The 82nd Annual Meeting of the American Association of Physical Anthropologists (2013)

Withdrawn. Applying statistical classification methodologies to morphological dental trait data in forensic studies


1Dipartimento di Biologia Ambientale, Università di Roma "La Sapienza", 2Sezione di Antropologia, Museo Nazionale Preistorico Etnografico "Luigi Pigorini", 3Facultad de Ciencias Antropológicas, Universidad Autónoma de Yucatán

Saturday All day, Clinch Concourse Add to calendar

Dental morphological studies have clearly indicated their potential in the analysis of both extant and past populations. However, when dealing with single samples, especially forensic, the potentiality of dental traits is not applied to its full strength since the studies tend to consider only a limited number of the available characteristics and their conclusions are drawn on the basis of the presence or absence of single traits. In the present study we build the database framework that enables the comparison of forensic samples to the worldwide distribution of the traits in their whole.

The initial effort implied the structuring of the reference databank. Available dental traits from autochthonous living and archaeological samples from all continents were scored, using the Arizona State University Dental Anthropology System, and the resulting data pooled into ten regional groups. Statistically relevant traits were thus isolated and used to define standardized individuals for each group that enable the comparison with dental traits from isolated individuals.

At its current level of detail the databank leverages relevant data points for most of the regions considered, sample individuals cluster with the standardized individual of the same area or with those from its wider macro region, and has thus far provided results that are very encouraging. Initial testing shows that it consistently provides highly discriminating results for individuals from autochthonous populations, confirming its value for genetic classification.

comments powered by Disqus