Testing the utility of dental morphological traits commonly used in the forensic identification of ancestry

Few human variants are truly population specific, with 100% frequency in one group and 0% in others. However, for traits to be of use in forensic identification they must be as specific to a population as possible. Forensically, several dental morphological traits have been described as useful for determining an unknown individual's ancestry. For these traits to be of value, they should occur in their associated group in proportions statistically different from all other groups. Furthermore, ancestral groups not associated with the trait should have no significant frequency differences among them. To test this, frequencies of dental morphological traits listed in the forensic literature as useful for ancestry determination were compared among samples of African Americans, European Americans, Hispanic Americans, and Native Americans (n=1625). chi2 tests were conducted on dichotomized frequencies of ten trait observations, including incisor shoveling, Carabelli's trait, canine mesial ridge, and cusp seven. Results were mixed. For example, Native Americans have statistically different frequencies of shovel shaping from all other groups. However, statistical differences are seen among other groups as well. Only canine mesial ridge was consistently different in African Americans and not different among the other three groups. Unfortunately, this trait is not common in any group (2-21%), so lack of the trait is not indicative of ancestry. Overall, these commonly used traits may not be of much actual value in ancestry determination. Combining these traits with other observations and more sophisticated statistical tools may be of more practical value.