As someone who was asking questions in criminal investigations a long time ago, I am interested in the advances of forensic science. These have made the lives of investigators much easier. But - and isn't there always a But - they have made legal matters more complicated. Defence lawyers love having opportunities to cast reasonable doubt on evidence. With advances at the cutting edge, these opportunities are many. Maybe too complex for the jury made up of the average layman. I personally see little difference if the chances of a DNA match are that only one person in, say, six billion will match are reduced by defence work to one in six million. There will be other evidence.
Just don't believe all you see on the CSI series of TV shows.
Running in America is a case where a woman alleges she was raped by a number of members of a sports team. It has raised considerable contraversy regarding DNA evidence. Someone who admits to being a militant feminist put a number of questions to experts in the DNA field. She has highlighted one of the responses as being - sorry to do this - seminal to the matter.
Simon Ford
Trained primarily in molecular biology and biochemistry, Dr. Ford is the Founder and President of Lexigen Science and Law Consultants, a firm that specializes in providing advice to lawyers about genetic evidence since 1988. He has personally reviewed the DNA evidence from thousands of criminal investigations and has conducted numerous workshops for agencies on the analysis of STR test results.Barry:For your information, I've not been following the reports of the Duke rape case allegations, all I know is that the incident allegedly involves multiple potential assailants, and that the initial DNA report failed to establish a link with any of the individuals tested from the Duke team — that's about all I know.
One important issue underlying the media coverage of a story like the Duke incident is that TV shows like CSI have raised unrealistic expectations in the general public of what forensic science, and specifically DNA testing, can do. People get the idea that you run a test, it takes just a few hours, and you get a definitive answer yes or no. In truth DNA testing takes much longer, can produce ambiguous results and has always got to be considered in the framework of the specific question asked and in the context of other biological tests, such as microscopic observation of cell types (for example sperm in a rape case) or serological tests for body fluids (such as semen). Also there are different types of DNA tests. Many labs start with the standard autosomal STR test on sex assault cases. This test looks at regions of DNA which differ from person-to-person and are scattered across the non-sex chromosomes. In some instances, for example samples which contains large amounts of (female) victim DNA, the assailant's DNA may get swamped out, and so the lab can use one of the Y-STR tests, which homes in on variable regions only found on the Y-chromosome (which men have and women do not). In cases of this type, the initial (autosomal) STR DNA report may fail to report a match with the suspect but a later Y-STR DNA report may incriminate. There can also be other simpler explanations for a negative initial report being followed by a later incriminating report, such as refining the choice of samples to test. I would not read too much into a negative report until the whole testing scenario has been completed.
With regard to your specific questions. I don't know much about testing for latex and detection of condom use. The question regarding mixtures is easiest to deal with. A mixed DNA profile will consist of the DNA profiles of the contributing individuals superimposed, one on top of another. It can be quite complex, because not all contributions are going to be in equal amounts, causing unevenness in the profile, and the DNA profiles of the individuals who contribute least to the mixture may well drop below detectable levels. So the first potential problem is that minor contributors may be missed. Beyond that though, it is still possible to answer the question as to whether a specific profile can be excluded as being a potential contributor to the mixed profile.
Once you have determined that a particular profile cannot be excluded the problem is how to express the significance of that observation; and this is were statistics comes in. The statistical calculations for mixture cases can be quite complicated. In single source cases labs often report stats in the trillions and quadrillions but in mixture cases the stats can be dramatically eroded — the more contributors, the less powerful the stats — it is not uncommon to see stats in the tens or hundreds in mixture cases. The presence of multiple contributors can erode the statistical power of the test to the point whereby, if there are enough contributors, the test really doesn't have any useful meaning any more. All the lab would be able to say is that there were a large number of contributors, but no further conclusions could be drawn.
Your two remaining questions are more difficult to answer, because they deal with DNA evidence without giving the context of the specific question being asked for that case or information about other biological testing. Every case is different and has its own nuances. With regard to DNA it all hinges on whether the specific allegations claimed by the victim are amenable to being proven right or wrong by DNA. If the claim is such that one would expect to see biological material and none is found, then sure it may be an indication of a false claim, but there are really so many other potential explanations, particularly when just dealing with the first round of DNA testing, such as issues like condom use, vasectomy, choice of test (autosomal STR v. Y-STR), choice of samples to test, many other things like this can all play a part.
Sorry I can't be more definitive, but I hope these comments help illustrate the difficult challenges that labs and attorneys face with regard to the real world usage of DNA evidence in criminal cases.
All the best,
Simon Ford
No comments:
Post a Comment