Reconstructing human population structure and history from dental phenotypes: Theory, methods and application to the ancient Greek colonization of southern Italy

Abstract

A central objective in many archaeological and anthropological studies is the reconstruction of past human population structure and population history. Biological distance (biodistance) analysis is a powerful tool to infer genetic relationships across samples of human skeletal remains using cranial and dental phenotypes. The underlying assumption of biodistance analyses is that skeletal samples which share a set of phenotypic features are presumed to be genetically more closely related, while those that differ in their phenotypes are presumed to be genetically less similar. Teeth have become a favored dataset for biodistance analyses primarily because dental size and shape are assumed to be highly heritable and selectively neutral. Moreover, dental form remains unchanged after full formation and teeth are generally bestpreserved in the archaeological record, even when associated skeletal and DNA preservation is poor.

This cumulative dissertation has two primary objectives: (1) to quantify the utility of dental phenotypes as a reliable proxy for neutral genomic data in a population and quantitative genetic framework; and (2) to apply a dental biodistance analysis to an archaeological case study: the Greek colonization of southern Italy.

To address the first objective, I compare existing large phenotypic and genomic datasets sampled from worldwide modern human populations. Specifically, I generate biodistance estimates from two commonly employed dental phenotypic data types (metric and nonmetric traits) and two neutral genetic marker types (SNPs and STRs) using the R-matrix model, and explore their relationship using Mantel tests. Results show that biodistances based on dental phenotypes are significantly correlated with those based on neutral genetic data (on average r = 0.574, p < 0.001), validating tooth form as an efficient proxy for nuclear DNA data. Dental metric and nonmetric traits give concordant but varying results, indicating that combining both data types increases performance compared to using the features separately. Future work seeking to quantify the association between dental phenotypic and neutral genomic variation has great potential to identify dental data combinations that are most useful for tracking human population structure and history.

To address the second objective, I collected a new dataset comprising 481 human skeletons with well-preserved dentitions from six archaeological sites along the coastal area of the Gulf of Taranto, southern Italy, dating to precolonial (900-700 y BC) and postcolonial time periods (700-200 y BC). For both periods, I infer population structure using adonis, betadisper and isolation-by-distance models based on interindividual Gower distance coefficients using a mixture of dental metric and nonmetric traits. For the postcolonial period, I furthermore determine individual ancestries using naïve Bayesian classification based on dental nonmetric traits. Results indicate that precolonial southern Italy was characterized by moderate levels of population stratification. During postcolonial times, the area became a place of high mobility hosting ~ 10 % Greek newcomers and their descendants. Interestingly, individuals of Greek ancestry were equally distributed across Greek colonies and indigenous settlements. These findings support a gradual colonization model with substantial involvement of local populations and contradict the theory that Greek colonies were homogenous enclaves within conquered territories. Future work expanding the study area to Calabria and Sicily has great potential to generate a detailed picture of the colonial history of whole Magna Graecia. The case study in this dissertation provides a conceptual template for this and the provided raw data allow for repeatability.

This dissertation bridges questions and analytical approaches of physical anthropology, population genetics and classical archaeology. Thereby it promotes multidisciplinary synergy between these research fields and highlights productive areas for future research.