CeleryKills

Even If We Hike Into the Hoh Rain Forest With a Syringe

I love that Jurassic Park is practically a regional documentary here: ferns, fog, and an optimistic entrepreneur who definitely should have listened to his paleontologists. But if you hand me an amber nugget at a roadside rock shop and ask whether there’s dino DNA inside waiting to boot, the answer is still a friendly no. The short version is chemistry wins. The long version is much more interesting.

The cleanest measurement we have for DNA survival comes from a moa graveyard in New Zealand, where researchers analyzed mitochondrial DNA from 158 radiocarbon‑dated leg bones and fit the decay to an exponential curve. They reported an average half‑life of 521 years for a 242‑base‑pair fragment, and warned that even under cool burial the sequence would become unreadable on million‑year timescales. At an effective burial temperature around 13 °C, total bond breakage would arrive by about 6.8 million years, with sequence legibility lost long before that, around 1.5 million years. That is decisive for dinosaurs that vanished more than 65 million years ago (Allentoft et al., Proc. R. Soc. B, 2012; Kaplan, Nature news, 2012).

If I sound a little delighted by the half‑life number, it is because it explains so much in one stroke. A half‑life of 521 years means that every five centuries about half of the remaining intact backbone bonds pop, and the fragments get shorter and junkier. Nuclear DNA fares even worse than mitochondrial DNA in the same bones, degrading at least twice as fast, which further shrinks the odds of reconstructing full genomes from deep time (Allentoft et al., Proc. R. Soc. B, 2012).

The moa study also offered a reality check about the messiness of the world. Only about 38.6 percent of variation in DNA preservation could be explained by age. The rest depended on local taphonomy, burial chemistry, moisture pulses, and microbial tunneling through bone. Think of age as the metronome and environment as the jazz solo. This is why two bones of the same age can yield very different aDNA outcomes (Allentoft et al., Proc. R. Soc. B, 2012; Emmons et al., PLOS ONE, 2020).

Temperature and water are the big dials. Hydrolysis keeps clipping the sugar‑phosphate backbone, and the rate climbs as heat and liquid water rise. Microbes bring enzymes and acids, introduce tunnels through mineral matrices, and solubilize phosphate, which speeds skeletal and DNA deterioration. In buried remains, the bone microbiome does not even match the surrounding soil cleanly. It reflects a shifting mix of soil colonizers, gut taxa in anoxic pockets, and diagenetic chemistry that vary across a single skeleton. The result is a preservation lottery that age alone cannot predict. I have handled bones that looked perfect and yielded nothing, and crumbly pieces that surprised us (Emmons et al., mSystems, 2025; Emmons et al., PLOS ONE, 2020; Rollo et al., Ancient Biomolecules, 2002).

So how old is “too old” in practice? Permafrost can push the clock back. The oldest genomic DNA from physical remains so far sits a little past a million years in Siberian mammoth teeth, a triumph of cold storage and careful assembly from tiny fragments. Environmental DNA bound to minerals in Greenlandic sediments has stretched to roughly two million years, which is astonishing, but those are microscopic snippets used to reconstruct ecosystems, not a passport to resurrect an animal. Both results depend on deep cold and unique binding conditions that slow decay and shield the fragments, and even then the sequences are shreds that require heroic bioinformatics to interpret (Callaway, Nature, 2021; The Conversation explainer, 2022; ScienceDaily summary, 2022; New Scientist, 2022).

What about the amber fantasy that has lured many of us into science? Multiple independent studies have failed to replicate the early 1990s claims of Mesozoic DNA in amber entombments. Using modern high‑sensitivity methods, researchers have shown that copal and amber are poor DNA vaults. Resin is permeable, water works its way in, contaminants are common, and even in sub‑fossil copal only years old, obtaining authentic DNA is hard. The upshot is that ancient DNA from amber is not considered viable, and reports that suggested otherwise are now classic examples of contamination. It stings to say goodbye to that mosquito, but the data are the data (Penney et al., PLOS ONE, 2013; ScienceDaily, 2013; SYFY explainer, 2020; New Atlas, 2020).

Here is the part that gets misreported. The chemical doom for DNA does not imply that nothing organic survives in deep time. Some biomolecules and structures can persist as altered remnants, or they can leave chemical fingerprints that are diagnosable. Mary Schweitzer and colleagues famously demineralized a Tyrannosaurus rex femur and a hadrosaur bone and reported pliable vessel‑like structures and collagen peptide signatures using mass spectrometry. Those results ignited debates about chemistry, contamination, and mechanisms of preservation, including hypotheses about iron-mediated crosslinking. The field has since added spectroscopy and immunological tools to chase “deep‑time organics.” Even recent work has reported hemoglobin‑like signals in vessel‑shaped voids with resonance Raman methods, while others caution that microbial biofilms and diagenetic iron can mimic biology. The headline is not “soft tissues survive like fresh steak.” The headline is that under rare geochemical regimes, trace proteins or protein fragments and original structures can be stabilized, recognizable, or templated, even if DNA is catastrophically shattered (Schweitzer et al., Science, 2005; Phys.org summary of collagen sequencing, 2007; NCSE commentary on skepticism and dating, 2005; C&EN news on hemoglobin signals, 2025).

I sometimes get asked, usually over a cup of Olympia coffee, whether these soft‑tissue reports mean DNA might hide nearby. It is tempting to hope. But collagen is a ropey structural protein that crosslinks and can mineral‑template, while DNA is a fragile, water‑sensitive information polymer that snaps, depurinates, and deaminates. Those are different chemical fates. The half‑life math still applies, and the most optimistic survival calculations under subzero conditions give you fragments suitable for ecosystem reconstruction or limited phylogenetic questions, not a genome worth cloning. The mammoth and Greenland records show the upper bound for what cold can do. They also prove that age plus environment sets a hard limit on informational content (Callaway, Nature, 2021; New Scientist, 2012 report on moa half‑life implications).

Could we ever get lucky with a dinosaur in some perfect chemical cocoon? I would never say never in science, but luck would have to be extraordinary. You would need uninterrupted cryostorage for tens of millions of years, no thaw pulses, mineral binding that outcompetes hydrolysis, zero microbial colonization, and a time machine to keep volcanoes, groundwater, and plate tectonics out of the way. Even then, the 521‑year half‑life number follows you like a very patient raincloud. I stake my field notebook that we will keep finding fascinating molecules, crosslinks, and mineral ghosts, and that those will tell us about physiology, color, phylogeny, and diagenesis. I will also stake my boots that we will not be growing a hadrosaur in a lab from DNA retrieved out of amber or bone (Allentoft et al., Proc. R. Soc. B, 2012; Penney et al., PLOS ONE, 2013).

If you want a hopeful ending, here it is. The same chemistry that dooms dinosaur DNA is exactly what makes the ancient DNA we can recover so precious. It is fragile, local, and biased, which forces us to be careful, humble, and inventive. That mindset has already given us Pleistocene genomes, million‑year mammoth histories, and a two‑million‑year snapshot of an Arctic forest in Greenland. It has also taught us that oxygen, water, heat, and microbes are not villains. They are the rest of nature, doing what it always does in the Salish air. The trick is to meet it where it cooperates, not where we wish it would.

References