I read an article this week which described an analysis of ancient DNA samples from Portugal to understand how prehistoric populations mingled and merged to form the modern Portuguese gene pool (Martiniano et al. 2017). This type of article is not unusual. Similar studies have been conducted on populations from every inhabited continent.

What interested me about this study was how the researchers merged this research with polygenic scores. As I have explained before, polygenic scores are scores derived from specific DNA base pairs that can be used to make a prediction about someone’s phenotype. Martiniano et al. (2017) used polygenic scores for height and estimated that prehistoric hunter-gatherers in Europe had a genetic propensity for being taller than the later waves of agricultural peoples. As the agricultural groups replaced the early hunter-gatherer groups, the polygenic score for height decreased. Some members of these groups intermixed, though, and later individuals tended to be taller if they had more hunter-gatherer DNA, while later individuals who were shorter generally had more agriculturalist ancestors.

Height is often a phenotype that scientists use when first applying polygenic scores to different situations or populations. It is easy to accurately measure, highly influenced by genes, and many DNA segments are known to contribute to accurate predictions of height. When polygenic scores produces results that make sense and match other observations (like here), then it is a proof of concept that polygenic scores for other phenotypes could work, too.

Figure S32 from Martiano et al. (2017). These five scatterplots show the correlation between the amount of ancestry a person had from different groups and their height polygenic scores. The top three scatterplots show that more ancestry from ancient European hunter-gatherer groups (WHG, CHG, and EHG) was associated with taller predicted height, while having more ancestry from the Anatolian agriculturalists was associated with shorter height.

This is what is so exciting to me about this article. If scientists can use polygenic scores to understand the height of preshistoric individuals, then we can use polygenic scores to make estimates about psychological traits, like personality, mental health diagnoses, or intelligence. These traits usually leave no trace in the archaeological record, and scientists are on the cusp of having this data. It is amazing that one day we might actually know whether a skeleton at a dig site belonged to someone who likely had an outgoing personality or a high risk for depression.

In addition to understanding past peoples better, this information will shed light on 21st century people and how they came to be the way they are. The genes of modern individuals are not a representative sample of the genes of ancient peoples. This is because human evolution never stopped, and the environment continues to select some traits and not others. Contrary to popular belief, people do not have a “caveman brain,” though modern people’s brains are influenced by the traits that their ancestors were selected for.

Thus, comparing ancient and modern genomes will help scientists understand how people are similar to humans of past eras, and how they are different. I would be interested, for example, in understanding whether ancient peoples had a greater propensity towards violence, physical strength, and schizophrenia. If there are differences in these traits, then it may have implications on how cultures and people change with time.

Once the traits that have been selected for in different populations are identified, evolutionary psychologists can start hypothesizing about the environments that caused average group differences in traits to develop over time. Using polygenic scores to understand ancient humans is a breakthrough that has massive implications for differential psychology, evolutionary psychology, and the study of group differences.

If there are similarities between modern and ancient humans on these traits, then it lets scientists know which aspects of human psychology might be universal and constitute a general “human nature.” This can help bind humanity together and help people understand the similarities they share with one another—and with humans in the past.

Postscript: Yes, I am aware that polygenic scores do not make perfect predictions. In fact, they cannot make perfect predictions, unless heritability for a trait is 100%. But they can make predictions that are better than not having an estimate for a phenotype. That means they are still very useful for learning about people (both living and dead).


Martiniano, R., Cassidy, L. M., Ó’Maoldúin, R., McLaughlin, R., Silva, N. M., Manco, L., Fidalgo, D., Pereira, T., Coelho, M. J., Serra, M., Burger, J., Parreira, R., Moran, E., Valera, A. C., Porfirio, E., Boaventura, R., Silva, A. M., & Bradley, D. G. (2017). The population genomics of archaeological transition in west Iberia: Investigation of ancient substructure using imputation and haplotype-based methods. PLOS Genetics, 13(7), Article e1006852. https://doi.org/10.1371/journal.pgen.1006852