Natural Selection Shaped Hundreds More Human Genes Than We Thought, Massive Ancient DNA Study Finds

The new work, which analyzed DNA from nearly 16,000 people spanning around 18,000 years, found that natural selection had acted on 479 gene variants across West Eurasia – far more than the 21 instances previously identified from ancient DNA.

More than 60 percent of those variants have known links to traits we can measure today, including skin tone, disease risk, blood type, and even susceptibility to alcoholism. Selection appeared to accelerate after the introduction of farming, as a new way of life brought new pressures to bear on human biology.

"This single paper doubles the size of the ancient human DNA literature," said senior author David Reich, Harvard University, in a statement. "It reflects a focused effort to fill in holes that limited the power of previous studies to detect selection."

Natural selection is a cornerstone of biology that drives evolutionary change. While individual gene mutations are random, pressure from the environment weeds out those that harm survival and boosts those that help an individual stay alive long enough to reproduce, resulting in advantageous genes spreading through populations at rates higher than you would expect by chance.

The classic example from recent human prehistory is lactase persistence: a mutation that keeps the lactose-digesting enzyme switched on into adulthood. It spread rapidly through dairy-farming populations in Europe and Africa, presumably because the ability to drink milk conferred a significant nutritional advantage.

You might have thought this effect would be easy to spot across lots of genes, but it’s a little more complex than that. For one thing, researchers have had to rely on extrapolating backwards from modern genomes, essentially doing detective work on the signals that selection leaves behind.

If a variant is strikingly common in Europeans but rare in East Asians, for example, a geneticist can guess that something must have driven it up in one group but not the other.

The difficulty is that this something doesn't have to have been natural selection. There are other scenarios that can leave similar traces. Take the founder effect, for example. This is a phenomenon where a population goes through a "genetic bottleneck" – a period when there were very few individuals in the population.

You can imagine this happening when, for instance, a small group of families breaks off from their ancestral population by crossing a mountain range and spreads into the lands beyond.

This group won’t be perfectly representative of the population they left behind, so there will be differences that they pass on to their descendants. Despite not conferring any kind of advantage or disadvantage, some of these genes will end up being more or less common in this new population than the old one.

Looking back from the present, a modern scientist who spots this disparity will have a tough time determining whether the difference between the two populations was the result of selection or purely down to chance.

That’s where this new study sets itself apart. Having spent seven years building a collection of DNA sequences from ancient people living in West Eurasia – what is now Europe and parts of the Middle East – the researchers could observe how genetic changes occurred over time and place, rather than having to infer everything retrospectively from today’s genomes.

Their work drew on new data from 10,016 ancient individuals from West Eurasia, to which they added 5,820 previously published ancient sequences and 6,438 modern ones.

In addition, they developed new computational methods to isolate the signal of directional selection from other causes of gene frequency changes, such as migration, population mixing, and the kinds of random fluctuations that occur with small populations and cause the founder effect.

“With these new techniques and large amount of ancient genomic data, we can now watch how selection shaped biology in real time,” said first author Ali Akbari. “Instead of searching for the scars natural selection leaves in present-day genomes using simple models and assumptions, we can let the data speak for itself.”

One of the 479 gene variants for which they found evidence of directional natural selection was CCR5-Δ32. This variant confers complete resistance to HIV-1 for people that carry two copies, and it is thought also to confer some resistance to the Black Death.

The researchers found that it was positively selected for between 6,000 and 2,000 years ago, increasing from approximately 2 percent to 8 percent of the population. Obviously this is way out of sync with the famous medieval outbreak of the plague, but there are some signs that the pathogen behind the disease, Yersinia pestis, was endemic to Europe during the Stone Age.

The authors emphasize that, as a proportion of all human genes, finding strong directional selection for 479 of them is still only a faint signal, and it accounts for only 2 percent of total gene frequency changes over the time period. They also stress that it is difficult to know whether the trait we associate with any of those genes is the reason for its change in frequency.

For example, one of the genes for which they found evidence of directional selection is associated with red hair, but you can’t know from this alone whether red hair was advantageous to Neolithic humans in West Eurasia. Rather than that trait being advantageous in and of itself, it might be that the gene happens to be near another that is beneficial and it got carried along for the ride. The gene might also have a secondary effect that we simply aren’t aware of, such as the case of the genes that cause sickle-cell disease and favism (a form of acute hemolytic anemia), both of which confer a negative and positive effect.

The study is published in Nature.