Ancient DNA reveals pervasive directional selection across West Eurasia

[DylanY684]

Human evolution did not magically stop when humans stopped living in caves. In fact, new research from David Reich’s lab at Harvard shows that for some traits it has sped up in the past 10k years (at least in western Eurasians).

By comparing differences in DNA ancient remains with living humans’ DNA, Reich et al. were able to identify traits that were under evolutionary selection in western Eurasians during recent evolution. One of these traits is higher intelligence, which has been under positive selection. In other words, in the past 10k years, people with DNA variants that today are seen in smarter people were more successful in their reproduction.

Other traits that were under positive selection pressure include B blood type, celiac disease susceptibility, multiple schlerosis susceptibility, lighter hair and skin tone, walking pace, household income, and years of schooling. Traits selected against Type A blood antigens, include bipolar disorder, schizophrenia, being a current smoker, having a health decline in old age, and body fat percentage (see 2nd and 3rd images).

What’s interesting is that this is a mix of purely biological traits (e.g., blood type) and purely psychological traits (like intelligence). So, evolution worked on both body and behavior in recent times. Additionally, some of these traits would have shown up differently thousands of years ago than they do today. For example, ancient Europeans and west Asians didn’t have access to tobacco until 500 years ago. But for thousands of years before that the behaviors that today lead to people smoking were selected against–even though those people never saw tobacco in their life. SOMETHING biological and/or behavioral was being selected against… we just don’t know what that was. The same is true with intelligence: no one ever took an IQ test before 1905. Yet, higher intelligence was selected for long before the tests were invented.

The article also confirms a lot of previous studies, such as the finding that intelligence is influenced by a large number of genes, each with a small effect. The authors also found evidence supporting the idea that a partial reason IQ is positively correlated with household income, years of schooling is that these traits share genes (see 4th image).

This is a landmark article that tells us so much about the genetics of many different traits (including intelligence), local human evolution, and how individual differences develop over the generations.

Original post: https://x.com/RiotIQ/status/2051647600609419768

Access the full article here:
https://doi.org/10.1038/s41586-026-10358-1

[M.Evanta]

A landmark study from David Reichs lab at Harvard published in Nature earlier this year directly addresses this using ancient DNA from over 15000 West Eurasian individuals spanning 10000 years. By tracking allele frequency changes over time in ancient remains and comparing them to modern DNA, they identified traits under consistent directional selection. Intelligence polygenic scores increased steadily over the period, meaning people carrying variants associated with higher cognitive ability today were reproducing more successfully across thousands of years before IQ tests existed. The same analysis found selection against variants associated with schizophrenia, bipolar disorder, body fat, and health decline, and selection for variants linked to walking pace, years of schooling, and household income. The genetic correlations between these traits are also substantial, which the study interprets as partial evidence that intelligence and socioeconomic outcomes share underlying genetic architecture rather than just correlating through environment.

The honest caveat the authors include is that we do not know exactly what phenotype was adaptive in ancient environments. Nobody took an IQ test in 4000 BC. The variants being selected for are defined by their associations in modern industrialized populations, and those associations may not map cleanly onto what was actually fitness-relevant in the past.

[VeronicaTale]

The smoking finding is the one that makes this concrete. Genes associated with current smoking behavior were being selected against for thousands of years before tobacco reached Europe. Something behavioral was under pressure long before the specific behavior existed. The same logic applies to intelligence selection. The trait precedes the measurement by millennia.