« View all Articles

DNA Evidence in the Dock

By Bob Woffinden on

DNA Evidence in the Dock

From Inside Time February 2010 

At the end of last year, the Court of Appeal produced a very important judgment on the safety of DNA evidence. Unsurprisingly, perhaps, given the timing – judgment was given at the start of the Christmas period, on 21 December – hardly anyone noticed.

Bob Woffinden writes...The appeal concerned two separate cases: a murder case in which two brothers, David and Terry Reed had been convicted; and a sexual assault in which the appellant was Neil Garmson. The issue ostensibly at the heart of this linked appeal was the reliability of low template DNA – that is, DNA which has been found in such minute quantity that a standard DNA test would not be expected to yield reliable results.

Routine DNA profiling can be carried out on a sample as small as a nanogram of material (that’s a billionth – a thousand millionth – of a gram). However, from 1999, UK scientists began to work with even smaller quantities and to develop a technique termed low copy number (LCN).

With LCN, it was theoretically possible to produce a DNA profile from a single body cell. Current estimates of the number of cells in the human body range between fifty trillion (that’s fifty million million) and ten thousand trillion.

Obviously, the DNA needs to be amplified. This is achieved, for LCN purposes, by copying the material in a 34-cycle process, as opposed by the standard one of 28 cycles, during which it is treated chemically so that markers (alleles) become coloured and are then displayed as peaks (loci) along a DNA graph. Ideally, there should be up to two markers at ten different loci. The resulting chart will require expert interpretation.

However, there were problems with LCN technology, partly because of what were termed the stochastic effects (that is, random data, which make interpretation difficult if not impossible), and features such as drop-out (that is, some alleles not being visible or reproducible). Additional difficulties arise where there were mixed profiles (that is, the sample contained the DNA of more than one person), if different samples had been pooled, or where there were doubts about the integrity of the sample.

Further, because of the smallness of the sample, nothing could be inferred about the nature of the original material from which the DNA had been extracted.

Some argue that the LCN process was pressed into judicial service before its worthiness had been fully assessed and, indeed, that it has still not been tested sufficiently. The technology has never been accepted in the United States or in the majority of jurisdictions in continental Europe, where scientists believed that LCN techniques were too uncertain for courtroom use.

So, in 2007, the incoming forensic science regulator, Andrew Rennison, commissioned a study from Brian Caddy, emeritus professor of forensic science at the University of Strathclyde, into the reliability and forensic value of low template DNA.

While this study was still being undertaken, however, a judicial flashpoint occurred. 
In December 2007, Crown evidence based on LCN DNA was given in Belfast at the trial of Sean Hoey, who was charged with the Omagh bombing of August 1998. Defence experts who challenged this evidence included Dr Dan Krane, professor of biology in Dayton, Ohio, and Professor Allan Jamieson of the Forensic Institute in Glasgow.

The case was thrown out, primarily because the judge, Mr Justice Weir, did not receive satisfactory assurances about the integrity of the scientific exhibits. However, he also queried ‘the present state of the validation of… the LCN process and in consequence its reliability as an evidential tool’.

This underlined the significance and, indeed, urgency of the issues before the Caddy inquiry. Trials in which LCN DNA evidence played a key part were going ahead throughout this period; Crown expert witnesses gave no indication to juries that the scientific process was under review.

The report was published in April 2008. Caddy recommended that the Forensic Science Service should, prior to testing, quantify the amount of DNA material available. The FSS began to do this routinely in the summer of 2009.

Caddy concluded that the LCN process was valid, albeit under certain conditions. He said that there should be national agreement about how LCN DNA profiles were interpreted (international agreement was clearly an impossibility). He said that the potential pitfalls of the science should be clearly explained to juries – especially, that nothing could be inferred about when the DNA material was deposited; and that the possibilities of secondary transfer were greatly increased. In other words, the suspect may never have been in direct contact with the surface from which the DNA had been lifted. Secondary transfer is, as virtually all scientists acknowledge, not well understood even today.

Such was the background to the Reed-Reed-Garmson appeal, in which the scientific experts for the defence were Professor Jamieson, again, and also Dr Bruce Budowle, who was until last year senior laboratory scientist with the FBI in Washington.

Ironically, however, by the time it was heard, this wasn’t an LCN appeal at all.

The DNA recovered from a crime scene is routinely divided into three parts, which are termed aliquots. Two are tested and the third held in reserve. Because, during the period between trial and appeal, Professor Caddy had recommended quantification, the FSS had measured the amount of material in the remaining aliquot and discovered that there was 2.5 nanograms of DNA material – more than sufficient for standard DNA testing.

This meant that the original judgment of the Crown’s forensic scientist, Valerie Tomlinson, about the quantity of material available had been wrong, so that the FSS had initially performed the wrong tests on the material.

Dr Budowle therefore asked for the material in the remaining aliquot to be tested by the conventional DNA process, and this was satisfactorily completed. There was no longer any issue in the case about LCN DNA. However, the Court of Appeal had been tasked to provide fresh guidelines on the admissibility of such evidence and so, notwithstanding its irrelevance to this appeal, that is what they did.

In his evidence, Budowle testified that LCN was not a robust technology, that the raw samples contained too little DNA and that the results were inherently non-reproducible. The FBI would not consider it safe to rely on such evidence.

However, the judges – Lord Justice Thomas, Mr Justice Kitchin and Mr Justice Holroyde – concluded not merely that LCN DNA was admissible at trial, but that challenges to the validity of the method should no longer be permitted, providing that the amount of DNA was above the stochastic threshold.

So what is the stochastic threshold? The judges said that it ranged between 100 and 200 picograms. (A picogram is one trillionth of a gram.) Bearing in mind that there must be a safety margin in all science, this should mean that it is currently 200 picograms. The judges added, ‘It may be that an opinion can be expressed on profiles [from] below 200 picograms, but that question was not before us and we express no view’.

Yet this is not as straightforward as it seems. Some scientists argue that the judges misunderstood the issue. What matters is not whether the threshold is crossed but whether there are stochastic effects at all; if there are, the profile is inherently unreliable. For example, there could be more than 200 picograms of material, but if it is a mixed sample, there may well be stochastic effects.

However, the judges also emphasised that the jury must evaluate the DNA evidence ‘in the context of all the other evidence in the case’. In other words, it would be unwise for juries to regard LCN DNA evidence itself as definitive.

In the Reed brothers case, which concerned the murder of Peter Hoe in Grangetown, Teesside, plastic fragments of two knives found beside the body bore traces of each man’s DNA.

Giving trial evidence, Valerie Tomlinson said that this demonstrated that the brothers were ‘handling the knives when they broke’. The defence argued that, in saying this, she had gone beyond her remit and in effect usurped the judicial process.

The appeal court judges found her ‘an impressive witness’, but nevertheless criticised those comments, saying that there was ‘no reliable scientific basis’ for them and that they were ‘impermissible’.

The judges said that her evidence should have been fully challenged at trial; but it wasn’t. ‘The decision not to call that evidence cannot be revisited’, they declared, ‘it is not in the interests of justice now to admit evidence simply on the basis that, if it had been called, it might have affected the decisions of the jury’.

Call me pedantic, but I’d have thought it was absolutely in the interests of justice to do precisely that.

One of the concerns which legal observers will have about this judgment was the way in which any mistakes of Crown expert witnesses were treated indulgently while the expertise of defence witnesses seemed to be disparaged at every turn. It is unlikely that Dr Budowle returned to Washington with a high opinion of British justice.

It is the other case, the sexual assault case, that illustrates real difficulties with LCN DNA evidence. Neil Garmson, the appellant, was convicted of disturbing a couple having sex in their car. The attacker pushed the boyfriend out of the car and forced the female to give him oral sex. Neither the victim nor her partner identified Garmson. Indeed, both described the attacker as having long hair; Garmson had virtually no hair. So the Crown asserted, somewhat improbably, that he had been wearing a wig.

Garmson must accordingly have been convicted on the DNA evidence. There was, for example, a profile from a saliva stain on the woman’s knickers. The Crown believed that some of the alleles matched the boyfriend, others matched Garmson, and one that didn’t match either was ‘contributed by an unidentified male’.

Yet this DNA chart could have been read differently. Dr Sara Short, in a report for the defence which was not used at trial, said that one explanation of the profile was that the alleles all came from one source – who was not Garmson. At appeal, Jamieson supported this interpretation, saying that what was certain was that the DNA of another male (who was not Garmson) was present.

The judges dismissed this case also, arguing, again, that ‘the evidence Professor Jamieson gave at appeal could and should have been given at trial’.

So the judgment places a heavy onus on all parties at trial – the scientists, the lawyers, and the jurors – to get it right at first instance. This appeal made it crystal clear that mistakes will not be rectified afterwards.

The gravest responsibility, of course, rests with the jury. They have to comprehend the scientific issues, no matter how seemingly impenetrable, and assess them accurately.

Yet juries have tended to be blinded by science, and in particular by DNA science. The popular presumption is: if they’ve got your DNA, they’ve got you. That’s the impression that has been propagated. Prosecutors have done nothing to rectify it and have neglected to point out to the public what was emphasised here: that DNA evidence must be considered in the context of all the evidence in the case. As we now know, this especially applies when the DNA is low copy number.

The Forensic Science Service maintains that there have been no convictions on LCN DNA evidence alone. Yet it seems to me that there are many cases where such evidence has been crucial. For example, in the still-controversial case of James Hanratty (who was executed for murder in 1962) the LCN DNA is the only extant evidence; everything else has been discredited.

There are three remaining points. The first is how frighteningly simple it has become to frame someone. If someone has criminality in mind, but wants to avoid detection, they need only take with them to the scene something bearing someone else’s DNA – a dirty glass, a smoked cigarette butt – and the chances are that they can commit the crime with impunity.

Of course, that is precisely why the scientific evidence must be considered alongside all other evidence.

The second issue of huge importance is obvious: what about all those cases heard during the ten-year period from 1999 in which LCN evidence was given, but the amount was never quantified? Given what we know now, how reliable was that evidence?

Where there is additional cogent evidence, the conviction may well be safe; but where the other evidence is slender or non-existent, and perhaps the weight of the scientific evidence was overstated at trial, then those cases may need to be looked at afresh.

Finally, LCN was specifically designed to deal with extremely low levels of DNA – as low as 100 picograms. What is going to be the courtroom approach from now on if quantification of samples shows them to be less than 200 picograms? Will the CPS now have to abandon such cases, even though they would previously have progressed them?

In giving this judgment, the judiciary wanted above all to put to bed the continuing DNA debate. It seems likely, however, that they have merely achieved the opposite – and heightened concerns about a significant number of past cases and introduced difficulties for future ones.