11
The Challenge of DNA
Saturday-morning cinema at the Gaumont, Tally Ho Corner, North Finchley, was always a moment of light relief in the school week. Not so much for the films – Champion the Wonder Horse, The Lone Ranger (‘Hi ho, Silver!’) and Tom and Jerry cartoons – but for all the other elusive attractions, such as girls. It was also a period when cinemas had projection rooms at the back, of the kind featured in Cinema Paradiso, when reels of film might slip a little or need an inconvenient break for a change of reel, or inexplicably show a sequence of a black background with white stars and dots flickering past. It was at moments like these that I would become absorbed by the beam of light cast by the projector through the darkened auditorium towards the screen. (This might well explain why I remained celibate for so long.) The beam contained, or so it seemed, millions of tiny specks or motes floating in frenzied random motion, none of them visible once the lights came up, or in ordinary daylight. So crowded was the beam that I wondered how the picture could penetrate such density. Then I wondered how much of the area surrounding me was saturated with this space debris, which presumably got sucked in by my nose and mouth. Since the film did get shown, and no one died from speck suffocation, I relegated my thoughts to the zone in my brain reserved for trivial pursuits.
Over the last four or five years these thoughts have been resurrected and, I hope, put to good use. I have been encouraged to join a circuit of speakers in performances called An Audience with. . ., which have included such luminaries as Tony Benn and the late John Mortimer, and which enable me to address audiences I would not normally reach, in a variety of places and venues throughout the country – sometimes theatres, at other times cinemas. I do it Dave Allen-style on a stool, with minimal props – part talk, part questions – and my hope is that I can give the law a human face and demystify some of its quaint habits. The audiences are incredibly well informed and always ask highly pertinent questions about up-to-the-minute topics. DNA is one of them: ‘Do Not Ask,’ I respond! They want to know about the process, its reliability and whether it’s safe for all of us to contribute to a national databank.
I assume that they, like me – and like jurors, for that matter – are pretty down-to-earth and far from stupid, but in need of visualising what lies at the heart of it all. In comes the projector beam and all the unseen particles. I ask each member of the audience to start by considering the seat on which they sit. Once they leave, it is highly likely that a part of them remains on that seat. Beside traces from the fibres of their clothes, particles deposited by their bodies may end up on the surface of the seat, as well as on the surrounding surfaces, including the floor. Those particles could emanate from blood, sweat or tears; from combing or stroking hair; from scratching or rubbing the skin; from licking, spitting or kissing; from nose-blowing or sneezing. The list is endless – and there is always some bright spark in the audience who can think of an activity that I haven’t mentioned. Once a chain of transference is envisaged, this can extend to people sitting nearby, who in turn may inadvertently carry your trace to a completely different place; or you may do so yourself, or the trace may become airborne. The longer the chain, from primary contact through secondary onwards, the smaller the amount being transferred. Therefore the presence of your DNA at a particular scene may have significance, and equally it may not.
Suppose a crime is committed in the vicinity of your seat shortly after you’ve gone. Your DNA – either alone or mixed with another person’s – may be recovered. The other DNA trace could have come from the perpetrator of the crime, or it could have come from someone who’d sat in the seat before you, or from someone who had carried the trace from somewhere else. Scientists sometimes refer to a prior deposit as ‘incidental’.
A further complication arises once the crime scene is investigated, and while much of this may seem obvious, all too often the ramifications have not been thought through carefully enough. The individuals who examine the scene bring with them a potential array of DNA (both theirs and other people’s), which they have picked up accidentally. It comes with their own bodies, their clothes, the tools of their trade – the equipment they use for detecting, collecting and packaging samples. Is the scene-of-crime kit they are using DNA-free? Has there been a control test in order to assess this? This may seem unnecessarily laborious, but when we come to consider the size of the particles we’re dealing with, and the sensitivity of the tests now being employed, its relevance will become apparent.
In one of the early cases in which DNA played a major part, I sought to demonstrate the practical difficulties facing any crime-scene examiner.
Question to witness: What do you wear?
Answer: Overalls.
Question: Are they new or laundered? Where have they been stored? Where did you put them on? What clothes were you wearing underneath?
The same questions were repeated in relation to gloves and shoe covers, and the interesting bit came at the end of the witness’s examination of the scene. I asked where he had taken off his protective clothing. He replied that he had done so outside the premises, after the exhibits had been packaged, and that he took everything off, leaving the gloves till last. However, he accepted that the gloves may have touched surfaces contaminated with the DNA, and that in doing so the outside of the gloves may have retained DNA particles. The witness went on to point out that he had thrown the gloves away.
But, I asked, how did you take the gloves off?
And as he demonstrated in the witness box exactly what he did, he suddenly realised that he had used a bare hand to remove the second glove. He knew very well what was coming, because the next question was: What did you touch with that hand before you washed it (if he did), before leaving the scene? The risks are that DNA particles on his hands might then be transferred to other surfaces, including central exhibits in the case (even though they may have been packaged), because a particle on the outside may end up on the inside.
Since that first case, protocols and procedures have been tightened considerably, but creating any contamination-free zone is a huge mountain to climb, and the best you can do – short of eliminating DNA evidence altogether – is to ensure that contamination is reduced to a minimum.
So far I have not even got beyond the scene of the crime, because obviously any collected sample has to be transported and stored at the laboratory. Then it has to be unpacked and examined, and the DNA extracted, amplified and analysed. At each stage the risk of contamination has to be contemplated; in other words, there is the danger of a spurious or a false-positive finding.
The magnitude of the challenge facing forensic science is compounded by the pressure to derive a DNA match from ever-decreasing trace amounts recovered from the scene, which requires employing ultra-sensitive techniques. To understand how sensitive these methods have become, it helps to know what it is these scientists are looking for, once the swab with a sample on it arrives in the laboratory.
The human cell, most of which is water, is the basic unit of life. At the centre of the cell is a nucleus containing forty-six chromosomes (in twenty-three pairs), which carry hereditary information, or the genetic code. The chromosomes themselves are made up primarily of proteins and nucleic acids, specifically DNA (deoxyribonucleic acid), which is structured like a twisted ladder or helix, with something in the region of three billion rungs. You might imagine that a cell housing a mind-bogglingly big ladder could be seen by the naked eye or even spotted in a projector beam, let alone by some hugely perceptive microscope, but the fact is that we are dealing with dimensions at the other end of the universe or infinity spectrum, and we have moved from the inconceivably grand scale to the incomprehensibly petite. Since 1999 1ng of DNA – one billionth of a gram, equivalent to somewhere between 150 and 160 cells of DNA – has been the standard starting template for the purposes of obtaining a profile. To get a handle on this, imagine you are looking at a minute speck of blood, yet what is currently being contemplated in the laboratory is a process that recovers a DNA profile from an even smaller quantity, namely less than 200 picograms: a picogram is one million millionth of a gram.1
If you’re still reading this, just try to imagine what it’s like trying to write it – and it doesn’t get any easier with what comes next, because scientists (like lawyers) love appearing inscrutable by using in-house vocabulary, in this instance acronyms. If I said, ‘MM is an OAP who would need GM before running faster than a BMW’, you would probably be able to translate most of it, but in case you ever run across any of this DNA stuff again, I’ve inserted the most commonly used acronyms alongside the explanations.
Having established the amount, what is the comparison you are trying to make? The majority of the human genetic code is the same in all individuals, which explains why we all have a head, arms and legs, and so forth. There is less than 1 per cent of each individual’s DNA that is different, and interestingly it’s the stretches of DNA on the ladder for which there is no known purpose that provided the key to understanding this. In 1984 British geneticist Dr Alec Jeffreys (now a Professor and a Sir) identified locations or loci on the ladder where there were hyper-variable repeat patterns of DNA. Currently in the UK the Forensic Science Service (FSS) uses ten loci, where the variability is described as Short Tandem Repeats (STR). This was at one time misleadingly compared to a fingerprint; rather, it is a profile, and by no means the whole of it.
From the beginning it has been recognised that the whole business would be easier if a bigger sample could be obtained, and this was achieved by the introduction of a method capable of copying the original sample millions of times, which then enables the scientist to run tests on the copies: it uses a substance called polymerase, which causes a chain reaction (PCR) that reproduces the DNA. The amplification process (called a cycle) is undertaken twenty-eight times using a pre-prepared kit (SGM Plus). In the drive for greater detection, the number of amplification steps has been increased from twenty-eight to thirty-four, and this is where we encounter amounts smaller than 200 picograms. At these levels, described as ‘touch DNA’ or low-template/low-copy number (LT/LCN DNA) analysis, the sensitivity is such that the test may reveal DNA not related to the alleged incident – in other words, the DNA already referred to as either ‘incidental prior deposit’ or deposit due to contamination. Alleles (the name of a variant of a gene) may ‘drop out’ and by chance not get amplified; alternatively, one may ‘drop in’ (stochastic effect – statistically random).
Sounds like some sort of glorious genetic tea party, doesn’t it? Therefore, if you’re going to be sure who’s at the party you have to keep going back in the room, as it were (or rerunning the test), and maintain a high degree of caution when interpreting what you see.
It may be that the test itself is at fault if it is unable to achieve reliable reproducibility, which is precisely the issue I have raised, under the heading of inter-verifiability. Unless the test is validated and accepted by the scientific community, its admissibility in court is questionable.
I represent one of the families who were victims of the Omagh bombing in August 1998, and it was the judgment in the trial of Sean Hoey, charged with involvement in that bombing, that addressed this problem in a very public way. In December 2007 Mr Justice Weir delivered a scathing critique of the evidence, which resulted in the acquittal of the defendant,2 and part of that critique revolved around the alleged finding of the defendant’s DNA on a number of devices. The judge was particularly concerned at the wide variance in expert opinions, not only between the prosecution and the defence, but also between the two experts called by the prosecution. The defence case was that the LCN system had not been validated by the international scientific community. For example, it has only been adopted for evidential purposes in two other countries in the world, the Netherlands and New Zealand, whereas in the United States it is used only for intelligence purposes. During the trial it was revealed that the amplification process had produced a DNA match not only with the defendant, but also with a fifteen-year-old boy from Sussex who had never been to Northern Ireland.3
Much play was made by the FSS that this was a technique that had in fact been used in the Peter Falconio murder trial in Australia, where the defendant’s ‘touch DNA’ had been recovered from the gearstick of the van in which Falconio and his girlfriend Joanne Lees had been travelling. Bradley Murdoch was a mechanic who approached the couple’s van in the outback and, based largely on the DNA evidence, was convicted in 2006 of shooting Falconio and assaulting Lees.
However, in the light of Mr Justice Weir’s observations on Omagh, a full review was carried out by independent scientists headed by Professor Brian Caddy, whose work in relation to the Birmingham Six I describe in Chapter 17. His report in the spring of 20084 accepted the thrust of the judge’s points about validation, especially the need to assess the ability of a procedure to obtain reliable results, and outlined twenty-one proposed recommendations, foremost among which is the need for a ‘national education programme setting out the advantages and limitations of Low Template DNA in order to establish a conformity of approach to crime scene work’.
One of the particular problems in the Omagh case was that no one in 1998 appears to have thought about the possibility that items being recovered, transported and stored from the scene might be susceptible to DNA examination and analysis and the hazards of contamination. The older the incident, the more likely that this will be the case.
That was one of a number of grounds of appeal I put forward on behalf of James Hanratty’s family at a posthumous appeal in 2002. Hanratty had been hanged in 1962 for the notorious A6 murder of Michael Gregsten and the rape and shooting of Valerie Storie in a Bedfordshire lay-by, and a small sample of her underwear had been retained by the police. Hanratty’s body was exhumed and a match of his DNA was made with DNA on the sample. A similar result was achieved in relation to a handkerchief found with the murder weapon at the back of a double-decker bus. We argued, unsuccessfully in this instance, that continuity and integrity had been compromised throughout the significant lapse of time the sample had been stored. Besides the risk of contamination, a sample of DNA may degrade, depending upon the atmospheric conditions under which it is kept. Prior to the development of DNA techniques in the mid1980s, exhibits officers and others would have been unaware of the potential for cross-contamination – for example, with regard to the storage of Hanratty’s clothes when they were taken to court for the trial in 1961.
Despite the dismissal of the appeal, doubts remain and were best expressed by investigative journalist Paul Foot5 before his untimely death. He had researched and campaigned in relation to this case over many years, and in the late 1960s had carefully interviewed over fourteen witnesses who in various ways were able to corroborate the alibi evidence that James Hanratty gave for the first time at his trial. It included fine detail with regard to the Ingledene boarding house in Rhyl where Hanratty had stayed and the evidence of Margaret Walker, a landlady in a neighbouring guest house, who was certain of the date when a young man matching Hanratty’s description came to her house looking for lodgings, the night of the A6 murder. This evidence was not seriously undermined in the Court of Appeal. Paul Foot’s view was that if there was DNA to show that a man staying in Rhyl committed a murder 200 miles away, then there was something seriously wrong with the DNA.
DNA works both ways. It may exonerate or exclude someone just as easily as it may assist in the proof of guilt.
In 1992, after sixteen years in prison, Stefan Kiszko was released after it was demonstrated that semen on the dead victim could not have been his. In 2007 Ronald Castree was convicted on DNA evidence. Also in 1992 the Cardiff Three had their convictions for murder quashed, and in 2003 Jeffrey Gafoor was convicted as a result of DNA analysis. Colin Stagg was originally charged and put on trial for the notorious murder of Rachel Nickell on Wimbledon Common in July 1992, but the trial judge stopped the case in 1994. Fourteen years later DNA evidence led to Robert Napper, Rachel’s real killer.
The implications of these results go far beyond just convicting the guilty; they expose the real risks of convicting the innocent, who may have confessed falsely, been vulnerable or mentally unstable, or been erroneously targeted by over-zealous investigative methods. These are serious fault lines in our system of criminal justice and no area is sacrosanct.
A striking example of this relates to the role of circumstantial evidence – that is to say, indirect evidence inculpating a suspect in the commission of crime. More often than not there is no direct evidence, no eyewitnesses, no admissions or confessions, no contemporaneous audio-visual recordings. Instead what is relied on is a combination of circumstances, a spectrum which may range from forensic science at one end through to opportunity and motive at the other.
The traditional approach to such material has tended to be for the defence to underplay its significance and for the prosecution to overplay it. All too often, the trial judge has been prone to remind juries that, whilst care is required, circumstantial evidence can provide the most compelling form of proof. Analogies are drawn with links in the chain of guilt, but more commonly with strands in a length of rope. Even if one strand falls away, the rope does not necessarily break; much depends on where in the spectrum the strand lies.
No jury should convict unless it can be sure – this is known as the ‘standard of proof’, sometimes described by the words ‘beyond reasonable doubt’. Inevitably these are highly subjective terms of art, not science. It is absolutely essential for a jury to appreciate the meaning and significance of these seemingly simple concepts, which are not always easy to apply. In addressing a jury about the critical nature of this test, I like to employ phrases such as ‘driven to the conclusion’ or ‘as certain as is humanly possible’. It is not the mathematical certainty of the slide rule, but it is the certainty that each of us expects whenever a life-changing decision in human affairs is undertaken. Nothing less is acceptable where the liberty of the subject is at stake.
What this entails, so far as circumstantial evidence is concerned, is that the inference of guilt must be unequivocal and inexorable. The circumstances must point in one direction only. If there’s more than one explanation, or more than one possibility – such that the defendant might be guilty, but on the other hand might not – the necessary standard of proof has not been reached.
The dangers that lie just below the surface of this extremely delicate exercise were graphically illustrated by the intervention of DNA in the case of Michael Shirley – a true victim of circumstance.
In 1986 Michael was an eighteen-year-old able seaman in the Royal Navy, whose ship HMS Apollo had docked at Portsmouth. While on shore-leave he went to Joanna’s nightclub, where he met a woman called Deena Fogg. He left the club with her in a taxi between midnight and 12.30 a.m. on the night of 8/9 December. They parted company after arriving near Deena’s home address.
Not far away, between 12.30 and 1 a.m. on the same night in December, another young woman, Linda Cooke, was raped and killed by a man stamping on her head and neck, on waste ground at Merry Row. Michael was charged and convicted of her murder at Winchester Crown Court on 28 January 1988. The basis of the case against him rested on four circumstances. The first was semen found on intimate swabs taken from the victim, which came from the same blood group as Michael (a group he shared with almost 25 per cent of the British adult male population); the second related to injuries on Michael’s face, arms and back, which could have been inflicted by the victim; the third was a right shoe impression with the word ‘Flash’ in the heel, which was discernible on the abdomen of the victim (Michael possessed a pair of shoes bearing this word in the heel, which he admitted he could have been wearing that night); the fourth was the fact that he was in the general area at about the time of the attack.
Whilst any one of these strands alone would have been insufficient, their combination provided the jury with a rope upon which they hung a verdict of guilty. Therein lies the danger: the infusion or cross-transference of the stronger strands into the weaker. In a nutshell, it’s like adding two and two and making five. My argument has always been that if an individual strand has little or no probative value, then it cannot and should not be enhanced by adding to it one that has.
Michael remained in prison for sixteen years, maintaining his innocence, forgoing parole and serving an extra eighteen months beyond his tariff or earliest day of release; he went on hunger strike five times (one lasting forty-two days, which nearly ended in an irreversible coma). He told the Daily Mirror after his release in July 2003, ‘I’d rather have died in jail than admit a murder I didn’t do.’6
Earlier that year I came on board for the first time to represent Michael at a renewed appeal. His first application had been turned down in 1989. Over the intervening years both the Criminal Cases Review Commission and my instructing solicitors, Nelsons, had pursued further lines of enquiry. One of them followed an insistent request from Michael for a DNA test to be carried out, something that had not been available at the time of his arrest and trial. (This was hardly the reaction of a guilty conscience.) Enough material had been retained for this to be done, and the result opened the prison gates.
The profile of the crime-scene DNA obtained from the victim’s swabs was ‘mixed’ (as might arise in the theatre-seat example I’ve already mentioned). The profile in this case may be visualised as akin to a barcode on a supermarket product – a series of parallel straight lines. These are called ‘bands’. Some of the bands were attributable only to the victim; some were common to both the victim’s DNA profile and Michael’s. There was, however, an array of ‘foreign’ DNA bands which could not have come from either the victim or Michael, save one band consistent only with Michael. This single band, however, occurs in the profile of approximately one in three unrelated individuals among the population at large.
When this was discovered, it put an entirely different complexion on the prosecution case. It had never been suggested that there were two male attackers, and there was no evidence of any prior intercourse between the victim and anyone else that night. There was only speculation. Therefore the killer could well have been the person who deposited the array of ‘foreign’ bands. No reasonable jury, properly directed by the judge, could safely have convicted Michael if it had known all this. It cast each of the four original circumstances in a far less sinister light.
The blood grouping was large and inconclusive; the defendant’s injuries could not be forensically linked to the victim, let alone accurately dated; the shoe mark was similar to any one of 1,058 pairs sold by Marks Shoes in 1986, or to 1,721 shoes manufactured by Melkrose. It had been estimated, conservatively, that fifty-one pairs had been sold in Portsmouth alone. By the time of the appeal, further evidence had been uncovered about timings and movements, especially the taxi log. This enabled us to restrict the window of opportunity in such a way that there was no missing time within which the murder could have been committed by Michael.
All well and good, but his young life (like that of Linda Cooke) had been shattered, albeit (unlike Linda’s) not ended. His mother barely recognised the serious man on his release as the same carefree son who had been imprisoned sixteen years before.
Michael summed up his experience: ‘It’s like crying in the dark when there’s nobody there to hear you. You just sit there knowing you are innocent, asking why people don’t believe you . . . I would love to meet the jury now and very gently ask them what convinced them to convict me and whether they would do so now.’7
This all bears an uncanny resemblance to the case of Sean Hodgson, who served twenty-seven years before DNA finally supported his long-standing refutation of confessions he made at the time, to the rape and murder of a part-time barmaid in Southampton.8
Having now examined all the DNA stages from deposit through to final analysis, there is one more hurdle to clear.
Assuming that a reliable match is achieved on analysis, and given the enhanced specificity of the test, there is a further question relating to the chance that it may have come from someone else, unrelated to the matched person. This probability used to be expressed in thousands, but is now expressed in millions. Herein lies the pitfall – not so common now – which became known as ‘The Prosecutor’s Fallacy’. Having brushed up against logic and semantics as part of my philosophy degree, this was a little easier for me to fathom than the minutiae of the science, although not without considerable assistance from expert statisticians.
Suppose there’s a murder and my DNA is found on the murder weapon, and suppose that this match is calculated to have a probability of one in a million out of a population of fifty million. The probability that my DNA would match that of the murderer is one in a million, and therefore the probability that I didn’t do the murder is one in a million. Or is it? The probability of it matching me is one in a million if I’m a randomly selected person. But this also means that there are another forty-nine people out of the population of fifty million who also match the DNA. So the probability that I did the murder is actually only 2 per cent. Everything depends on what is called ‘conditional probability’, encapsulated by the phrase GIVEN THAT.
So question A, GIVEN THAT I am innocent, what is the probability that my DNA will match the DNA profile from the crime sample? is not the same as question B, GIVEN THAT my DNA matches the DNA profile from the crime sample, what is the probability that I am innocent? It is an easy elision to make, whereby the answer to the first question is given as the answer to the second.9 This was canvassed in an appeal I did in 1994 before the Lord Chief Justice Lord Taylor, where there had been an allegation of sexual offences against women students in Manchester, and a retrial was ordered.
The elevation of DNA science to the status of some kind of Magic Bullet is understandable. In one of its first applications in 1986 it helped to exculpate a young man who had falsely confessed to the rape and murder of Dawn Ashworth. There have been many similar cases since. The danger – as ever – is the assumption of infallibility, and recently this has had the effect of reviving both the arguments in favour of a national databank containing the DNA of every UK resident and the arguments in favour of the return of capital punishment. This is why I have felt it necessary to trespass upon the detail in this sphere, lest these arguments should prevail in the absence of the shortcomings I’ve attempted to identify. At least one member of the higher judiciary, Lord Justice Sedley, has already publicly supported the databank proposal. On the other hand, Professor Sir Alec Jeffreys has opposed the retention of thousands of innocent people’s DNA, saying it raises ‘significant ethical and social issues’.
English and Welsh police already have wider sampling powers than the police of any other country, and per capita we have the world’s largest database, holding the DNA profiles of 4.5 million people, including some 24,000 samples from young persons aged between ten and seventeen who were arrested but never convicted, plus a disproportionate number of Afro-Caribbeans: 40 per cent of black men in England and Wales have their DNA stored on the database, compared with 9 per cent of white men, and there are real concerns that it could be open to abuse. The President of the National Black Police Association, Commander Ali Dizaei, has declared: ‘Black men are disproportionately targeted right across the criminal-justice system where there is no evidence whatsoever that they disproportionately commit crime. We see the current data as a classic example of institutional racism.’10
The Criminal Justice Act 2003 provided for the indefinite retention of fingerprints and biological samples from all persons arrested for recordable offences in England and Wales, even if those persons had been acquitted. In 2007 the government proposed an extension so that bio-information can be kept from anyone who has been arrested, whether they were charged or prosecuted or not. Professor Sir Bob Hepple, ex-Chair of the Nuffield Council on Bioethics, considers: ‘The establishment of a population-wide forensic DNA database cannot be justified at the current time. The potential benefits would not be great enough to justify the cost and intrusion to privacy.’11
Reassuringly in December 2008 the European Court of Human Rights in Strasbourg ruled that keeping innocent people’s DNA records on a criminal register breached Article 8 of the Human Rights Convention, covering the right to respect for private and family life. As a consequence,* police forces in England and Wales could be forced to destroy the DNA details of hundreds of thousands of people with no criminal convictions.12
The public has been treated to the spectacle of numerous government departments managing to leave important files in railway carriages or on office desks for inexplicable periods of time, or relying on pigeons to convey super-sensitive personal information from north to south. With security like this, goodness knows where my DNA might end up. And that is exactly the problem. If the police are going to be allowed to trawl through the databank following a serious crime, I am then going to have to prove why and how my DNA might have arrived at a particular crime scene, and why I am not the perpetrator. Since it may have got there merely by sitting in a cinema seat, and because the match comparisons may have occurred a long time after the crime, this is going to pose considerable difficulty for the innocent citizen, leaving aside the fact that it reverses the normal burden of proof in criminal cases.
DNA is undoubtedly a remarkable discovery, but like any scientific advance it should be treated with respect, and we should all be vigilant about its limitations.
* It took the Home Secretary five months (to May 2009) to propose very modest measures which barely satisfy the ruling. In short, the DNA profiles of the unconvicted can still be retained for between six and twelve years depending on the gravity of the allegation.