Somewhere between 2% and 19% of the genetic ancestry carried by present-day West African populations traces back to an archaic hominin group that split from the modern human line before Neanderthals and Denisovans ever diverged. No skull, no tooth, no fragment of bone has been found to represent this population. The only proof it existed lives inside the genomes of people alive right now, making it one of the most striking “ghost” lineages in human evolutionary science.
Why a fossil-free human lineage rewrites assumptions about Africa
For decades, the story of ancient interbreeding centered on Eurasia. Neanderthal DNA in European and Asian populations, and Denisovan DNA in Oceanian and Southeast Asian groups, dominated the conversation because scientists had bones and cave sediments to work with. Africa, where modern humans originated, was largely treated as a genetically simpler picture. That assumption collapsed when researchers modeled site-frequency spectra across West African genomes and found a substantial archaic signal from a population with no known fossil match, as described in a Science Advances analysis. The estimated contribution, ranging from roughly 2% to 19%, rivals or exceeds the proportion of Neanderthal ancestry found in many non-African populations.
The tension is straightforward: Africa’s fossil record and ancient DNA preservation are far sparser than Eurasia’s, largely because tropical climates degrade organic material faster. That means genetics may be the only tool capable of detecting entire branches of the human family tree that lived, reproduced, and disappeared on the continent without leaving physical traces. If one ghost lineage has already surfaced through statistical modeling alone, others could be hiding in populations that have not yet been densely sampled, particularly across Central and East Africa.
This realization also challenges older narratives that framed African populations as a single, panmictic source for all modern humans. Instead, the emerging picture suggests a continent-wide patchwork of semi-isolated groups that sometimes merged and sometimes remained separate for long stretches of time. In that framework, archaic lineages in Africa are not anomalies; they are expected outcomes of a long, complex demographic history.
How statisticians found a population that left no bones
The West African signal was not detected by comparing modern DNA against an archaic reference genome, because no such genome exists for this lineage. Instead, researchers used computational models that look for unusual patterns in the frequency and distribution of genetic variants across living populations. When certain stretches of DNA look too old or too divergent to have come from the known modern human family tree, they stand out as candidates for archaic introgression.
A key advance came from methods that do not require a fossil genome at all. One such tool, introduced in a methods-focused study, is ArchIE, a reference-free statistical approach designed to identify introgressed archaic segments without needing a sequenced archaic genome for comparison. It works by training on simulated genetic data under different demographic scenarios, learning what introgressed segments should look like in terms of length, divergence, and surrounding variation. The trained model then scans real genomes for segments that match those expectations.
In practice, researchers combine such machine-learning approaches with more classical population genetic tools. They examine the site-frequency spectrum, which summarizes how common different genetic variants are in a population, and test whether observed patterns can be explained by known population splits and migrations. When no reasonable model fits the data unless an unknown archaic contributor is added, the ghost lineage becomes the most parsimonious explanation.
The Neanderthal case offers a useful benchmark for what these methods can reveal. By aggregating the fragments of Neanderthal DNA scattered across millions of modern genomes, researchers have reconstructed a substantial portion of the Neanderthal genome itself, even though direct fossil sampling remains limited. That work showed that living people collectively preserve large portions of an archaic genome. The same logic applies to the West African ghost lineage: even without a single bone, the genetic footprint left behind in millions of living people can reveal the rough shape, timing, and scale of an entire population’s contribution.
Ghost lineages beyond West Africa
The West African signal is not the only ghost in modern human DNA. Analyses of Eurasian and Oceanian genomes indicate that some archaic segments cannot be explained by known Neanderthal or Denisovan contributions. In Asia and Oceania, for instance, approximate Bayesian computation paired with deep learning has uncovered patterns consistent with at least one additional archaic source. This implies that populations in multiple parts of the world carry genetic material from archaic groups that remain unidentified in the fossil record.
An even deeper layer exists in Eurasia. A study in Science Advances concluded that the ancestors of Neanderthals and Denisovans themselves interbred roughly 700,000 years ago with a “superarchaic” population that split from other humans about 2 million years ago, based on demographic modeling of genomic data from archaic and modern humans. The authors inferred this ancient contact by showing that the observed genetic relationships among Neanderthals, Denisovans, and modern humans could not be explained without invoking gene flow from a very divergent group, as detailed in their superarchaic admixture study.
No fossil has been definitively assigned to that superarchaic group either, placing it in the same ghost category as the West African lineage, though at an even greater time depth. It might correspond to one or more early Eurasian hominin populations known only from fragmentary remains, but without ancient DNA from those fossils, the match remains speculative.
The Denisovan story provides a template for how ghost lineages can eventually gain physical evidence. Denisovans were first identified from mitochondrial DNA extracted from a tiny finger bone in Siberia’s Denisova Cave, long before researchers could say much about their appearance or geographic range. Years later, a mandible from the Tibetan Plateau’s Baishiya Karst Cave was linked to Denisovans through protein analysis, extending their known distribution by thousands of kilometers. That progression, from a single genetic signal to scattered fossils across vast distances, shows that today’s ghost lineages could become tomorrow’s named species if the right excavation happens in the right place.
Open questions about archaic DNA still shaping living people
Several gaps remain. The 2% to 19% range for the West African ghost lineage is wide, reflecting genuine uncertainty in the statistical models rather than a single neat value. Different assumptions about population sizes, migration rates, and the timing of admixture can all change the inferred contribution. More extensive sampling of African populations, particularly in regions that remain underrepresented in genomic datasets, will be crucial for narrowing those estimates.
Another open question is functional: what, if anything, are these archaic segments doing in living people? In non-African populations, Neanderthal and Denisovan DNA has been linked to traits ranging from immune responses to altitude adaptation. It is plausible that some ghost-lineage variants in West Africa and elsewhere also influence physiology, disease susceptibility, or environmental adaptation. Identifying such effects will require careful association studies that can distinguish archaic ancestry from other sources of genetic variation and from social and environmental factors.
There is also the issue of taxonomy and language. Calling these contributors “ghost lineages” emphasizes our ignorance, but it does not resolve whether they should be classified as separate species, subspecies, or simply deeply divergent populations of Homo sapiens. Genomic data alone may not settle that debate, because species definitions hinge not just on genetic distance but also on morphology, behavior, and reproductive isolation-traits that are hard to reconstruct from DNA fragments embedded in modern genomes.
Finally, the growing list of ghost lineages forces a rethinking of how we narrate human origins. Rather than a single, clean branching tree, the picture looks more like a braided river, with channels that split, rejoin, and sometimes vanish. Modern humans are the product of that braided history, carrying within them traces of populations that left no bones behind. As genomic technologies and statistical tools improve, those invisible ancestors are becoming increasingly visible, reshaping our understanding of who “we” are and how many different kinds of humans once walked the Earth.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
