Rasudas

A New Web-Based Application for Estimating Ancestry from Tooth Morphology

Dados Bibliográficos

ID5397272
AutoresG R Scott (0000-0003-2328-8003, University of Nevada, Reno, autor correspondente), M A Pilloud (0000-0002-1704-3707, University of Nevada, Reno), David Navega (0000-0003-4495-3762, University of Coimbra), João D’oliveira Coelho (0000-0003-0871-1926, University of Coimbra), João D''Oliveira Coelho, João D'Oliveira Coelho, E Cunha (0000-0003-2998-371X, University of Coimbra), Joel D Irish (0000-0001-7857-8847, Liverpool John Moores University)
Ano2018
Volume1
Fascículo1
Data de publicação2018-01-19
Peer ReviewedSim
Open AccessNão
TipoARTICLE
PeriódicoForensic Anthropology (JOURNAL)
Identificadores do periódicoISSN: 2573-5020 • E-ISSN: 2573-5039
EditoraUniversity Press of Florida (PUBLISHER)
DOI10.5744/fa.2018.0003
OpenAlexW2785282562
IdiomaEN
Citações recebidas23
Referências citadas23

The use of crown and root morphology to estimate population relationships has a long history in dental anthropology. Over the past two decades, methods employing dental morphology within forensic anthropology have been formalized with the incorporation of statistical models. This paper presents a new web-based application (rASUDAS) that estimates the ancestry of unknown individuals based on their suite of tooth crown and root traits. The application utilizes 21 independent traits that were scored following the Arizona State University Dental Anthropology System (ASUDAS). The reference sample represents approximately 30,000 individuals from seven biogeographic regions. A naive Bayes classifier algorithm was created in the R open source programming language to assign posterior probabilities for individual group assignment. To test the application, 150 individuals were selected from the C. G. Turner II database with the proviso that an individual had to be scored for a minimum of 12 of the 21 traits. In a seven-group analysis, the model correctly assigned individuals to groups 51.8% of the time. In a four-group analysis, classification accuracy improved to 66.7%. With three groups, accuracy was at 72.7%. It is still necessary to validate the program using forensic cases and to augment the reference sample with modern skeletal data. However, results from the beta version of rASUDAS are presented as proof of concept on the potential of dental morphology in ancestry estimation in forensic contexts.KEYWORDS: dental morphology, dental non-metrics, quantitative analysis, Arizona State University Dental Anthropology System, Bayesian statistics

Archaeology · Forensic anthropology · Geography · Machine learning · Naive Bayes classifier · Population · Sociology · Statistics · Computer Science · Demography · Forensic and Genetic Research · Forensic Anthropology and Bioarchaeology Studies · Mathematics · Artificial Intelligence