How fast can evolution occur? More precisely, how fast can adaptation occur? The rough answer is pretty fast. For humans that’s clear when you read a book like 10,000 year explosion, or see the results from ancient DNA in papers, that selection (~1% coefficient) on variation can drive allele frequency changes rapidly (~0% to ~100%).
A new paper in PNAS on Yakutian horses is interesting because it highlights the power of selection and how quickly it can change populations when there’s enough genetic variation for adaptation to proceed. These horses are a stocky and hairy breed optimized for life in extreme cold, as Yakutia is in the heart of Siberia, as everyone who has played Risk knows. As a Palearctic species (with origins in North America!) horses have been present in this region since the last last Ice Age. So one model is that the Yakutian horses are a long present population adopted by the Yakut Turkic people, who arrived form the Altai in the last ~1,000 years. The alternative model is that they’re descended from populations brought by the Yakuts, and have undergone recent adaptation. Finally, there is a synthetic model which allows for admixture between the local and intrusive groups, so that adaptation can occur through introgression of alleles from the former to the latter.
To answer these questions they took a genomes from the extant population, as well as ancient or historical genomes. The TreeMix plot makes it pretty clear that the Yakutian horses are intrusive, and do not descend from the ancient populations, represented by Batagai. CGG101397 is from the 19th century, and you can see that is placed near modern individuals from the region, as you’d expect if they arrived with the Yakuts. Interesting, the Batagai individual is even further out that the Mongolian wild horse, which itself is rather diverged from the domestic population, which is not derived from it. They tested for non-trivial admixture between the Batagai lineage and the modern Yakutian horses, and did not find any evidence. This tilts the playing field toward the idea that Yakutian adapatations occurred from their own variation. Mind you, I don’t think they’ve totally eliminated the population of introgression with their sample sizes, as you can have alleles move between populations even if there is almost not detectable admixture.
But there are other elements which suggest that ancient variation wasn’t used. They note that selected variants seem to often be cis-regulatory elements, and not nonsynonymous ones. The former alter gene expression, while the latter presumably impact the function of the proteins, as they’re ultimately transcribed and translated from genes. There is a long argument in the geneticists community around this topic, and Sean B. Carroll’s book from the 2000s, Endless Forms Most Beautiful makes the case for cis-regulatory elements. Here’s a paper from Hopi Hoekstra and Jerry Coyne taking the other side. The authors argue that because the population didn’t have enough standing genetic variation there was a focus on these noncoding elements, which can alter regulation in a quantitative fashion.
The list of genes are what you’d expect when it comes to climate adaptation in this region. What is interesting, and noted in the abstract, is that mammoth and human also seem to have used the same mechanisms and pathways. What this implies is that there are certain genetic constraints, or at least low hanging adaptive fruit, at the level of the mammalian taxonomic rank. I’d go further judging by what we know of pigmentation. Basically, genetic variation is not endlessly pliable and infinite. The wheel is always reinvented, but from the same parts that have long been handy.