When he was a doctoral student at the University of Copenhagen, Eske Willerslev could not access the few fossils that might have contained traces of ancient DNA. But in 2000, inspiration struck in an unexpected form: a dog depositing its morning droppings. Willerslev realised that if DNA could persist in animal excrement, it might survive in the wider environment as well. Even after rain and decay, traces of genetic material could remain in soil, offering an entirely new way to study organisms long gone. At the time, his idea was dismissed by professors, but history would prove him right.
In a 2003 study published in Science, Willerslev and colleagues successfully extracted plant and animal DNA from a Siberian permafrost core dating back 400,000 years. Later, in New Zealand, he identified DNA from the extinct moa bird in 600-year-old cave sediments—demonstrating that complex organisms could be identified using sediment alone. Over the past two decades, this technique, known as sedimentary DNA (sedaDNA), has matured into a transformative tool for archaeology and palaeogenetics. Laboratories that once relied on precious fossil specimens are now turning their attention to dirt, revisiting soil samples collected decades ago to uncover new evidence about past life.
The potential of sedaDNA is vast. In 2022, Willerslev’s team extracted fragments of DNA from two-million-year-old permafrost at the northern tip of Greenland, the oldest genetic material recovered to date. “You have humans, you have animals, you have plants, you have the whole bloody ecosystem,” he told Nature. Researchers like Matthias Meyer at the Max Planck Institute for Evolutionary Anthropology in Leipzig describe the field as “scratching the tip of the iceberg,” suggesting that sediments could eventually replace the need for bones at many sites.
Sedimentary DNA has been particularly revelatory in the study of ancient humans. It has illuminated the presence of Neanderthals, Denisovans, and early modern humans in places where no bones were ever found. At Denisova Cave in Siberia, for example, DNA recovered from sediments confirmed that Neanderthals arrived some 170,000 years ago—30,000 years earlier than fossil evidence suggested—and that early modern humans occupied the site around 45,000 years ago. This method also links distinct DNA sequences to particular types of stone tools, potentially identifying the makers of artefacts long before skeletal remains can be found.
The breakthrough in human sedimentary DNA came in 2017, when scientists successfully isolated human sequences from ice-age soils. This discovery opened the floodgates for palaeogenetic research, with teams worldwide eager to apply the technique. To overcome the rarity of human DNA compared with microbial and animal sequences, researchers such as Svante Pääbo’s team at the Max Planck Institute developed “molecular fish hooks” that selectively capture human mitochondrial DNA (mtDNA), which is more abundant in cells and easier to retrieve than nuclear DNA.
Yet the holy grail remains nuclear DNA, which carries far more information about ancestry and population intermixing. Benjamin Vernot and colleagues developed a set of 1.6 million probes to extract nuclear sequences from sediments, targeting Neanderthals, Denisovans, early modern humans, and even unknown hominins. The process was painstaking: even the best samples yielded only a fraction of usable sequences, necessitating months of computational work to interpret. Using this approach in Galería de las Estatuas, a Spanish cave, Vernot could distinguish DNA from single individuals and mixtures of several, revealing population replacements and interactions among Neanderthals over 100,000 years ago.
Other researchers, such as Pere Gelabert, have experimented with “shotgun” sequencing, reading all DNA in a soil sample indiscriminately. While this method can capture a broad spectrum of genetic material, it produces far fewer useful sequences for population studies than targeted capture, making it laborious and inefficient. As a result, nuclear DNA extraction from sediments is likely to remain reserved for exceptional sites, while mitochondrial DNA provides sufficient insights in most cases.
Despite remarkable successes, sedaDNA remains unpredictable. Scientists cannot yet reliably identify which soils will preserve ancient DNA, particularly in hot and humid climates, such as parts of Africa. Yet discoveries continue to defy expectations. In 2022, DNA from deer and hyenas survived in an Israeli cave for 70,000 years despite warm conditions. Researchers are systematically studying soil properties, such as pH, mineral composition, and disturbance history, to improve predictive models and ensure accurate interpretation.
Concerns over contamination and interpretation remain. Soil particles can move over time, and DNA may leach between layers. However, research suggests that once DNA binds to a substrate—whether bone fragments, faeces, or mineral grains—it tends to remain stable. Combined with climate models and archaeological context, these insights give researchers confidence in the age and provenance of DNA recovered from sediments. Notably, Willerslev’s team used sedaDNA to argue that woolly mammoths survived in northern Siberia until roughly 4,000 years ago, challenging previous assumptions about human-driven extinctions and highlighting climate as the likely cause.
Ethical considerations also shape the field. Some Indigenous communities in Australia, for instance, view DNA extraction from sediments as potentially disturbing ancestral remains or culturally significant species. Researchers, including Dawn Lewis of Adelaide University, emphasise early collaboration with communities to navigate sensitivities and intentions. In some cases, dirt is considered a non-invasive substrate; in others, its collection still requires careful consent.
Today, sedimentary DNA is revolutionising our understanding of human history and palaeoecology. As Willerslev notes, the potential to “drop the bones and just go to the dirt” could transform archaeology, extending the fossil record, clarifying the movements of ancient populations, and reconstructing entire ecosystems. But the technique requires meticulous methods, thoughtful interpretation, and respect for ethical and cultural concerns. For a field barely two decades old, Nature observes, the discovery of DNA in dirt is reshaping the study of our past in ways that fossils alone never could.
