What does ghost lineage mean in human evolution?

A ghost lineage is a population whose existence is inferred from genetic signals in living people's DNA, but that has never been documented in the fossil or ancient-DNA record. The term describes the state of the evidence, not the population itself.

How did scientists find ghost ancestors without ancient DNA?

A method called TRACE reconstructs ancestral recombination graphs — diagrams showing which stretches of DNA trace back to which shared ancestor — across hundreds of living people's genomes. Segments with unusually ancient or divergent ancestry point to interbreeding with populations no longer in the fossil record.

Could there be more unknown human ancestors hidden in our DNA?

The Berkeley team expects more ghost lineages to be found as genome databases expand, especially from underrepresented populations in Africa and island Southeast Asia. The computational method scales with data availability and has so far been limited by the size of existing genome archives.

Human DNA hides two ghost ancestors — one from a species 1.8 million years old

Q: How much archaic DNA do modern humans carry?

Modern humans carry approximately 2% of their genome from archaic populations, including Neanderthal (~1% in non-Africans), Denisovan (up to 4% in Oceanian populations), and the two newly identified ghost lineages (~0.5-1% from each). This archaic DNA concentrates in genes linked to immune function and metabolism.

Every modern human carries roughly 1.5% of their genome from two populations of hominins that left no known fossils, have no formally assigned species name, and have never had their ancient DNA sequenced. Both populations mated with our ancestors at different moments in prehistory, in different parts of the world, before vanishing entirely — leaving only their genes behind.

Detecting them required no excavation. A computational method called TRACE — TRacking Archaic Contributions via ARG Estimation — developed at the University of California, Berkeley, reconstructs the genealogical trees embedded within the genomes of living people and can detect the signatures of populations that disappeared from the fossil record before modern archaeology could reach them. The technique, reported in Science, does not require ancient DNA.

The paper’s first author, Berkeley graduate student Yulin Zhang, explained: “We were actually able to find and map genomic locations in modern humans that are from this ghost lineage and show that this ghost ancestry is in all modern humans, not only in Africans.”

How TRACE reconstructs extinct family trees from living genomes

The logic behind TRACE is that any two people who share an identical segment of DNA are descended from a common ancestor. The older that ancestor, the longer and more distinct the genealogical branch that records them. Modern statistical algorithms can map these ancestral recombination graphs — diagrams of which stretches of chromosome trace back to which shared ancestor — across a full genome.

The Berkeley team, led by Zhang and associate professor Priya Moorjani, analyzed more than 500 present-day human genomes from around the world. They identified segments where the inferred ancestral lineage was so divergent from the main human family tree that it could not have originated within any known modern human population. Those segments point to interbreeding with an archaic population genetically distinct enough to register as an outlier across hundreds of genealogical branches.

The method was validated before the new search began. TRACE correctly recovered Neanderthal DNA at approximately 1% of non-African genomes — a proportion established independently by ancient-DNA analysis. It also correctly identified Denisovan ancestry at up to 4% in genomes from Oceanian populations. Having confirmed both known signals, the team searched for anything else.

Two lineages, two different histories

The analysis found two distinct ghost populations, each with its own evolutionary timeline.

The first is an African lineage. Its divergence from the ancestors of modern humans dates to approximately 800,000 years ago — the same broad period when Neanderthals and Denisovans branched off — but it interbred with our species at some point before 50,000 years ago, before the large-scale human migration out of Africa. Every person in the study, African and non-African alike, carries between 0.5% and 1% of their genome from this lineage. Smithsonian coverage noted a possible connection to Homo heidelbergensis, a hominin present across Africa and Europe between roughly 200,000 and 700,000 years ago, but the researchers offer this only as speculation: no genetic sequence from that population has been recovered to confirm it.

The second lineage is older and reached modern humans indirectly. Its origin traces back 1.8 million years. This super-archaic population did not interbreed directly with Homo sapiens; it interbred with Denisovans somewhere in Eurasia more than 200,000 years ago. When Denisovans later mixed with the ancestors of modern humans — primarily in southeast Asia and Oceania — fragments of the super-archaic genome came along for the second transfer. Today it is most visible in Oceanian populations. Arjun Biddanda, a Johns Hopkins co-author, described the finding as “particularly exciting because it reveals genetic contributions from a human lineage that lived over a million years ago, despite the absence of any sequenced DNA from that population.” He and the team cautiously point to Homo erectus as a candidate, though no ancient DNA has ever been recovered from a specimen old enough to confirm it.

Together, the two lineages account for approximately 2% of the modern human genome — comparable in scale to the Neanderthal contribution. Much of that archaic DNA concentrates in genes linked to immune function and metabolic pathways, suggesting the introgressed sequences helped ancient modern humans adapt to new pathogens and food sources encountered during migration.

What this leaves unanswered

The two populations are inferences, not identifications. TRACE detects that ancient DNA entered the modern human genome from a source that diverged at a specific moment in evolutionary time. It cannot describe what that source looked like, how it behaved, or how it relates taxonomically to named hominins. No skull, no tool assemblage, and no protein sequence has been definitively matched to either ghost.

The African lineage is especially puzzling. An 800,000-year-old divergence date puts it in the same temporal bracket as Neanderthals and Denisovans, yet it left no usable fossil DNA anywhere in the African record. One reason is taphonomic — the chemistry that degrades organic molecules works faster in tropical Africa than in the cool caves of Siberia and Croatia where Neanderthal and Denisovan DNA survived. Another is that the African fossil record for this period is fragmented and largely undescribed at the molecular level. A third is simply that TRACE’s date estimates carry uncertainty windows: the real divergence could fall within a range of tens of thousands of years.

The 1% contribution from the African ghost is also approximately equal to the Neanderthal signal in people outside Africa, which raises a direct question the paper does not answer: why have earlier genome-analysis tools missed it? The authors argue that previous methods required ancient DNA or lacked statistical power to distinguish deep-ancestry signals from background genomic variation. Reconstructing genealogical trees appears more sensitive than the population-level statistics used before. Whether that explanation fully accounts for the gap is an open question that independent replication will need to address.

Moorjani characterized what the TRACE approach ultimately amounts to: “These new computational methods that allow us to reconstruct genealogical relationships are really the next frontier in this field because they are allowing us to uncover hidden episodes from our past without requiring ancient DNA.”

Common questions about ghost lineages and ancient human ancestry

What does “ghost lineage” actually mean? A ghost lineage is a standard evolutionary term for a population whose existence is inferred from indirect evidence — in this case genetic signals — but has never been documented in the fossil or ancient-DNA record. The word describes the state of the evidence, not the population itself.

How much archaic DNA do modern humans carry in total? Across all confirmed and newly discovered sources — Neanderthal, Denisovan, and the two ghost lineages — modern humans carry approximately 2% of their genome from archaic populations. That proportion is not spread evenly: archaic segments appear most often in regions governing immune responses and metabolic function, consistent with adaptive advantages gained from interbreeding.

Could more ghost lineages be waiting to be found? The Berkeley team expects so. Expanding the genome database — particularly from African populations and island Southeast Asian groups currently underrepresented — and adding more Denisovan reference sequences would likely expose signals from additional archaic sources. The computational approach scales directly with the volume of genomic data available.

The next step for the Berkeley group is applying TRACE to larger and more geographically diverse datasets, with particular emphasis on regions in sub-Saharan Africa and the Pacific. Moorjani also pointed to recent work extracting ancient proteins from Homo erectus fossils as a potential bridge: proteins survive in bone far longer than DNA, and matching those sequences to the super-archaic genomic signal might eventually attach a name to the second ghost population. The Science paper was published July 30, 2026.

Reference: Zhang et al., “Identifying archaic introgression using ancestral recombination graphs,” Science, 2026. DOI: 10.1126/science.aef8874

Tags: archaeology, ancient DNA, human evolution, UC Berkeley, Homo erectus, Priya Moorjani