonsense that we have, the more sites you look at, the more unique they get. So that, in a given population, by looking at X number of sites, the numbers of repeats found in each of those locations will start to differ and become more and more unique. Looking at more and more sites along the molecule gives an increasingly better indication of who that person is.
The database kicked off with the second generation Multiplex in 1995, looking at 6 sites of junk DNA, followed in 1998 by the upgrade to SGM Plus looking at 10 sites (which included sex determination). The Multiplex Upgrade Program went live in August 2015, allowing us to look at 17 sites – 16 sites of short tandem repeats and sex determination. The extra 6 sites added to those we already had with SGM Plus were chosen because they are smaller sections. The advantage of choosing smaller sections is that they are easier to find when a sample has been exposed to extremes of temperature or if it is degraded, because a smaller chunk will be more robust. As a result, we obtain more usable profiles with ever decreasing amounts of DNA.
However, because the technology is so sensitive in the first place and because we are only looking for these tiny sites, every result we obtain is a mixture. I tried an experiment of buying a packet of cigarettes (filthy habit) from a newsagent, opening it, smoking a cigarette myself, putting it into an ashtray and then testing it in an extraction buffer – the result being a mixture of DNA. I was the only person who had touched the cigarette, the only person who had smoked it, so the contamination came from the something in the cigarette itself, or someone from the factory, or from the ashtray, or, disastrously in legal terms, from the lab.
Single source profile
This is now exacerbated by the fact that we are looking at 17 sites, plus the extra ones which show up because they are only small in the first place. In other words, sites which were once not picked up at all because the DNA was old or degraded, or had been exposed to heat or moisture or nibbled by bacteria: any number of bad things could have happened to it and we would not see it. Now we see the whole mess. Everything we have is a mixture, which means that the amount of time it takes to interpret a profile to give the police a definitive answer that would help them in their investigation has extended. Time is money for the forensic providers and we don’t want to pay them any more for it – but there is no alternative. The DNA Strategy Board, governed by the Home Office, decided “We are going to get something better”, and, on paper, 17 sites are better than ten, but it takes longer, is more complicated and does not necessarily give a better result.
Case study: The red carpet
(For colour images please see online version of this paper)
This case occurred just at the cusp of being able to use DNA for useful purposes, although there were still some issues.
We investigated this particular scene in North East London after the London Ambulance Service received a call from a man saying his wife had fallen down the stairs. On the way there, the paramedics looked at the address and remembered they had been called there many times, usually by the wife saying that her husband had attacked her. They phoned the police and I went round to have a look.
Because the carpet is a powerful shade of red, the tiny bloodstain at the bottom of the stairs cannot be seen.
In the kitchen I saw, on the floor by the washing machine, a blue t-shirt with bloodstains on the back all around the neck. By this time the woman had been taken to hospital, where they had tried to resuscitate her but she had died. The post-mortem had not yet taken place because we were still examining the scene and taking photos, but I did know that she had received a major injury to the back of her head. This was consistent with the blue t-shirt, but we were not yet sure of the cause of death.
The spare room, with futon bed, where the couple's grandchildren would come to stay.
The main bedroom, showing that the valance doesn’t match the set – more relevant than might initially appear.
A tiny piece of hair which, from the dye patterns and the regrowth of the woman's hair, did not seem to equate with this particular incident, but did match up with a statement she had provided a couple of weeks before in which she said that her husband had grabbed her hair, pulled some of it out and assaulted her.
The sheet which matched the valance upstairs, along with a red pair of shorts which house-to-house enquiries suggested the woman had been wearing during the day with the blue t-shirt. It seemed highly unlikely that any man whose wife had just fallen down the stairs would start stripping the bed for the purposes of putting a wash on.
According to the husband, he was watching television that evening when his wife announced she was going to have a bath. Following her bath, she returned downstairs with a towel around her and sat with him for a while before deciding to go to bed. She set off upstairs and he heard a crash and a scream. Rushing into the hall, he found his wife lying at the bottom of the stairs, not breathing.
There are a couple of things wrong with the above story. The living room was right next to the front door and the woman would had to have fallen almost into the room if she had tumbled down the stairs. More importantly, if she had been dressed only in a towel, how did she get so much blood on the back of the blue t-shirt she had been wearing that day?
I thought it would be worthwhile going to extra lengths to see if the husband had, in fact, cleaned up, and to see what else we could find. This occurred around the time when the Met Lab had transferred to be the Forensic Science Service, and I was unable to get somebody to come out and do a presumptive test for blood at that scene for love or money. In the end, I got some people from the Netherlands Forensic Institute to presumptive test the entire house with 3-aminophthalhydrazide – Luminol – but in the right solution so that we did not break the cells down altogether. All I wanted to do (a phrase I use a lot) was exploit the peroxidase-like activity of haemoglobin to reveal the blood at the scene and, crucially, to show that it was from the victim. So I persuaded the team to come over from Holland, and I put them up in a hotel during Wimbledon week – and it was still cheaper than having the FSS out. (The FSS was closed down for good in March 2012. Forensic work is now, as mentioned above, mainly contracted out to private companies.) 
Results of the presumptive testing for haemoglobin on the stairs.
The Luminol testing revealed splashes of fluorescence on every stair all the way down along with large white patterns at the bottom where the victim's head was resting. In addition, we found blood in a similar drip pattern around the edge of the Futon bed. I concluded that the victim had suffered a bleeding injury before she reached the top of the stairs, meaning that the injuries which caused her death could not have been caused by the fall down the stairs.
However, one potential issue with presumptive testing is that it is not species specific. A suspected killer might claim they had previously had a barbeque with lots of guests to feed: “We ran out of room so I took some meat on a tray up to the bathroom sink to wash it, and then as I was coming down obviously it dripped and there was blood everywhere.” You have to be sure that the blood is your victim’s, but I was confident about this.
The argument put forward by the husband's defence team was “Well, you only came round here because you know I beat my wife up all the time, so her blood will be everywhere”.
Cold case: The fertiliser sack
In 1966, Yolande Waddington, who was working as a nanny in a small village called Beenham, in Oxfordshire, was found murdered. Her body was covered by a fertiliser sack.
Blood on the victim's hairband did not match her blood group, so it was possible that the attacker had been injured. Beenham became the scene of the first mass blood screen: every male in a 20-mile radius was asked to give a sample, but the killer was not found. The police's main suspect was ruled out during that mass blood screen; that individual went on to murder two nine-year-old girls in the same village the year after and was caught and convicted. In 2011, he was coming to the end of his tariff, and was even applying for release, so Thames Valley Police thought it would be worth having another look at the case in question.
We re-examined both the fertiliser sack and the hairband to see if we could get something. The image above shows visible blood upon which the Force had again performed the blood grouping test on. Because we can now do far more sensitive tests, such as the test with 6 nanograms, we decided to swab an area where we couldn’t see any blood at all and see if anything came off it. We obtained a profile which matched the man who was applying for his release.
It was clear that something had gone fundamentally wrong with that first grouping screen. Either he had, like Colin Pitchfork – the first person convicted on DNA evidence – got somebody to go and give a sample for him, or something had gone wrong with the blood grouping test. However, it was a good outcome after 25 years.
Case study: The snowy road
During a very snowy Christmas a few years ago, a young woman's body was found after an extensive search. Her name was Joanna Yeates, and her neighbour was initially the prime suspect. The case illustrates the effect of commercialism on the way that forensic investigation is conducted. As mentioned above, there is now no Home Office Lab; the forensics labs are all run by private companies, and their primary function is to make a profit.
The reason this can be an issue is that it is expensive to have an experienced forensic scientist working on a case. A lot of the analysis is automated, and it is very much ‘factory forensics': one person will know one bit of their job. In a lab fairly recently I saw somebody whose job it was to put cigarette butts into an extraction buffer, which would then agitate and spin them – rather like a washing machine – with the result of obtaining DNA at the bottom of the machine. However, this technician had stuffed a substantial piece of cigarette butt into the tube – so substantial that the extraction buffer was hitting the bottom of the cigarette butt, with the outcome that it was not washing through it at all. This person didn't know why they were doing what they were doing. They had a piece of paper that said “put object in tube, agitate, give to the next person”, with the result that items of forensic significance were not being properly analysed.
Joanna Yeates' body was found at the side of a road. This case was similar to the 1992 Rachel Nickell case in that we believed that she had been placed there – the road was not the attack site. As in 1992, we decided to take tapings from exposed areas of skin to see if anything cropped up. However, in a nightmarish re-enactment of the Rachel Nickell case, we found nothing to start with – nothing at all. This time around we knew that this could not be right – we had to have the victim's DNA on there at the very least. So what was going wrong?
The answer was that the extraction chemistry was not appropriate for the circumstances. Imagine a body being deposited at the side of a road in winter. What happens to roads in winter? The council puts salt on them. Salt is a PCR (preliminary chain reaction) inhibitor, which means that if a taping taken from the skin is contaminated with salt, the preliminary chain reaction – the genetic photocopying – is not going to work.
Money and the factory line
The ‘factory-line' approach to forensics came particularly to my attention recently. I received a statement from a forensic provider which said they had examined a pair of knickers full of body fluids, with a negative result. In the circumstances of that case I did not believe that this could be true. The victim’s body fluids had to be on there, if nobody else’s. “How long did you spend looking at it?” was my question. Eventually – and reluctantly – I was told that they had spent 15 minutes looking at it, because that is how long they will allow themselves to look at it in order to be able to guarantee they will make a profit so they can pay their shareholders. Now, a Phadebas Test, which is a presumptive test for saliva, takes 45 minutes. So they could not know if there was any saliva on there, which, in a statement coming back to a police officer, would make him or her think there was nothing there. A negative statement has changed its meaning now. It means that in the time allowed, they haven’t found it. I am slightly inclined, with every single item I send to these forensic services providers, to get it back and test it myself.
So a lot of our successes at the moment are not because we have got these improvements in terms of DNA profiling and chemistry and better profiling methodology; it is because, in cases such as that of Stephen Lawrence, a result is found after spending the money and taking the time to examine an item properly. Yet it is no lab's fault that they are profit-driven or that everybody wants everything tomorrow. 
On this jacket you can see a tiny patch of blood on the right-hand side. This blood had been looked at probably once a year for nearly 20 years before it came in to LGC Forensics, the team I was running at that particular time. It came to us because the Metropolitan Police were prepared to spend enough money on it to give somebody not 15 minutes to look at it, not a day, but as long as it took. Eventually we found that tiny, tiny blood fragment, which is Stephen Lawrence’s blood, on one suspect's collar, and which led eventually to the only two prosecutions that there have been.
This is a direct knock-on from the Rachel Nickell case because the same thing happened to the end of that crook-lock as happened to Rachel Nickell's legs: there was over-amplification of the handle. This is the only miscarriage of justice case I have worked on where we have actually found a profile which has led to a successful conviction.
The only reason this case, the murder of Rachel Manning, was brought to me at all was because it was being investigated by the BBC's Rough Justice programme. (Note: Rough Justice was cancelled by the BBC in November 2007 as a cost-cutting measure.) Barri White and Keith Hyatt had been convicted of the murder and were in prison at the time; they both served six years before being exonerated in 2007.
They were convicted, in terms of forensic evidence, on scientific suggestions that the particulate matter on the victim’s skirt was absolutely unique to their vehicle, so she must have been killed and then propped up on the front seat of this van before she was deposited. Rough Justice showed that, in fact, the van was dirty inside, and that this mixture of particulate could be found in many other vehicles that were in similar need of a clean. As a result of the programme, we were able to reinvestigate the evidence. But the process took a long time: ten years from start to finish. If that programme had not been broadcast, there is a strong chance the two men would have been in prison a lot longer than they were.
Touch DNA
Even now, we believe that a lot of the DNA that we are looking for is not in a cell at all. If it’s not in a cell, we can’t do a presumptive test for it.
Some excellent research work is being done at King’s at the moment. This “touch” residue that we’re leaving around is actually cell-free DNA deposited in our sweat. If we don't know what the source of the body fluid is – if we don’t know where the DNA came from in the first place – then how can we possibly say whether it has evidential value?
The last case I worked on at LGC was on a pair of swimming trunks. A woman had taken her young son to the swimming baths, where she turned around to see a man with his hands on the band of the boy's trunks. She yelled and the man ran off, but he was picked up straightaway. The swimming trunks were submitted for examination to corroborate her version of events, and we found three sets of DNA on them: the boy's, as we would expect to; we found his mother's, because she had dressed him; and we found our suspect’s girlfriend. We didn’t find the suspect's DNA. Even though he had been alone, he had carried his girlfriend's DNA with him to the swimming baths and he had left it on the band of those trunks. If nobody had seen the incident or if the child, being small, had said “Somebody touched my trunks” and we had examined that item and searched on the database, perhaps come up with a name, the investigation could have set off on the wrong road from the start.
The point of how investigations work has changed subtly because you now have to have a defence of where you were because you have to explain how your biological material got there. When I think that a police investigation can use biological material in a conviction, or a potential conviction, unless we can explain ourselves how it got there, I feel some concern.
The future
On the plus side, what is next? Rapid DNA means that we can do preliminary chain reactions in two hours. Reality is finally catching up with television! You can get four profiles from somebody in a couple of hours if you are prepared to pay for it. At the moment the cost is £2000; I think by the end of 2016 it will be the standard rate. It speeds things up enormously.
Next Generation Sequencing is definitely the way forward, because we are currently stuck with looking at short tandem repeats, which means that everything becomes reduced to whatever number you have at that point. Looking at the whole genome has its difficulties, because immediately we have stopped looking at junk. We are now looking at things that will actually show themselves. There will be some legislative problems surrounding that, but at least we have the capability to analyse the whole genome. It used to take years, then it took months, then weeks and now it takes a couple of days, and it will get quicker than that.
I am also particularly fond of Single Nucleotide Polymorphisms, or “SNiPs”, essentially looking at the most basic building block level of a variation. It is incredibly accurate and not that expensive, but we will have to move the whole industry, every single private provider, away from their comfort zone of “Oh, but I like short tandem repeats because I can do those cheaply and I can make a lot of money out of it”. My reply of “Yes, but it’s actually not very good. We want you to do this” inevitably leads to the response: “Well, pay me”, and I am not quite sure where that money is going to come from.
Returning to the problem of finding out how old blood is: I worked on a case where a woman was either pushed, fell or jumped from a balcony. Blood was found everywhere, but I believed it came from an incident a week before. Methylation, which is aging of the cells, is quite a substantial amount of research, but I was delighted to find that Dr Maura Isles, from Rizzoli & Isles, is working on Tryptophan, so this is my next research project. I have great hopes for it. We will be able to tell how old blood is. I think there is hope for the future.
I've talked about Colin Stagg, wrongly accused in the Rachel Nickell case, and I mentioned Barri White and Keith Hyatt, who both served years in prison for a crime they did not commit. These cases of alleged miscarriage of justice are covered by the Criminal Cases Review Commission (CCRC), and two I am interested in at the moment are the cases of Kevin Nunn and Roger Kearney. I am rather bitter about Roger Kearney's case because I appeared for the Defence and he was convicted. Effectively, I believe the issue here is that the Supreme Court ruled that, post-conviction, the rules of disclosure do not apply. Fair enough. But if somebody is prepared to pay for their own forensic investigation of items that they think will actually prove their innocence, because those tests hadn’t been done before, then I think again we need a change in the law to be able to do that. The CCRC is a fantastic organisation, but how are they going to get through all of those cases? There are thousands of them and just one file of DNA results, with all the little EPGs sticking out of it, would take me a week or so to go through. We will never get to the end of being able to tell if there has been an error, as in the Rachel Manning case or the Rachel Nickell case, because it is too complicated and there is too much to investigate. So if we have the opportunity for people who are absolutely convinced of their own innocence, and they are prepared to put their money where their mouth is, then maybe a change in legislation could make that happen a little more quickly.
Discussion
(Applause)