Law Report: DNA test use questioned 9 June 1994: Regina v Gordon (Michael). Court of Appeal (Criminal Division) (Lord Taylor of Gosforth, Lord Chief Justice, Mr Justice Owen and Mr Justice Latham). 26 May 1994

Courts should be careful to ensure that DNA identification evidence, and in particular the use of probability figures running into millions, whose dramatic quality could exert a strong influence on jurors, could properly be sustained. Expert opinion as to the accuracy of the tests employed was admissible to assess such evidence.

The Court of Appeal (Criminal Division) quashed the conviction and ordered the retrial of Michael Gordon on two rape charges.

David A Poole QC and James H Gregory (assigned by the Registrar of Criminal Appeals) for the appellant; Michael Shorrock QC and Howard Bentham (CPS, Manchester) for the Crown.

LORD TAYLOR LCJ, giving the court's reasons, said the two rapes had occurred in similar circumstances in ground-floor flats in Manchester. The Crown's case was that there was a match between the semen found on a duvet cover in one case, and from vaginal swabs in the other, with the DNA profile taken from the appellant's blood. The first stage of a DNA test involved a laboratory comparison between the DNA profiles. Any unexplained discrepancy would produce a mismatch and so exclude the suspect from complicity.

But even if a match was found, a DNA profiling was not unique: the second stage was therefore the statistical evaluation of the match. However, since any scientific analysis or measurement would produce some variation even between identical samples, allowance allowance was made for an acceptable range when assessing a match.

Dr Sullivan, from Cellmark Diagnostic Laboratories, Abingdon, on whose evidence the Crown principally relied, had adopted a variation of plus or minus a millimetre and on that basis found a visual match between the crime stains from the two rapes, and the appellant's blood sample.

On this appeal, the court also heard Dr Greenhalgh from the Metropolitan Police laboratory, who said the measurements and the basis on which they were arrived at were insufficiently precise or secure to justify the use of a statistical database to calculate the match probability.

The database was calculated by taking a sizeable number of persons of the relevant section of the community, in this case Afro-Caribbeans, and establishing what proportion of them had the characteristics found in the critical sample. Figures had then been put before the jury as to match probability.

In his summing up, the judge said: 'Using a database of 300 Afro-Caribbeans, the probability of six such similar results on Afro- Caribbeans is less than one in 88 million . . . Dr Sullivan said that as a conservative measure . . . the probability is reduced to one in less than 10.5 million' and that the 'probability of a trans-match in two unrelated persons was less than one in 159 million'.

Their Lordships did not doubt the validity and value of DNA evidence in general. But in this case it raised some arguable questions on whether the match probabilities put to the jury and summed up to them by the judge could properly be sustained. Figures running into millions of the kind put before the jury had a dramatic quality which might well exert a strong influence upon them.

Their Lordships could not be sure the jury would necessarily have reached the same result had they heard the additional expert evidence of Dr Greenhalgh.