In the DNA of the people living today there are small fragments left inherited from ancient human groups of which we do not possess any genome. It’s about populations still without a precise identitythat’s why researchers call them “ghost”. A study published in Science identified these traces in all modern populations analyzed. An unknown first human lineage would have left between 0.5 and 1.1% of the genome and would have interbred with the ancestors ofHomo sapiens before the most recent dispersal out of Africa, therefore over 50,000 years ago. A second, much older signal also emerged among the inhabitants of Oceania, probably arriving through the Denisovans (common ancestor among Sapiens hey Neanderthal). The discovery does not start from a fossil nor from a new sample of prehistoric DNA, but comes from a computer method capable of reconstructing, piece by piece, the genealogical history of modern genomes.
What are “ghost” ancestors and how are they found
In genetics, a ghost ancestor is not necessarily a species completely unknown to paleoanthropology. It’s one ancient population that he has contributed to the DNA of later groupsbut of which we don’t have a genome to use as a comparison. His presence then comes deduced from the traces left in the descendants. When populations that have been separated for a long time meet again and have children, part of their DNA passes from one group to another. This phenomenon is called introgression.
To understand what they were dealing with, the researchers developed TRACEacronym for TRacking Archaic Contributions via ARG Estimation. This system searches archaic traces in modern genomes without having the ancestor’s DNA nor a population without archaic ancestry to use as a comparison. The method reconstructs i Ancestral Recombination Graphs (ARG). Because chromosomes are shuffled through recombination with each generation, each region of the genome can have a different family tree. ARGs reconstruct these trees and estimate when two sequences trace back to a common ancestor, i.e. their time of coalescence.
The DNA coming from a population that has remained isolated for a long time produces very deep genealogical branches. Furthermore, if the crossing occurred later, the segments may still be quite long. These two clues distinguish introgression from the incomplete assortment of evolutionary lines, in which an ancient variant survives without new crosses. In the latter case the fragments are generally shorter, because recombination has had more time to break them.
The TRACE method and ghost traces found in 503 human genomes
TRACE was first verified with simulated genomes of 100 Africans and 100 non-Africans, inserting approximately 2% Neanderthal DNA and a bottleneck linked to exiting Africa. A bottleneck is a temporary reduction in population that causes the loss of some genetic diversity.
With an already known evolutionary history, TRACE achieved 92% accuracy, recovering 71% of true archaic segments and keeping false positives below 0.25%. The results remained robust even when changing the parameters. When the family trees had to be reconstructed, the accuracy remained around 90%, but TRACE found about half of the segments present. The estimates obtained in real genomes are therefore conservative: the method is more likely to lose a trace than to invent it.
The method was applied to the genomes of 503 people from the project 1000 Genomes. The researchers used 15,000 generations as a reference, approximately 420,000 years, preserving only the segments at least 50,000 base pairs long, the “steps” of the DNA double helix.
After identifying them, they compared the archaic traits with three Neanderthal and one Denisovan genomes. Similar ones were attributed to the two groups, while a third set matched neither. Furthermore, more than 90% of the Neanderthal segments and more than 70% of the Denisovan ones coincided with regions already recognized by other methods, confirming the reliability of TRACE.
An unknown ancestor common to Africans and non-Africans
Segments without attribution constituted on average 0.5 to 1.1% of the genome. Most of those found outside Africa also appeared in sub-Saharan populations: the interbreeding would therefore have occurred before the more recent dispersal of Homo sapiens towards Europe and Asia.
These fragments were almost equally similar to Neanderthal and Denisovan DNA, so they may have come from a lineage that separated before the two groups diverged. The genealogical reunion with the modern human line dates back to approximately 830,000 years ago: it is not the date of the crossing, but it indicates how ancient the separation between the populations was.
The ghost regions also had greater heterozygosity, that is, more differences between the two copies of the chromosomes, and shorter segments than the Neanderthal and Denisovan ones. Both of them clues they are compatible with DNA arriving from an isolated population and with an older cross.
The share of 0.5-1.1% concerns the average single genome, but each preserves different fragments. Summing those found in all populations, the tracks covered 71.54% of the analyzable genome. This does not mean that a person has 71% phantom DNA: it is the result of joining fragments present in hundreds of individuals.
1,932 peaks emerged, averaging 61,000 base pairs long. Ghost tracks also appeared in all five “archaic deserts” examined, regions considered devoid of Neanderthal and Denisovan DNA.. These areas may therefore not be exclusively sapiens: natural selection may have eliminated above all the inheritance of the two known groups, leaving that of other lineages.
In Oceania a appears genealogical branches even older
The researchers also analyzed 92 people from Oceania: 25 from Papua New Guinea and 67 from Vanuatu and the Santa Cruz Islands. TRACE estimated on average 0.73% Neanderthal ancestry, 0.66% Denisovan, and 0.33% Ghost.
In the Denisovan segments appeared much deeper genealogical branches than Neanderthal ones. The data suggests that the Denisovans interbred with an even older populationthen transmitting its DNA to the ancestors of the inhabitants of Oceania. In the simulations the signal disappeared by eliminating the super-archaic ancestor and reappeared by inserting this cross. The super-archaic lineage and the modern human lineage would have shared an ancestor about 1.77 million years ago, with an estimate between 1.69 and 1.83 million: much earlier than the ghost lineage identified in other populations.
We still don’t know who they were
Genetic data do not allow us to assign a certain name to these populations. For the phantom ancestor shared by modern humans, the authors cite as possibilities some Middle Pleistocene African human groups or populations attributed to Homo heidelbergensis. For the super-archaic lineage, the separation estimated around 1.8 million years ago would be compatible with Homo erectus. However, they are hypotheses, not identifications.
The possible biological consequences of phantom regions, including those related to immunity and metabolism, have also not yet been clarified.
These ancestors therefore remain “ghosts” not because they are mysterious creatures, but because an ancient genome is still missing with which to compare their traces.
