As we advance what we know about where we came from as a species, we slowly open up new sights and vistas looking back into our human story. We’ve also been using the conveniences brought about by modern science to get as complete a picture as we can achieve.
We’ve got DNA analysis, which shows instructions on how to make you and your body as it was passed down through centuries and generations, like when we used just that to identify Japanese ancestry and the origins of the famous and ancient Etruscans.
However, technology is increasingly becoming more competent in doing the analysis that we humans couldn’t hope to achieve in a reasonable amount of time. This case appears to be all the more evident with findings published in the journal Nature Communications back in 2019, where scientists were able to determine that we may have a “ghost” lineage, from an as-of-yet unknown species of human ancestor, present within our genome—and they found it with the help of artificial intelligence.
Said author and Universitat Pompeu Fabra evolutionary biologist Jaume Bertranpetit, ancient humans interbred with other extinct hominids as our species made an exodus out of Africa “about 80,000 years ago, […] when part of the human population, which already consisted of modern humans, abandoned the African continent and migrated to other continents, giving rise to all the current populations.”
It was thought that our species only made contact—and interbred—with Neanderthals and Denisovans, whose genomic heritage is still visible in modern humans today. However, the study by Bertranpetit and co-authors found a third “introgression” in our lineage from Eurasian DNA samples.
They found this “ghost” lineage by employing the help of deep learning algorithms and a statistical inference technique called Bayesian inference. The authors believe the “ghost” lineage to be either “related to the Neanderthal-Denisova clade” or a lineage which “diverged early from the Denisova lineage.”
Temple University evolutionary genomics expert Joshua Schraiber described the deep learning method used in the study as “fitting a more complicated shaped thing to a set of points in a bigger space.” “[You’re] fitting some squiggly thing to a set of points in [a] much bigger, thousand-dimensional space. Deep learning says, ‘I don’t know what squiggly shape should fit to these points, but let’s see what happens.’”
The team trained the algorithm using eight (8) known models of early Eurasian human evolution. From there, it projected how evolution would progress in such a way that it maintains the Out-of-Africa aspect of the theory, as well as reach the current population diversity that Eurasia has today. In doing so, the team found evidence pointing towards the new “ghost” lineage.
“Our theory coincides with the hybrid specimen discovered recently in Denisova, although as yet we cannot rule out other possibilities”, said by co-author and University of Tartu genomicist Mayukh Mondal in a press statement.
Bertranpetit and team know that their research methodology is basically at its infancy, and results from using these technical methods are pretty few and far between. However, the team believes they are on to something pretty remarkable in their findings.
In Bertranpetit’s statement to the Smithsonian Magazine: “We thought we’d try to find these places of high divergence in the genome, see which are Neanderthal and which are Denisovan, and then see whether these explain the whole picture. As it happens, if you subtract the Neanderthal and Denisovan parts, there is still something in the genome that is highly divergent.”
References
- Dockrill, P. (2021, October 25). Artificial intelligence has found an unknown “ghost” ancestor in the human genome. ScienceAlert. https://www.sciencealert.com/artificial-intelligence-finds-an-unknown-ghost-ancestor-in-the-human-genome
- Handwerk, B. (2019, February 8). Artificial intelligence study of human genome finds unknown human ancestor. Smithsonian Magazine. https://www.smithsonianmag.com/science-nature/artificial-intelligence-study-human-genome-finds-unknown-human-ancestor-species-180971436/
- Interbreeding. (n.d.). The Smithsonian Institution’s Human Origins Program. Retrieved October 25, 2021, from http://humanorigins.si.edu/evidence/genetics/ancient-dna-and-neanderthals/interbreeding
- Mondal, M., Bertranpetit, J., & Lao, O. (2019). Approximate Bayesian computation with deep learning supports a third archaic introgression in Asia and Oceania. Nature Communications, 10(1), 246. https://doi.org/10.1038/s41467-018-08089-7
