Our new paper 'Comparing the Accuracy of Osteological Sex Estimation Against Genetic Sex Obtained From Human Dental Calculus in 12th–19th C. British Individuals' is out now in the American Journal of Biological Anthropology.

This paper compares conventional macroscopic methods of sex estimation against genetic sex determined from dental calculus samples. We found almost complete agreement between macroscopic methods of sex estimation from the pelvis and genetic sex, but greater variability for the skull. This is the first paper to use genetic sex obtained from dental calculus, which turns out to be a remarkably good source of endogenous DNA for sex determination.

I collected all the macroscopic sex estimation data for this paper, as well as conducting the statistical analyses, compiling the results, contributing to the writing and editing of the manuscript, and producing the figures for it.

Abstract:

Objectives

For anthropologists and forensic scientists, sex estimation using skeletally dimorphic traits is fundamental in biological profile reconstruction. Archaeogenetic research on past populations allows us to identify genetic sex generating opportunities to validate the accuracy of commonly used skeletal traits. We compare genetic sex estimates, derived from aDNA analysis of dental calculus, with macroscopic skeletal observations to ascertain the accuracy of trait-based estimates on British skeletons.

Materials and Methods

Dental calculus was sampled from 146 individuals from two British sites (Barton-upon-Humber and St James's Gardens [London]), and genetic sex was obtained. A combination of frequently analyzed skull and os coxae traits was assessed to determine overall osteological sex; individual skeletal traits were also evaluated separately. These were then compared to genetic sex. A regression analysis of trait scores was used to identify the best trait combination for sex estimation.

Results

We found high agreement between overall osteological sex and genetic sex (97.9%); the os coxa estimate demonstrated almost perfect agreement (97.8%). Skull-based estimates were less accurate (79.0%). Individual Phenice traits demonstrated the greatest accuracy (80.2%–97.5%), whilst multiple skull traits performed poorly (58.9%–68.8%).

Discussion

Overall, application of osteological sex estimation techniques to British remains was highly accurate, especially using combined traits of the os coxa. However, we identified multiple skeletal features that are not reliable in isolation, but provide new sectioning points and regression equations to improve skull estimate accuracy. We emphasize the need for further work on factors affecting pelvic morphology, and the benefits of dental calculus as aDNA source.