The oldest DNA ever recovered comes from roughly two-million-year-old sediments in northern Greenland, where environmental DNA from an entire lost ecosystem was preserved in frozen clay and quartz.1PubMed Central. A 2-million-year-old ecosystem in Greenland uncovered by environmental DNA For DNA tied to a specific organism rather than scattered in soil, the record stands at about 1.1 million years, pulled from mammoth tooth specimens found in Siberian permafrost.2PubMed Central. Million-year-old DNA sheds light on the genomic history of mammoths Those numbers would have seemed absurd to researchers a generation ago, and pushing beyond them depends on a collision of chemistry, climate, and increasingly clever lab techniques.
Why DNA Falls Apart
DNA is a tough molecule by everyday standards, but geological time is merciless. The main enemy is water. Even at room temperature, water slowly attacks the sugar-phosphate backbone that holds the double helix together, snapping it into shorter and shorter pieces. Laboratory measurements of the chemical bonds that link one nucleotide to the next suggest they resist water attack with a half-life on the order of tens of millions of years at moderate temperature, which sounds reassuring until you realize that other damage pathways work much faster.3PubMed Central. The time required for water attack at the phosphorus atom of simple phosphodiesters and of DNA
The most important of those faster pathways is depurination, in which the chemical bases (the “letters” of the genetic code) pop off the backbone. Once a base is lost, the strand breaks easily at that spot. A study of 158 dated bird bones from New Zealand estimated that under the burial conditions at those sites, the half-life for a short stretch of mitochondrial DNA was roughly 521 years. That means every five centuries, half of the remaining intact copies of that sequence would be broken.4PubMed Central. The half-life of DNA in bone: measuring decay kinetics in 158 dated fossils At that rate, even under ideal conditions, every bond in a DNA molecule would be destroyed in a few million years at best. In warm, wet environments the clock runs much faster.
On top of breakage, surviving fragments accumulate chemical errors. Cytosine bases lose an amino group and start to resemble a different base entirely, a change called deamination. Ancient DNA sequences are riddled with these lesions, especially near the broken ends of fragments. Analysis of Neandertal DNA showed that this type of miscoding error is overwhelmingly concentrated at fragment tips, creating a recognizable damage signature that researchers now use to distinguish genuine ancient sequences from modern contamination.5PubMed Central. Patterns of damage in genomic DNA sequences from a Neandertal The damage is real and cumulative, but its predictability has turned it from a nuisance into a useful authentication tool.
What Keeps DNA Alive for Thousands or Millions of Years
If the chemistry of decay is so relentless, how does any DNA survive long enough to be read? The short answer is that a few environments slow the clock dramatically. Cold is the single biggest factor. Permafrost and polar ice keep temperatures so low that the chemical reactions responsible for breakage and deamination barely proceed. Every record-breaking ancient DNA recovery has come from a cold environment: Siberian permafrost for the mammoth specimens, Greenland’s frozen sediments for the two-million-year-old environmental DNA.
Dryness helps for the same basic reason. Water drives most of the damage chemistry, so removing it slows everything down. Cave systems in arid regions have yielded surprisingly old DNA, though not as old as permafrost finds. Deep-sea sediments are another archive, where cold bottom-water temperatures and low oxygen combine to slow degradation. Ancient eukaryotic DNA has been recovered from abyssal ocean-floor sediments tens of thousands of years old in the South Atlantic.6PubMed Central. Ancient DNA complements microfossil record in deep-sea subsurface sediments
Then there is the mineral effect. Bones and teeth are largely made of hydroxyapatite, a calcium phosphate mineral. DNA binds to hydroxyapatite with a specific chemical affinity, and once bound, it becomes markedly more resistant to both chemical decay and enzymatic digestion by microbes.7PubMed. DNA binding to hydroxyapatite: a potential mechanism for preservation of microbial DNA Experiments using synthetic apatite confirm this: DNA strands that interact with mineral surfaces degrade far more slowly than free-floating DNA in solution.8Applied Surface Science. Adsorption of DNA on biomimetic apatites: Toward the understanding of the role of bone and tooth mineral on the preservation of ancient DNA This is why teeth and the dense inner-ear bone, known as the petrous bone, are the go-to targets for ancient DNA extraction from skeletal remains. The mineral matrix in those tissues essentially locks DNA in place and shields it from the outside world. Extracting that DNA requires dissolving away the mineral first, a demineralization step that is a standard part of the lab workflow.9PubMed Central. Bone Type Selection for Human Molecular Genetic Identification of Skeletal Remains
For plant DNA preserved in lake sediments, the chemistry of the water matters. Studies of hundreds of lake-bottom samples from China and Siberia show that the water’s pH and electrical conductivity are the strongest predictors of how well plant DNA survives in the mud beneath, more important even than temperature or lake depth.10Environmental DNA. Preservation of sedimentary plant DNA is related to lake water chemistry Certain mineral-rich, slightly alkaline conditions seem to stabilize DNA in sediments just as hydroxyapatite does in bone.
The Headline Discoveries
The field’s biggest milestones give a sense of where the limits currently stand. In 2022, a team published the recovery of environmental DNA from the Kap København Formation in North Greenland, dated to about two million years ago. The sequences revealed an open boreal forest of poplar, birch, and thuja trees alongside Arctic shrubs, with DNA from mastodons, reindeer, rodents, geese, and even horseshoe crabs and marine algae. The ecosystem had no modern equivalent anywhere on Earth.1PubMed Central. A 2-million-year-old ecosystem in Greenland uncovered by environmental DNA A separate analysis reconstructed microbial communities from the same formation, identifying bacteria, archaea, and viruses that thrived at the site two million years ago.11bioRxiv. Two-million-year-old microbial communities from the Kap København Formation in North Greenland
For organism-specific DNA from identifiable specimens, the record belongs to mammoth molars from Siberian permafrost. In 2021, researchers reported genome-wide data from three mammoth specimens dating to the Early and Middle Pleistocene, two of which were over a million years old.2PubMed Central. Million-year-old DNA sheds light on the genomic history of mammoths More recently, analysis of ancient microbial DNA from 483 mammoth remains extended that timeline, including sequences from a 1.1-million-year-old steppe mammoth.12PubMed. Ancient host-associated microbes obtained from mammoth remains
For hominins, the deepest reach is more modest. DNA from the Sima de los Huesos site in Spain, where fossils are around 430,000 years old, represents some of the oldest hominin DNA recovered. Researchers extracted a complete mitochondrial genome from an extremely degraded cave bear bone at the same site by developing improved methods to capture ultrashort DNA fragments.13PubMed Central. Complete mitochondrial genome sequence of a Middle Pleistocene cave bear reconstructed from ultrashort DNA fragments That cave’s cool, stable interior helped slow decay, but the fragments were still astonishingly short, many under 40 base pairs.
How Lab Methods Keep Pushing the Boundary
A huge part of the story is not about finding better-preserved specimens but about squeezing more information out of terrible ones. Ancient DNA fragments are short, chemically damaged, and vastly outnumbered by microbial contamination. Twenty years ago, standard lab methods lost most of those fragments before they could ever be read. A wave of methodological advances has changed that.
The most impactful shift was the development of single-stranded DNA library preparation. Traditional methods require double-stranded DNA, which means every ancient fragment that has lost its complementary strand is invisible. Single-stranded library techniques capture both strands and, critically, recover the shortest fragments that older methods discarded. One widely adopted protocol, called ssDNA2.0, uses a specialized ligation strategy to convert single-stranded fragments into readable libraries with less bias and higher efficiency than earlier approaches.14PubMed Central. Single-stranded DNA library preparation from highly degraded DNA using T4 DNA ligase Comparisons show that single-stranded methods consistently convert more molecules into usable libraries than double-stranded approaches, sometimes recovering several times more DNA from the same extract.15Journal of Heredity. A Fast and Efficient Single-stranded Genomic Library Preparation Method Optimized for Ancient DNA
Meanwhile, the rise of sediment DNA has opened a completely different source of ancient genetic material. Instead of extracting DNA from a bone or tooth, researchers take a scoop of cave floor or lake-bed mud and pull DNA from whatever organisms once shed cells there. A single 25,000-year-old sediment sample from a cave in the Republic of Georgia yielded genome-scale data from humans, wolves, and bison without any targeted enrichment, just shotgun sequencing of whatever was in the dirt.16Current Biology. Genome-scale sequencing and analysis of human, wolf, and bison DNA from 25,000-year-old sediment A newer high-throughput method pools multiple sediment extracts together for rapid screening, discarding empty pools and focusing detailed analysis only on those that show a positive signal. The approach has been applied to sites spanning the Middle and Upper Paleolithic across Europe, Asia, and Africa.17Scientific Reports. Maximizing efficiency in sedimentary ancient DNA analysis: a novel extract pooling approach
Telling ancient DNA apart from contamination is its own technical challenge. Modern human DNA is everywhere: on the hands of excavators, on lab surfaces, floating in the air. Software tools now exploit the characteristic damage patterns of ancient DNA to estimate how much modern contamination is present in a given dataset. One tool, AuthentiCT, predicts contamination levels based purely on the deamination patterns expected in genuine ancient sequences from single-stranded libraries.18PubMed Central. AuthentiCT: a model of ancient DNA damage to estimate the proportion of present-day DNA contamination Another, mapDamage, maps sequencing reads against a reference genome and computes the fragmentation and miscoding patterns expected from real ancient sequences.19Bioinformatics. mapDamage: testing for damage patterns in ancient DNA sequences These tools have become standard gatekeepers. If a sequence does not show the right damage profile, it gets flagged as likely contamination, no matter how old the sample is supposed to be.
The Amber Myth and Why Millions of Years Remains Unlikely
In the early 1990s, a wave of spectacular claims suggested that DNA could survive tens or even hundreds of millions of years inside insects trapped in amber. The timing coincided with “Jurassic Park,” and the idea captured the public imagination. But rigorous attempts to reproduce those results failed. When researchers applied strict contamination controls and tried to replicate DNA extraction from amber- and copal-preserved bees and flies, they found no authentic ancient insect DNA at all.20PubMed Central. Problems of reproducibility–does geologically ancient DNA survive in amber-preserved insects? The consensus today is that the original results were contamination artifacts. Despite amber being one of the most promising fossilization environments in theory, it does not preserve DNA over millions of years.
This is a useful reality check. The two-million-year mark for environmental DNA and the roughly one-million-year mark for organism-specific DNA represent genuine, reproducible results from exceptionally cold environments. But the chemistry of decay imposes a hard ceiling. Even under ideal permafrost conditions, the fragmentation rate means that DNA becomes unreadable within a few million years. The half-life estimate from dated bird bones predicts that at a temperate burial temperature, every backbone bond in a DNA strand would be broken well before the ten-million-year mark.4PubMed Central. The half-life of DNA in bone: measuring decay kinetics in 158 dated fossils Colder conditions extend that window, but not infinitely. If you are hoping for dinosaur DNA, the chemistry says no.
When DNA Runs Out, Proteins Pick Up
Proteins are chemically tougher than DNA. Their peptide bonds resist hydrolysis better, and proteins trapped inside tooth enamel are especially durable because the enamel is nearly solid mineral. This means that when DNA is long gone, protein fragments can still be present and informative. A recent study recovered enamel proteins from fossils in Kenya’s Turkana Basin dating to 16 and 18 million years ago, including fragments from an Early Miocene rhinocerotid and several proboscideans.21Nature. Eighteen million years of diverse enamel proteomes from the East African Rift That is nearly ten times deeper than the oldest DNA, and it came from tropical Africa, an environment where DNA survival is essentially zero at those timescales.
Protein sequences carry far less information than full genomes. You cannot reconstruct an organism’s complete genetic blueprint from a few enamel protein fragments. But proteins still contain enough sequence variation to sort out evolutionary relationships, confirm species identifications, and sometimes distinguish closely related lineages. For the deep past, where DNA is physically impossible to recover, ancient proteomics is the only molecular tool available.
Sediment DNA Versus Skeletal DNA
An important distinction that often gets blurred in media coverage is the difference between DNA from identified specimens and environmental DNA from sediments. Skeletal remains give you DNA from a specific individual. You can determine sex, ancestry, kinship to other individuals, and even disease history. Sediment DNA gives you a community snapshot: which species were present at a site, and sometimes rough population-level genetic profiles, but generally not individual-level resolution.
The Greenland two-million-year record is environmental DNA. Nobody pulled a genome from a single mastodon bone at Kap København. Instead, fragments shed by many organisms over time became trapped in mineral grains and frozen in place. The result is an extraordinary catalog of biodiversity but not a genome you could use to clone anything. The mammoth molar studies, by contrast, produced genome-wide data from individual animals, which is why those results were so powerful for tracing mammoth evolutionary history, even though the DNA was “only” about a million years old.2PubMed Central. Million-year-old DNA sheds light on the genomic history of mammoths
Most ancient DNA research still focuses on the last 50,000 years, a window where techniques are mature and skeletal preservation is often good enough for individual-level genomics. Paleogenomic approaches are now reaching into the early Pleistocene, but the deeper you go in time, the more you rely on fragments, environmental sources, and statistical inference rather than clean, long reads from well-preserved bone.22PubMed Central. Deep-time paleogenomics and the limits of DNA survival
Teeth, Dental Calculus, and the Oral Archive
Teeth deserve special mention because they are disproportionately useful for ancient DNA work. Dense enamel and the sheltered pulp chamber inside protect genetic material from the environment. But the real surprise has been dental calculus, the hardite plaque that builds up on tooth surfaces during life. Calculus mineralizes quickly, trapping not just human DNA but also DNA from the bacteria living in the mouth, food particles, and even pathogen genomes. Researchers have used ancient dental calculus to reconstruct oral microbiomes, study ancient diets, and identify specific infections in individuals who lived thousands of years ago.
Because calculus entombs DNA in mineral so rapidly, it sometimes preserves genetic material better than the bone around it. For forensic and archaeological work where soft tissue is long gone, a single tooth can yield human DNA for identification, microbial profiles that tell you about health and diet, and sometimes DNA from animal or plant food sources. It is one of the richest single sources of ancient molecular information available.
Ethics and the Politics of Ancient Remains
Pushing the boundaries of ancient DNA recovery raises questions that are not purely technical. Much of the most informative ancient human DNA comes from Indigenous burial sites, and the communities connected to those remains have increasingly and justifiably demanded a voice in how that research proceeds. A growing body of ethical guidance emphasizes that engagement with descendant and stakeholder communities should be central to the research process, not an afterthought.23PubMed Central. Ethical Guidance in Human Paleogenomics: New and Ongoing Perspectives
One tension point is data sharing. Many genomics guidelines insist on open data, which makes sense for reproducibility but clashes with principles of Indigenous Data Sovereignty, where communities assert the right to control information derived from their ancestors. Critics have argued that framing community involvement in data decisions as somehow unethical conveniently serves researchers who want unrestricted access and ignores the legitimate interests of the people most directly connected to the remains.24PubMed Central. Community partnerships are fundamental to ethical ancient DNA research These are not abstract philosophical debates. They shape which samples get studied, which results get published, and whose history gets told.
Computational Reconstruction and Ancestral Inference
When physical DNA cannot be retrieved, computational methods offer a different kind of time travel. Ancestral sequence reconstruction uses evolutionary models and modern protein or gene sequences to infer what ancestral molecules looked like millions of years ago. Traditional approaches assume that each position in a sequence evolves independently, which misses the reality that mutations at one site can affect what is viable at another. Newer generative models trained on large families of related proteins capture those interdependencies and produce more diverse, more accurate reconstructions of ancestral sequences.25PubMed Central. Reconstruction of Ancestral Protein Sequences Using Autoregressive Generative Models
These reconstructions are not the same as reading preserved DNA. They are statistical best guesses, and they rely on assumptions about how evolution works. But they let researchers synthesize predicted ancient proteins in the lab and test their properties, offering a window into molecular function that predates any surviving physical DNA by orders of magnitude. For the deep evolutionary past, computational reconstruction is not a consolation prize; it is the only game in town.
What Determines Whether Your Specimen Has Anything Left
If you are working with or curious about a specific sample, a few practical factors determine whether DNA is likely to be recoverable:
- Temperature history: Continuous cold preservation is the single strongest predictor. A bone that has been frozen since burial is orders of magnitude more likely to yield DNA than one from a tropical or temperate site, even if the tropical specimen is much younger.
- Tissue type: The petrous bone (the dense bone surrounding the inner ear) and tooth roots consistently outperform other skeletal elements. Dental calculus is a dark horse that can outperform bone in some circumstances.
- Water exposure: Waterlogged specimens in warm environments degrade fast. Dry cave environments or frozen ground are far better.
- Microbial activity: Soil bacteria and fungi digest DNA. Acidic, biologically active soils are the worst-case scenario. Mineral-bound DNA in bone resists microbial enzymes better than free DNA, but heavy microbial colonization can still overwhelm that protection.
- Age: All else being equal, older is worse. But environment matters so much more than calendar age that a 50,000-year-old permafrost specimen routinely yields better DNA than a 500-year-old skeleton from a tropical grave.
For environmental DNA in sediments, the chemistry of the surrounding matrix matters most. Lake-bottom muds with favorable pH and conductivity preserve plant DNA far better than acidic, oligotrophic waters.10Environmental DNA. Preservation of sedimentary plant DNA is related to lake water chemistry In cave sediments, clay-rich layers tend to bind and protect DNA more effectively than sandy ones.
Machine-learning tools are also beginning to help with the analysis end. ARIADNA, for instance, uses a trained algorithm to distinguish genuine ancient mutations from the noise introduced by chemical damage, reducing false positives compared with standard variant-calling software.26PubMed Central. ARIADNA: machine learning method for ancient DNA variant discovery As both extraction chemistry and computational analysis improve in parallel, the practical limit of how far back DNA can be traced will likely keep creeping deeper into the past, though the laws of chemistry guarantee it will never reach the age of dinosaurs.