办事指南

Bearing false witness - Chance matches are much more likely with mtDNA tests

点击量:   时间:2017-09-25 03:02:10

By Jonathan Knight San Francisco A TYPE of genetic fingerprinting with a high chance of producing a false match has helped to convict six people in the US. The increasing use of this type of DNA evidence, based on DNA from the mitochondria of a cell, has sparked heated debate about whether it should be admissible in court at all. The genetic fingerprints most people are familiar with are based on the DNA in the nucleus of the cell. Forensic scientists analyse more than a dozen different chromosomal regions, so the chances of two people having the same fingerprints are extremely low—in the order of a million to one. But in some types of tissue found at a crime scene, such as hair without roots, teeth, excrement and old bone, the nuclear DNA is too badly broken down to analyse. In such cases, DNA in mitochondria, which carry their own genes, often remains intact. One highly variable region of this mtDNA can be sequenced to give an individual fingerprint. William Shields, an expert on mtDNA from the State University of New York in Syracuse, claimed last week that there are huge flaws in the way the FBI analyses the evidence and presents it to juries. FBI expert witnesses fail to emphasise the high probability of a random match, he says. “People assume this form of DNA testing carries similar weight to other forms of DNA evidence.” Based on the FBI’s own published data, however, the odds of mtDNA sequences from two people matching by chance are between 1 in 114 and 1 in 468. “I don’t call that very discriminating,” says Shields. Special agent Mark Wilson of the FBI’s DNA programme says that the FBI always explains these facts to juries. “We tell them straight up that this is not like nuclear DNA,” he says. One problem, Wilson admits, is that the database is still rather small, containing mtDNA sequences from only a few hundred people. “But we spell it all out to the jury,” he says. Since the FBI began using mtDNA profiles in casework in 1996, six juries in the US have heard cases in which mtDNA from either the accused or the victim was presented as evidence. All six trials resulted in conviction. A further criticism of the test is that there is now evidence that the variable mtDNA sequence used for forensic purposes can vary even in one individual. Shields says this phenomenon, called heteroplasmy, increases the chance of a random match between a suspect and a sample from the crime scene. In some cases, the mtDNA profile varies by one or two nucleotides from one end of a hair to another. “We do not yet have an adequate assessment of the frequency of heteroplasmy,” he says. But Wilson argues that, if anything, heteroplasmy is more likely to eliminate a suspect—wrongly—than incriminate an innocent person. Because of these limitations, Britain’s Forensic Science Service in Birmingham uses mtDNA only to eliminate suspects or to back up other pieces of evidence. Dave Werrett, director of research and DNA services at the FSS,