Saturday, April 10, 2010

DNA in the spotlight again

These issues were inevitable and certainly must be addressed. Questions about shared objects and multiple DNA samples, DNA data errors when the analysis time is shortened from months to days, contamination, laboratory procedural errors, and the Big Brother data base.

"The rise of DNA analysis in crime solving"

Its success in the Damilola Taylor and Rachel Nickell murder cases has thrust it into the spotlight, but how does it work, and how ethical is it?


Thomas Jones

April 10th, 2010

The Guardian

In June 2008, a 19-year-old man from Nottingham was arrested for careless and inconsiderate driving. The police took his photograph, his fingerprints and a swab from the inside of his cheek to get his DNA profile. This is routine procedure whenever someone's arrested for a recordable offence – any that is punishable by imprisonment, along with a few others, including those committed by football hooligans, prostitutes, beggars or poachers. Details of convictions and offenders are recorded on the police national computer. You don't have to be convicted, however, for your DNA profile to be kept on the national crime intelligence DNA database: it's enough just to be arrested for a recordable offence. A few months after the DNA profile of the 19-year-old careless driver was uploaded to the database, it was flagged as a close but not perfect match to the profile of the probable killer of Colette Aram.

Aram, a 16-year-old trainee hairdresser from Keyworth in Nottinghamshire, was abducted, raped and strangled on 30 October 1983 – five years before the careless driver was born. (Her murder was featured in the first ever episode of Crimewatch in 1984.) The police had a few circumstantial leads to go on: a stolen red Ford Fiesta; a handwritten message from the killer saying they'd never catch him; a paper towel recovered from the toilet of a pub, the Generous Briton, where shortly after the murder a man with blood under his fingernails had eaten a ham sandwich, drunk a pint of orange juice and lemonade, and told the landlady an unlikely story about having driven up the M1 to see some friends who weren't in. Twenty thousand people were interviewed in the course of the investigation, but the killer wasn't found.

In October 2008, on the 25th anniversary of the murder, Nottinghamshire police announced they had new evidence, derived using the latest forensic DNA analysis techniques. (DNA profiling didn't exist when Aram was murdered, and the national DNA database wouldn't come into existence for another 12 years.) They could now "say with certainty" that Aram had been in the red Fiesta, and that her killer had gone to the Generous Briton. They also had his DNA profile. But it didn't match any of the four million profiles on the database. A new tactic was called for.

The database was searched again, this time for "near misses": profiles similar enough to the killer's that they could belong to a member of his family. The DNA of the 300 closest (male) hits was then re-examined, this time looking at markers on the Y-chromosome: as all the DNA on this is passed from father to son, it's a very good indicator of familial relationships between men (allowing for mutations, my father, uncle, cousin and his son all have the same Y-chromosome as me, inherited from my grandfather). But all 300 near misses came back negative. As more profiles were added to the database, the same checks were carried out.

Eventually, after 600 near misses had been re-tested, the markers on the 19-year-old careless driver's Y-chromosome came up as a match for the killer's. His father and two uncles were arrested in April 2009. Their swabs were flown to the forensics lab by helicopter (the custody clock was ticking) and a positive match to the killer's profile confirmed within nine hours. The careless driver's father, Paul Hutchinson, a 51-year-old newspaper delivery agent, was charged with Colette Aram's murder. He pleaded guilty on 21 December, and on 25 January was sentenced to life imprisonment.

Hutchinson's DNA profile was generated from cells recovered from the car and the paper towel by the Home Office's Forensic Science Service (FSS); the later work – the familial testing and confirmation that Hutchinson's DNA matched the profile generated from the crime stains – was carried out by an independent competitor, LGC Forensics, a division of LGC Ltd. At one time the police would automatically have gone to the FSS for all their forensic work, but since 1999 the industry has been privatised, with different companies competing for police business. Once the state-owned Laboratory of the Government Chemist, LGC Ltd was sold off by the Major government for £5m in 1996. In February this year, LGC was valued at £257m. It's grown in other ways, too, since privatisation: staff numbers have increased from 270 to more than 1,500, and several other firms have been bought up – including, in 2005, Forensic Alliance, then the UK's largest private provider of forensic science services to the criminal justice system.

As well as the Colette Aram case, LGC Forensics has helped to solve other high-profile murders including Damilola Taylor, Rachel Nickell and Chantel Taylor. They helped to identify the bodies of victims of the London Underground bombings of 7 July 2005, and to exonerate the Cardiff Three. With improvements in technology and the growth of the national DNA database – it now contains the profiles of more than 5 million people – DNA analysis is becoming an ever more important tool in solving crime. It's even possible, according to Liberty, that "police may drop investigations if DNA evidence is not found at the crime scene". It's possible, too, that the database, which was set up by the FSS in 1995 and is now run by the National Policing Improvement Agency, is larger than it should be. Following a recent ruling by the European Court of Human Rights, legislation currently going through parliament will place (fairly modest) new limits on whose profiles can be stored and for how long. Yet the mere existence of the database raises questions about the fundamental relationship between citizens and the state – and the opening up of the forensic science market to competition between private providers raises more about whether it's appropriate for the private sector to have any kind of a role in law enforcement.

LGC's headquarters are in Teddington, south-west London, at the north-east corner of Bushy Park. It was overcast and drizzling when I visited last month, but inside all was light. Television shows such as CSI, which is where many people's preconceptions about forensic work (including mine) must come from, have scientists working in moody pools of light in otherwise darkened labs. But as James Walker, who leads LGC's specialised forensic DNA team, points out, the "right examination tools", when you're looking for very small stains on objects recovered from crime scenes, include "good lighting" and "magnifying glasses". There's no place for chiaroscuro effects in a real lab. Walker doesn't watch CSI, but "other people say it is actually quite good".

Emma Westacott, LGC's London head of DNA operations, doesn't watch CSI either, but she says there have been times when she's found herself shouting at the TV during a police procedural: "If only it was that simple!" The fictional forensic scientists seem to get everything done within the brief timespan of a single episode, don't tend to be working on many samples at once, and can sometimes get a "wonderful profile" in very unlikely circumstances. "There was a stage a few years ago where that possibly filtered through to some of our non-scientific police officer customers," Westacott says, "but that seems to have died down recently."

The science and technology have certainly come a long way. Since 2001, standard turnaround times – from crime scene exhibit to DNA profile – have gone down from three months to three weeks to three days. When they need to, as with Hutchinson's profile, they can do it in a matter of hours. And they're currently working on developing a portable kit the police could use to get a rough profile (with a one in a million chance of a match) at the crime scene within an hour. It's also become possible to generate profiles from ever smaller or more degraded DNA samples.

A lot of the material that LGC analyses comes in as swabs taken by police investigators, but finding and removing the human tissue from objects isn't always straightforward, and there are a couple of laboratories dedicated to item examination. Mostly it's a question of things such as cigarette ends, gloves, hats, drink cans, but sometimes there are larger objects, too: "The most challenging item we've had to get through the door so far was the entire back seat of a car," Westacott says.

Different objects present different challenges: a cigarette end is likely to have plenty of DNA on it, but it may come from more than one person, if the cigarette has been shared, and chemicals in the paper or filter can interfere with the profiling process. There's also the potential difficulty of linking the cigarette to the crime in court. Just because the suspect smoked a cigarette found at the crime scene doesn't mean he committed the crime, or that he was even there: the cigarette could have been brought along on someone's shoe. So the condition of the cigarette end has to be documented: does it look freshly stubbed out, or is it dirty and flattened? A blood-stained hammer may be easier to link to the crime, but getting the DNA profile of the last person to have wielded it presents a whole new set of problems.

One of the more notorious items from which LGC scientists have recovered DNA evidence was a trainer belonging to one of Damilola Taylor's murderers. They found on the shoe's heel three spots of the 10-year-old victim's blood that had been missed by the FSS. Walker and Westacott are careful not to disparage their colleagues there. "If somebody's got all day to look at an item," Walker says, then they're more likely to find something than someone who "has spent only perhaps half an hour" looking at it. "That case is not an issue of new technology be"it's just a second person having a look at an item and seeing something different."coming available," Westacott adds,

According to the independent review ordered by the Home Office into the way the forensic evidence was handled in the Taylor case, the shoe was one of 50 items submitted to the FSS on 4 December 2000. All were marked urgent. On 22 December, another 144 items, all marked urgent, arrived. On 16 January 2001, another 78 urgent items turned up in the lab, which was already dealing with the evidence from 200 other murders and 200 rapes. "The tension between the pursuit of excellence and the demand for urgent results is likely to have its effects," the review noted.

The industry is, of course, tightly regulated. All forensic labs in the UK have to be accredited and conform to rigorous trading standards. The conditions for being able to submit profiles to the national DNA database are even stricter. Two or three of the samples LGC receives each year from the police have been secretly submitted by the custodian of the database as a quality control test. This allows them to check the accuracy of the uploaded profiles; they'll also pick one of the samples and demand to see all the records and case files connected with it, to make sure the entire process is running as it should. And there's a lot of paperwork: every movement of every item, from delivery crate to fridge shelf to lab bench, along with the identity of the person moving it, is logged twice, electronically by barcode scanner and by hand on the back of the case file.

Once a DNA sample has been collected from an exhibit – and, hopefully, separated from any impurities or chemical inhibitors, such as the dark blue dye in a pair of jeans – the next thing to do is see how much human DNA you've got, and whether there's still anything else in there with it. Then, assuming you've got a decent amount of clean DNA, it's ready to be analysed.

To get someone's DNA profile you don't need to sequence their entire genome. (Only around 20 people have had their full personal genomes sequenced, including Archbishop Desmond Tutu and the actor Glenn Close.) The profiling system in use in Britain looks at 11 very small regions of DNA – about one millionth of the total. One of them tells you the person's sex; the other 10 are short tandem repeats or STRs. In lots of places in our so-called "junk" DNA (the vast majority of it which isn't involved in making proteins) there are repeating patterns of short sequences of base pairs (the molecules DNA is made up of). The number of times each short sequence is repeated varies from person to person, though within a limited range of, say, 10 to 25. However, the chances of two unrelated individuals having exactly the same number of repeats in all 10 regions used for DNA profiling are one in a billion.

I didn't go into the laboratories, but peered in through the glass in the doors, to see figures in green and red colour-coded outfits working on colour-coded equipment to reduce the risk of cross-contamination between different procedures. I had a mouth swab taken before I was even allowed in the corridor: LGC keeps a record of the DNA profiles of all staff and visitors as a further precaution against contamination.

For the most straightforward analyses – mouth swabs from people the police have arrested – the process has been fully automated and can produce a result in 12 hours. With that hypothetical hammer handle, however, where there's almost certainly not going to be enough DNA to process through the normal route, it's a much more intricate procedure. "It takes about four hours of cleaning before they'll start work," Westacott says. Is that what they're doing now in their space suits on the other side of the glass? No: "They're just cleaning down before they leave, then it will be cleaned again before the next piece of work starts."

Even – or especially – with the most advanced DNA profiling, Walker warns, "you've got to have a lot of caveats in the reporting and interpretation of the results because there can be lots of reasons why a profile has been generated." Westacott adds: "One of the key roles of the expert witness in such cases is to try to educate the jury, so they appreciate exactly what the evidence is."

One of the concerns expressed in the Nuffield Council on Bioethics 2007 report The Forensic Use Of Bioinformation: Ethical Issues was that "the 'prosecutor's fallacy' has compromised the use of DNA evidence for a fair trial". The fallacy here is the idea that as the DNA profile generated from stains found at a crime scene matches the suspect's DNA profile, and as there is only a one in a billion chance of it also matching someone else's profile, then there is a one in a billion chance that the suspect is innocent. But DNA evidence simply can't on its own tell you whether or not someone is guilty.

The prosecutor's fallacy stems, in general, from a misunderstanding of how statistics work. In the case of DNA, however, it may be exacerbated by an almost mystical belief in the molecules' power. Paul Debenham, LGC's director of innovation and development, suggests that there is a pervasive and "unjustified faith that human genome sequencing is going to cure the world".

The public's irrationally positive view of medical genetics goes hand in hand with an irrationally fearful view of the implications of the national DNA database. Even if, as Debenham puts it, "everyone likes it for catching criminals", no one likes the idea of their own DNA being on it. And, he says, "Even if they had all of your DNA today, and for the long foreseeable future, we won't have a clue what that tells us about you, and actually you should be more worried if someone gets hold of your supermarket purchasing history."

The current ethical problems with the database aren't in fact to do with what the government, the police or anyone else may or may not be able to learn about you from your DNA, but with the question of whose profiles are stored on it. According to a report published in November by the Human Genetics Commission, Nothing To Hide, Nothing To Fear?, the database now holds the profiles of approximately 5 million citizens, or 8% of the population of Britain. Among black men between the ages of 18 and 35, however, that figure rises to more than 75%. You don't have to have been convicted or even charged with a crime to have your DNA taken, remember: you just have to have been arrested. The report also quotes anecdotal evidence from a retired senior police officer to the effect that the police now arrest people simply to get their DNA profile.

The crime and security bill currently going through parliament "establishes new time limits for the retention of DNA samples, DNA profiles and fingerprints", following the decision of the European court of human rights in S and Marper v The United Kingdom that the retention of the applicants' fingerprint and DNA data when they hadn't been convicted of any crime was in violation of their right to respect for private and family life. Something that never seems to come up in arguments about whose DNA profiles should or shouldn't be on the database is whether after a certain time convicted criminals should also have the right to have their profiles removed.

There is a reasonable case to be made for having a universal database with everyone's profile on it: that way there'd be no discrimination, no stigma, the police wouldn't have to rely on the chance arrest of a careless driver to catch a murderer, and let's not forget that DNA evidence is a powerful tool for eliminating the innocent as well as pinpointing the guilty: fingering Robert Napper for the murder of Rachel Nickell also exonerated Colin Stagg. And yet… As I was leaving LGC, two plain-clothes detectives from the Met carrying a plastic bag – full, I assumed, of exhibits in urgent need of analysis – were ushered into the forensics labs. And thinking of the mouth swab I'd had taken, even though I knew they needed my profile precisely in order to exclude it, and despite everything else I'd been shown, I felt an irrational twinge of anxiety at the thought of my DNA somehow getting mixed up with the evidence in that bag.


Timothy said...

here in San Diego a young lady was raped and murdered...within a couple of days they had identified the killer thru DNA...that fast. so there are good sides to a database and i am hardly afraid of being on it...we need to remember here paranoia is treatable.

Mercury said...

Well Tim, I understand your reference but one has to be careful in accusing and convicting someone solely on DNA evidence for the very cautionary reasons I mentioned. All it takes is one error to cast suspicion on the whole process. The data base has little to do with paranoia and more with the protection of an individual's privacy and rights and to prevent misuse.