The oldest DNA ever recovered comes from sediments in northern Greenland, dated to roughly two million years ago. Reported in 2022, these fragments of environmental DNA preserved a snapshot of an entire ecosystem, including plants, insects, and even a mastodon relative, frozen into clay and quartz in the Kap København Formation on Greenland’s northernmost coast. That record shattered the previous champion by about a million years, and it reshaped what scientists thought was physically possible for DNA preservation.
Two Million Years in Greenland’s Permafrost
The Kap København eDNA was not pulled from a single fossil. Instead, researchers extracted it from ancient sediment layers that had been locked in permafrost for the entire duration of their existence. The DNA fragments were tiny and badly degraded, but by matching them against reference databases, the team identified a surprisingly rich community of organisms. The plant list included poplar, birch, and thuja (a type of cedar), alongside reindeer, hares, geese, and horseshoe crabs. Most strikingly, the sediments contained DNA from mastodons, placing these elephant relatives in Greenland for the first time. The ecosystem described by the DNA suggests that northern Greenland two million years ago was a boreal forest, far warmer than today’s polar desert.
1PubMed Central. A 2-million-year-old ecosystem in Greenland uncovered by environmental DNAWhat made this find possible was the mineral matrix the DNA was trapped in. The fragments had adsorbed onto clay and quartz grains in the sediment, and the permanently frozen ground kept temperatures low enough to slow chemical breakdown to a crawl. That combination of mineral binding and extreme cold is now understood to be the key recipe for ultra-long-term DNA survival.
The Oldest DNA from a Single Organism
If you want DNA tied to an identifiable individual rather than a scattering of environmental traces, the record belongs to a set of mammoth molars recovered from permafrost in northeastern Siberia. Two of the three specimens are over a million years old, making them the oldest genomic data ever retrieved from a single organism. The teeth were found in separate deposits, and their DNA told a story researchers did not expect: one of the specimens belonged to a previously unknown mammoth lineage that eventually contributed ancestry to the Columbian mammoth of North America.
2PubMed Central. Million-year-old DNA sheds light on the genomic history of mammothsBefore the mammoth study appeared in 2021, the oldest organism-level DNA on record came from Siberian permafrost sediment cores. Those cores, ranging from about 10,000 to 400,000 years old, contained DNA from at least 19 plant species along with mammoth, bison, and horse. At the time of publication, researchers called this the oldest authenticated ancient DNA then known.
3PubMed. Diverse plant and animal genetic records from Holocene and Pleistocene sedimentsThe Oldest Human DNA
Human DNA does not survive nearly as long as mammoth or plant DNA, mainly because most human remains were not buried in permafrost. The oldest human-lineage DNA recovered so far comes from the Sima de los Huesos cave site in northern Spain, where researchers sequenced an almost complete mitochondrial genome from a hominin fossil dated to over 300,000 years ago. The cave, whose name translates to “pit of bones,” held at least 28 individuals from the Middle Pleistocene. The mitochondrial sequence surprised everyone by aligning more closely with Denisovans, an eastern Eurasian group, than with Neanderthals, complicating what had been a tidier picture of hominin evolution in Europe.
4PubMed. A mitochondrial genome sequence of a hominin from Sima de los HuesosThe Sima de los Huesos site benefits from the cave’s cool, stable temperature and high humidity, conditions that slow DNA breakdown compared to open-air burial. Even so, the DNA was heavily fragmented and required careful authentication before the results could be trusted. More recent ancient human DNA studies have reached back tens of thousands of years for Neanderthals and early modern humans, but the 300,000-year mark from Sima de los Huesos remains the frontier for our own lineage.
Why DNA Can Survive for Millennia but Not Forever
DNA is a fragile molecule under everyday conditions. Water breaks the chemical bonds holding the bases to the backbone, a process called depurination that chops long strands into ever-shorter pieces. Meanwhile, a separate reaction converts cytosine bases into uracil, which the cell’s machinery would normally read as a different letter entirely. In ancient specimens, these cytosine-to-uracil changes pile up near the broken ends of fragments, creating a distinctive damage signature that researchers now use as a fingerprint of genuine antiquity.
5PubMed Central. Ancient DNA damage6PubMed Central. Patterns of damage in genomic DNA sequences from a Neandertal
One influential study estimated that under the conditions found in a New Zealand fossil assemblage (with an effective burial temperature around 13°C), DNA has a half-life of roughly 521 years for a short mitochondrial sequence. That rate is actually about 400 times slower than laboratory experiments on purified DNA would predict, suggesting that the cellular and mineral environment around a bone protects DNA far better than a test tube does. The same study found that nuclear DNA degrades at least twice as fast as mitochondrial DNA, which is one reason mitochondrial sequences dominate the oldest ancient DNA results.
7PubMed Central. The half-life of DNA in bone: measuring decay kinetics in 158 dated fossilsCritically, the half-life number is an average. The biggest variable is not time itself but the conditions during burial. A large meta-analysis of ancient DNA samples found that temperature and precipitation were far stronger predictors of DNA fragmentation than the sample’s age. Samples from hot, wet environments lose their DNA fast; samples from cold, dry, or permanently frozen ground can preserve readable fragments for orders of magnitude longer than the average half-life would suggest.
8Oxford Academic (Nucleic Acids Research). A new model for ancient DNA decay based on paleogenomic meta-analysisHow Minerals Shield DNA from Destruction
One of the reasons bones and teeth are such reliable sources of ancient DNA is the mineral hydroxyapatite, the main inorganic component of both tissues. Laboratory experiments have shown that DNA bound to hydroxyapatite remained detectable after three months of incubation in water, serum, and even in the presence of DNA-degrading enzymes. In contrast, unbound DNA in the same conditions broke down to undetectable levels within three weeks. The mineral essentially wraps around the DNA and blocks the enzymes and water molecules that would otherwise chew it apart.
9PubMed. DNA binding to hydroxyapatite: a potential mechanism for preservation of microbial DNASimilar protective binding occurs with the clay and quartz grains in sediment, which is why the Greenland eDNA record could reach two million years. Experiments using synthetic apatite crystals have confirmed that DNA strands interact strongly with the mineral surface, substantially limiting degradation. This mineral-binding mechanism explains a pattern that puzzled early ancient DNA researchers: sometimes a visually degraded, crumbly bone yields better DNA than a well-preserved one, because what matters is how tightly the mineral locked down the molecules, not how pretty the bone looks on a shelf.
10Applied Surface Science. Adsorption of DNA on biomimetic apatites: Toward the understanding of the role of bone and tooth mineral on the preservation of ancient DNADebunked Claims and the Amber Myth
The 1990s saw a wave of extraordinary announcements: DNA from insects in amber tens of millions of years old, even claims of genetic material from dinosaur-era bones. These reports captured public attention and directly inspired the premise of a famous film franchise. The problem was that none of them held up. Repeated attempts to replicate the amber insect results failed, and careful analysis showed that the supposed ancient sequences were modern contamination. As one review put it, the authenticity of these geologically ancient DNA claims was “questionable on theoretical and empirical grounds,” and no verified DNA from specimens millions of years old existed at the time.
11Trends in Ecology & Evolution. Palaeontology in a molecular world: the search for ancient DNAThe Greenland eDNA result at two million years now pushes into that territory, but with a crucial difference: it was recovered using modern sequencing technology with strict contamination controls, and the damage patterns match what you would expect from genuinely ancient molecules. The debunked 1990s claims relied on older amplification methods that were exquisitely sensitive to contamination and offered no way to distinguish real ancient sequences from modern ones that snuck in during handling.
How Researchers Tell Ancient DNA from Contamination
Contamination is the central anxiety of ancient DNA research. A single skin cell from a lab technician contains more intact DNA than a 50,000-year-old bone fragment. Modern labs address this with ultra-clean rooms, negative controls at every stage, and increasingly sophisticated computational tools that check whether recovered sequences look genuinely old or suspiciously fresh.
The key insight is that ancient DNA degrades in predictable ways. Those cytosine-to-uracil changes described earlier cluster at the ends of fragments, creating a characteristic damage profile that modern contaminant DNA lacks. Software tools can scan millions of sequencing reads and flag which ones carry this telltale pattern, essentially sorting the real ancient molecules from the modern interlopers.
12PubMed. mapDamage: testing for damage patterns in ancient DNA sequencesNewer tools extend this approach to work with environmental and microbial sequences, where there is no single reference genome to map against. These programs evaluate each assembled sequence for the expected damage gradient and assign a statistical confidence score, letting researchers filter large datasets of sediment DNA without manually inspecting every fragment.
13PubMed Central. PyDamage: automated ancient damage identification and estimation for contigs in ancient DNA de novo assemblyThe Library Methods That Made It Possible
Ancient DNA fragments are typically extremely short, often under 50 base pairs. Standard methods for preparing DNA for sequencing were designed for modern samples with much longer fragments and routinely lost the shortest, most damaged molecules. A breakthrough came with single-stranded library preparation, which works with each individual strand of a DNA fragment separately rather than requiring both strands to be intact. This approach captures fragments that double-stranded methods would miss entirely, dramatically increasing the yield from ancient and degraded samples.
14PubMed. A Method for Single-Stranded Ancient DNA Library PreparationRefinements to this technique have made it faster and more efficient. One streamlined protocol converts single-stranded DNA into sequencing-ready libraries in a single enzymatic reaction, cutting both the hands-on time and the opportunities for contamination. These methods were essential for the Greenland and mammoth studies; without them, the fragments would have been too short and too damaged to sequence at all.
15PubMed Central. A Fast and Efficient Single-stranded Genomic Library Preparation Method Optimized for Ancient DNAEnvironmental DNA and What Sediments Preserve
The Greenland record is part of a broader shift in the field toward sediment-derived environmental DNA. Instead of relying on finding a well-preserved bone or tooth, researchers extract DNA directly from layers of dirt, clay, or marine sediment. The advantage is enormous: you do not need a visible fossil. Organisms shed DNA into their environment constantly through skin cells, feces, pollen, root exudates, and decomposition. That DNA binds to mineral grains and, under the right conditions, persists for thousands or even millions of years.
Marine sediments have also yielded ancient DNA, though the preservation window appears shorter than in frozen terrestrial deposits. One study of eastern Mediterranean seafloor cores found that specific microalgal DNA was concentrated in the youngest layers (around 9,000 years old), while older layers dating to 80,000–125,000 years ago retained DNA mainly from fungi and terrestrial plants that had washed in via river systems. The lipid biomarkers of the original marine algae were still present in the older layers, but their DNA was largely gone.
16PubMed Central. Preservation potential of ancient plankton DNA in Pleistocene marine sedimentsThis contrast between terrestrial permafrost and marine settings underscores the point that preservation depends overwhelmingly on local conditions. Permafrost in Siberia and Greenland provides the perfect combination of cold, stable temperatures and mineral binding. Tropical soils, by contrast, are essentially DNA graveyards: high temperatures, abundant water, and thriving microbial communities destroy genetic material within years or decades.
Dental Calculus as a Surprising DNA Archive
Bones are the classic source of ancient DNA, but one of the richest and most unexpected archives is dental calculus, the hardite deposit that builds up on teeth during life. Calculus is essentially fossilized dental plaque: a mineralized biofilm that traps bacteria, food particles, and host cells in a hydroxyapatite matrix. Because the mineral seals the biological material inside during the person’s lifetime, the DNA is protected from post-mortem contamination in ways that exposed bone surfaces are not.
17PubMed. Ancient DNA analysis of dental calculusSequencing ancient calculus has opened a window into the history of the human oral microbiome, tracking how our mouth bacteria changed across major cultural transitions like the adoption of farming and the industrial revolution. It has also provided dietary information, with plant and animal DNA from food sometimes preserved alongside the microbial community. For researchers interested in health and disease rather than species identification, dental calculus has become one of the most information-dense materials in the archaeological record.
How Far Back Could DNA Theoretically Survive
The two-million-year Greenland result sits uncomfortably close to what many researchers had assumed was a hard physical limit. Under permafrost conditions with stable temperatures near or below freezing, DNA fragments get shorter and more damaged with time, but they do not vanish at a single predictable cutoff. The fragmentation rate depends so heavily on local temperature, moisture, and mineral context that a single “expiration date” for DNA does not exist.
The meta-analysis finding that age alone is not a significant predictor of fragmentation, once you account for environmental conditions, is the most counterintuitive result in the field. It means that a 500,000-year-old sample from deep permafrost could yield better DNA than a 5,000-year-old sample from a temperate cave. It also means that the theoretical ceiling for DNA recovery is not a fixed number of years but depends entirely on where the sample spent those years.
8Oxford Academic (Nucleic Acids Research). A new model for ancient DNA decay based on paleogenomic meta-analysisWhether DNA could survive beyond two million years under ideal conditions remains an open question. Greenland’s Kap København Formation has been continuously frozen for essentially the entire period since deposition, and the fragments recovered were right at the edge of what current technology can read. Colder, more stable environments might push the boundary further, but candidates are rare: deep Antarctic ice, perhaps, or subsurface permafrost on other planets. For Mars exploration, the Greenland result is directly relevant. If DNA-like molecules ever existed on Mars, the cold, dry conditions might have preserved them for geological timescales, and the sequencing methods developed for ancient Earth DNA could in principle be adapted to look for them. Whether that search would find anything is another question entirely, but the Greenland work proves that the chemistry of preservation can outlast even the most pessimistic theoretical predictions.