NEW YORK — Researchers have teased out signals of selection on polygenic traits such as pigmentation, body size, and dietary intake, that give insight into human evolution in a new analysis that drew on genome-wide association study data.
The team, led by investigators at Shanghai Jiao Tong University, aimed to quantify signals of selection for 870 polygenic traits — both disease and non-disease related — from GWAS summary statistic data from European populations and examined the strength of these signals across four time periods. As they reported on Monday in Nature Human Behaviour, they found traits associated with pigmentation, body measurement, and nutritional intake exhibited signals of selection across the different time scales: the present day, recent history or 2,000 to 3,000 years ago, the Neolithic period, or since human speciation.
"At first glance, the large proportion of traits under recent positive selection was a little bit surprising to me, but after careful consideration, I would say this result is reasonable," corresponding author Guan Ning Lin, an associate professor at Shanghai Jiao Tong University, wrote in an email.
"First of all, the distribution of strength of recent selection is actually quite close to zero, meaning that most of the traits underwent mild positive selection," Lin said, adding that the analysis included some highly similar traits that could be correlated. He further noted that the proportion of traits under positive selection during the ancient time period was also expected, as genes sensitive to loss-of-function mutations contribute to polygenic traits and most traits are known to be under purifying selection.
For their analysis of recent selection signals, the researchers collected GWAS summary statistic data, largely from the UK Biobank project, for 870 polygenic traits. These included disease traits like Crohn's disease, attention-deficit hyperactivity disorder, and high cholesterol as well as non-disease traits like height, skin tanning, and facial measurements.
Using a Mendelian randomization analysis, they found that body measurement-linked traits were associated with both male reproductive and mating success during the present-day period, but that fewer of them were linked to female reproductive or mating success. The researchers noted this difference is in line with an evolutionary psychology theory that males and females adopt different sexual strategies.
They then shifted their focus to the past 2,000 to 3,000 years using different statistical approaches. In this time period, dermatological traits were under the most significant selection, followed by nutritional intake-related traits, they found.
To peer even further back in time, the researchers examined the polygenic burden of these traits in ancient human genome datasets from the pre-Neolithic and Neolithic eras as well as from Near East farmers who lived 14,000 to 3,400 years ago. A number of traits related to hunter-gatherer ancestry, including ease of skin tanning, were under positive natural selection in the pre-Neolithic period and in Near East farming populations but were suppressed by natural selection during the Neolithic period.
The findings indicated that there has been selection for dark skin prior to the Neolithic period as well as in recent history, but that there was inconsistent selection during the Neolithic period.
Lin noted that the study of skin pigmentation in human evolution has been controversial, with many theories and exceptions to those theories. "I have no doubt that if we expand our analysis to cover more diversity in population and time scale, the direction and strength of positive selection would be even more diverged," he said. "Our analysis could only provide some phenomena, and the mechanism behind them would be difficult to tell from our result."
Additionally, polygenic disease traits were generally under negative selective pressure, though they still lingered, a finding the investigators plan to further pursue.
The researchers noted that their analysis has a number of limitations, including its reliance on data from individuals of European ancestry. Lin said that other groups have been able to examine signals of selection from BioBank Japan data, and that they plan to examine time scales that those studies have not covered.