Interpreting DNA Mixtures

• NCJ Number: 172960
• Author(s): B. S. Weir, C. M. Triggs, L. Starling, L. I. Stowell, K. A J. Walsh, J. Buckleton
• Date Published: January 1996
• Length: 22 pages
• Agencies: NIJ-Sponsored

Annotation

The use of genetic markers has proved invaluable in identifying perpetrators and exonerating falsely accused suspects in a large variety of crimes, but the interpretation of genetic profiles from biological samples requires care when the samples contain material from more than one person.

Abstract

Likelihood ratios are developed for a series of common situations involving people whose genotype is known and who are believed to have contributed to the profile under the stated explanation. The interpretation of mixed stains is straightforward in the likelihood ratio context. Alternative explanations for mixed strain profiles need to be specified and then compared on the basis of probabilities of the profile. When there are only two explanations for an item of evidence, the likelihood ratio is separate from prior probabilities. The general form of the Bayes' theorem, however, means that priors and likelihoods become intermixed in the same expression. Prosecution and defense may have different interpretations of numbers, but there should be as little disagreement on which numbers to present as there is for stains from single contributors. The mathematical analysis covers likelihood ratios in the context of conditional probabilities, the ethnicity of unknown people, population substructures, and effects of relatives. Detailed mathematical formulations are appended. 9 references