11.16.2020 07:00 AM

Ancient Dog DNA Reveals Their Enduring Connection With People

Genetic material from prehistoric dogs shows how early humans may have migrated with, bred, and cared for their canine companions.

siberian husky — Photograph: Craig Moore/Getty Images

For hundreds of years, there was one way to study human prehistory: Put on a pith helmet, go to a desert in Africa or the Middle East, dig up some skeletons and artifacts, and make inferences based on their physical attributes. Then modern genetics came onto the scene. Just by studying the genomes of living humans, it’s now possible to make inferences about how our ancestors populated the planet. The last big break in this case came only a decade ago, when scientists refined their techniques for extracting ancient, degraded DNA from prehistoric skeletons well enough to begin sequencing the genomes of ancient hominids. That’s how they discovered that Neanderthals and Homo sapiens interbred and that most of us have some Neanderthal DNA hiding in our genomes.

But there’s still a great deal that we can’t learn about prehistoric human culture by studying human bones. In a study published on October 30 in Science, a group of geneticists, evolutionary biologists, and archaeologists makes use of another source of data: ancient dog DNA. The study not only helps uncover the evolutionary history of dogs, but also provides some tantalizing clues about prehistoric human culture. It opens a window onto the close, millennia-long relationship between humans and their canine companions.

“It’s another layer to the understanding of human history,” says Anders Bergström, a postdoctoral researcher at the Francis Crick Institute and one of the paper’s lead authors. “We can uncover historical processes between human populations that aren’t necessarily visible in human DNA.”

Dogs make for interesting research subjects, from both a biological and an anthropological perspective. Domesticated dogs, wolves, and dingos all belong to the same species, even though dogs diverged evolutionarily from wolves sometime between 15,000 and 40,000 years ago. Since that split, people have bred dogs into the genetically distinct populations that we call breeds. By comparing dog genomes, we can learn about the human-made processes that produced these populations—and by studying ancient dog genomes, we can work to understand what those processes looked like far into the past.

Ancient DNA research is always a group effort, and this paper is no exception: It has 56 authors. Some of these people are archaeologists—in sites as far apart as Spain and Siberia, they dug up the bones of the 27 dogs examined in this study, who lived between 11,000 and 100 years ago. Other authors are the scientists who labored over these ancient specimens to extract and sequence their genetic material. And still others, like Bergström, analyzed the data for evidence of the evolutionary relationships among these 27 dogs, modern dogs, and wolves.

Studying ancient DNA poses some major methodological challenges. DNA degrades over time, so it’s difficult to be confident in any given A, C, T, or G in an ancient genome. So to analyze the dog genomes, Bergström and his colleagues used F-statistics, which compare the whole genomes of different specimens to one another in order to determine which pairs of animals are more or less closely related. “They tell you, with very few assumptions, that this [specimen] is closer to A than to B,” Bergström says. “And, of course, that doesn’t necessarily tell you what happened in terms of the historical process. But it tells you some basic facts about relationships.” Scientists can then use this relational data to infer what the dog family tree must have looked like. Using this technique, the team was able to determine, for example, that ancient East Asian dogs are, surprisingly, more closely related to ancient European dogs than they are to ancient Middle Eastern dogs.

By building out this tree, Bergström and his colleagues discovered something surprising: 11,000 years ago, before many human groups had adopted agriculture, domesticated dog populations had already formed at least five genetically distinct groups. “At the end of the Ice Age, before any other animal had been domesticated, dogs had already branched out into separate lineages and spread out across the world,” Bergström says. “A lot of these lineages are still represented today, in present-day dogs.” In other words, breeds like today’s Siberian huskies and German shepherds are descended from dog populations that were completely separate by around the year 9000 BC, if not earlier. “We can look at dogs in a park, and we can see the result of this process that started already before any humans had started farming, or before any other animal had been domesticated,” Bergström says.

This information isn’t just interesting from an evolutionary perspective. It also provides a clue about how early humans related to dogs, long before they had formed close relationships with any other species of animal. Laurent Frantz, professor of paleogenomics at the Ludwig Maximilian University of Munich and the study’s other lead author, believes that this discovery provides reason to think that hunter-gatherers may have deliberately bred dogs for certain traits. “Maybe artificial selection was already strong,” he says. “And maybe it was already something conscious that people were doing.”

The fact that separate dog lineages existed 11,000 years ago can’t single-handedly tell us what prehistoric humans thought about their dogs. But Frantz, Bergstöm, and their colleagues were able to probe their data for more clues about the relationship between ancient humans and dogs by comparing the dog family tree they had drawn to similar trees that can be drawn using ancient human DNA. They found that the overall shapes of the family trees were quite similar: For example, ancient humans in East Asia were also more closely related to ancient humans who lived in Europe than those who lived in the Middle East. At first glance, this result might not seem so surprising: If humans took their dogs with them when they migrated, then the most closely related human populations would have the most closely related dogs. And some of the human-dog correspondences can certainly be explained by people bringing their canine companions when they moved from place to place. But for the evolutionarily oldest patterns, there’s a big problem: The human groups in question split over 45,000 years ago, long before dogs were domesticated.

Bergström and Frantz have a theory to explain why dog migration patterns would mimic human ones from thousands of years earlier. Because of climate and geography, perhaps the routes that were hospitable to travel in 45,000 BC were also convenient for trade—of dogs, and of other goods—35,000 years later. “It suggests that there was actually more contact between early human groups than we can see from just the human DNA,” Bergström says. If this theory turns out to be true, ancient dog DNA may have helped us learn about features of prehistoric human culture beyond animal domestication.

But this trade hypothesis is still only speculative, at least until it can be corroborated by other archaeological evidence. “It’s a reasonable hypothesis, but I think it’s worth keeping in mind that the timescales are quite, quite different for the Out of Africa [human migration] and dog domestication,” says Kerstin Lindblad-Toh, a professor of comparative genetics at Uppsala University in Sweden who was not involved in the study. The geographical and climatic factors that contributed to human migrations over 45,000 years ago could very well have shifted substantially by the time that dogs were domesticated.

After that domestication event, some things do seem to have stayed constant. According to the team’s results, after dogs split off from wolves over 11,000 years ago, wolves never made a major reentry into dog populations (until, perhaps, the contemporary craze for wolfdogs). Given that dogs and wolves belong to the same species and produce perfectly healthy offspring, this discovery came as a surprise to the authors. They inferred this result from the observation that some wolves are equally related to all ancient and modern dogs, which indicates that all dogs have the same amount of wolf ancestry. The logical explanation is that wolves didn’t contribute substantially to the dog gene pool after domestication. If, instead, wolves had continued interbreeding with dogs, the team would have expected to observe that all wolves were more closely related to some dogs—which had wolves in their family trees post-domestication—than others, which only had dog ancestors.

But, for some reason, the opposite happened when it comes to the wolf genome: Dogs are universally more related to some wolves than they are to others, which indicates that dogs did in fact contribute genetic material to wolf populations. This asymmetry between dogs and wolves may have a simple explanation: humans. “It shows us,” Lindblad-Toh says, “that probably people held onto their dogs and took good care of them and made sure that they didn't let wolves in.” The wolves had no such guardians.

But Liisa Loog, a postdoctoral researcher in the Genetics Department at the University of Cambridge who was not involved in the study, believes that it is important to keep this result in perspective. She notes that the authors’ argument depends on some specific assumptions about how ancient wolves relate to modern wolves, assumptions that are impossible to confirm without studying ancient wolves directly. “The authors here rely on the assumption that this happened on a now-extinct wolf population that hasn’t been sampled, and that is equally related to all modern-day wolf populations,” she says. “This may be the case, but it also may not be the case.”

This assumption, and the assumptions about geographic and climatic consistency that undergird Bergström and Frantz’s trade hypothesis, do mean that their results and theories can’t be confirmed without additional research, like similar studies of ancient wolf DNA. But, ultimately, 27 dog genomes are a narrow window onto the past: When working with such a small amount of data, assumptions become necessary. “The DNA itself is just DNA,” Bergström says. “It needs that wider context of interpretation.”

The scarcity of evidence, coupled with the difficulty of extracting high-quality DNA from such old bones, might make ancient DNA research seem like a foolhardy endeavor—why not just obtain genetic samples from modern dogs and figure out the family tree from there? But ancient DNA also has some distinct advantages over modern DNA, especially when it comes to dogs. Many contemporary dogs owe their genetic profiles to the Victorian dog breeding craze, so the signatures of their more distant past may be difficult to discern. Looking for evidence about ancient dogs in the genomes of modern ones is like “searching for a needle in a haystack,” Loog says. So it can help to go directly to the source. “Ancient DNA,” Loog says, “literally gives us this time-stamped genetic picture of the past.”

So, while it may be difficult to learn about prehistoric dogs by studying their modern descendants, the special insights afforded by ancient DNA can provide invaluable context for understanding how humans relate to dogs today. “Dogs are kind of unique in that they are a predator, a carnivore. And they were domesticated by hunter-gatherers, way before agriculture, and they were also able to spread so quickly to most groups,” Bergström says. “It’s somehow a surprisingly good fit for the human species to take on this animal as a companion—even though, a priori, it seems like an unlikely candidate for domestication.” If Bergström and his colleagues are right, the human tradition of living with, breeding, and protecting dogs, and of treating canines not just as useful tools but as sources of social connection and emotional support, could have an 11,000-year history. Even before they figured out how to cultivate crops, humans may very well have known how to take care of, and be taken care of by, their animals.