Lurking deep in your DNA, you may have a ghostly remnant from a “super-archaic” protohuman that isn’t our ancestor. This is because over half a million years ago, “Neandersovans” – the common ancestor of Neanderthals and Denisovans – intermixed with a small-brained, super-archaic hominin. Hundreds of thousands of years later, Denisovans did it again. And after that, Neanderthals and Denisovans intermixed with ancestors of modern humans. Got it?
The ape-ish super-archaic was only distantly related to other Homo species, having split off from the Homo family tree about 2 million years ago. Going by the indirect evidence of the admixture, this creature with a brain the size of an orange evidently survived to grunt the tale for well over 1.5 million years.
These early mixing-and-matchings were deduced by Alan Rogers, Nathan Harris and Alan Achenbach from the University of Utah, based on an innovative analytic technique they call “Legofit” – the results of which were published Thursday in Science Advances.
The mating between the Neandersovans and the mystery super-archaic is the earliest known episode of gene flow in the human line, the authors say. Yet again, we discover that multiple human species coexisted, overlapping in time and space. While the mechanism of the interspecies attraction remains a mystery, it is starting to seem that different human species lost no opportunity to mate with each other. Or maybe they took whatever opportunity beckoned.
Population and group sizes of the super-archaic hominin is believed to have been small: They had migrated from Africa to a radically different environment and were “basically African apes trying to make a living in that environment,” Rogers describes.
Neandersovans had big brains, and so of course did their descendants the Denisovans. Yet if their sex-partner had split from the Homo tree 2 million years ago, wouldn’t it have had a very small brain? “I presume so, but I’ve never seen one,” Rogers laughs in conversation with Haaretz. “Judging by the time frame, they must have been much smaller than ours.”
To get this argument out of the way, are Neanderthals, Denisovans and We in fact separate species? Theoretically, because we interbred and had fertile offspring, we’re all the same species (albeit different types). By that criterion, this 2-million-year-old demi-ape is also the same species, though whether it could have fruitfully interbred with us today is an open question. For the purposes of clarity, we shall call each one a species.
- The last stand of Homo erectus
- Neanderthals migrated to Siberia – twice
- Early humans reached Greek island 200,000 years ago, changing picture of spread from Africa
- ‘Ghost’ ancestor detected in DNA of today’s West Africans
The brain in Spain
Our story starts around 2 million years ago in Africa, when a small-brained super-archaic hominin branched off from the Homo family tree. At least part of that population then migrated to Eurasia: the earliest evidence of hominin presence there, discovered in Dmanisi, Georgia, dates to about 1.9 million years.
Homo erectus georgicus had a small body, and a brain about a third to half the size of ours.
Meanwhile, back in Africa, the ancestors of Neandersovans split from our line about 750,000 years ago, it seems. That is earlier than most other scientists propose. By 700,000 years ago, the large-brained Neandersovans were making their way from Africa to Eurasia – where they met, and mixed with, the small-brained super-archaics.
Augmented with super-archaic DNA, the Neandersovans in Eurasia themselves split about 600,000 years ago – into Neanderthals in the west, and Denisovans in the east.
And apparently much later, perhaps half a million years, the super-archaics and Denisovans then mated too. The timing of this admixture is not clear, Rogers tells Haaretz. It could have been as recent as 50,000 years ago.
Who was this super-archaic? Inquiring minds want to know. Well, they can’t. It seems likely that the super-archaics who mated with the Neandersovans and Denisovans descended from that initial settlement of Eurasia 1.9 million years ago, Rogers tells Haaretz. Homo erectus also likely descended from that initial settlement.
There is, however, at least one other candidate: Homo antecessor, an archaic whose jaw was found in the Atapuerca Mountains, Spain, in the 1990s. It may have lived in England too (stone tools dating from its time, but nothing more, were found at Happisburgh) and France, about 1.2 million to 0.8 million years ago. Very little is known about Homo antecessor and there are hardly any remains, but it seems to feature a mix of archaic and modern characteristics.
Antecessor could have descended from Homo erectus or could have originated in a separate influx from Africa. We have no idea if it’s our ancestor or a side branch that died out.
Since it seems half a million years passed between the admixtures with the Neandersovans and Denisovans, maybe it’s not the same super-archaic? “They seem to be related,” Rogers answers. “I think they’re a related population, if not necessarily the same one.” Indeed, they ought to have evolved and changed somewhat in those half-million years.
So the super-archaics who left a faint signal in some modern people could have been wee Homo erectus georgicus; rather bigger Homo erectus; Homo antecessor, or some other hominin we aren’t aware of at this stage.
In any case, it seems that although super-archaic species, one or more, hung on grimly in Eurasia for well over 1.5 million years, eventually they were replaced by the smarter Denisovans and Neanderthals.
The same would happen to the Denisovans and Neanderthals when anatomically modern humans exited Africa and reached Eurasia. There was interbreeding: modern humans mated with both species, and then replaced them.
Since our ancestors admixed with Neanderthals and Denisovans, could we have faint signals of this new unknown super-archaic? “Yes, we could,” Rogers says – and in fact the team has detected two.
In fact, it’s possible we didn’t even get them from the Neanderthals or Denisovans but admixed directly ourselves with the last remnants of the super-archaic population in Eurasia.
One is the variant of immune system gene called OAS1 found in Melanesians, who have, relatively, very high components of Denisovan DNA.
But wait for it: Genetic statistics (based on length) indicate it was introduced into the Melanesian population in the last 20,000 to 30,000 years – but it’s very different from all the other haplotypes (gene variants) in that area of the genome, according to Rogers. The degree of difference implies that it had been separate from the other haplotypes for 2 million years. It may have reached the Melanesians from Denisovans, but seems to have originated in the super-archaic.
Ditto EPAS1, the famous altitude-tolerance gene found in Tibetans, which has also been said to originate in Denisovans. Indeed, it’s similar to the Denisovan genome, Rogers says. But it’s also wildly different and the separation time is indicated at a million years. “It could have started out in super-archaics and came to us through Denisovans,” Rogers adds.
Finally, parallel research found a super-archaic admixture in today’s Yoruba and Mende in West Africa, apparently about 43,000 years ago (give or take a big margin of error). That African super-archaic split from the human tree before the Neandersovans did, according to that research, and clearly also survived for a long time, though in Africa.
The differences in time and place are vast. But could it be the same archaic that we seem to find irresistible? It could be, or it could be a completely different one, Rogers says. “It will be interesting going forward to figure out if the population that admixed with Africans was the same as the one we are finding evidence of,” he adds.
All this was inferred using a new technique the researchers dubbed Legofit, fed by genetic data from modern Africans and Europeans and from Neanderthals and Denisovans.
The gist of Legofit, Rogers explains to Haaretz, is to ignore the noise of within-group genetic variance, in order to more clearly discern genetic variance between groups. Weeding out within-group variance enables them to reduce the complexity of the model.
“If you only use the between-group components of genetic variation, your model is insulated from everything that happened since the last interaction between the two populations,” he says. Internal variations within have very little information about the deep past anyway, he says: “We don’t lose much by throwing it away. But we gain something in simplicity. We can see farther into the past.”