How DNA forensics is transforming studies of ancient manuscripts

In May 2006, Tim Stinson travelled to England to tour the libraries of London, Oxford and Cambridge. At the time, he was editing a fourteenth-century poem for his PhD at the University of Virginia in Charlottesville, and after months of poring over grainy microfilm copies, he was eager to get his hands on an original. During a visit to Oxford’s Bodleian Libraries — a place so magical that scenes from the Harry Potter films were shot there — he was finally handed one of the manuscripts he had travelled all that way to see. But he found himself so riveted by the physical book that the text it contained became secondary.

The volume was about six centuries old, bound in worn brown leather and composed of 266 yellowed leaves of carefully crafted parchment. It bore the marks of heavy use — faint stains marked the pages and the edges were worn from repeated handling.

How AI is unlocking ancient texts — and could rewrite history

“It had its own biography, its own deep history. It seemed like an archaeological site between covers,” recalls Stinson, who is now a medievalist at North Carolina State University in Raleigh. “The parchment even had a vaguely animal smell, albeit a pleasant one.”

Stinson wondered whether DNA might survive in the animal skins used to make the book’s pages, and whether that DNA could offer fresh ways to date and contextualize manuscripts beyond the conventional markers of handwriting and dialect. His brother, a biologist, said that this was possible theoretically, but warned that the technological obstacles were daunting. The technologies needed — next-generation sequencing methods and associated computational tools for deciphering the data — were still in their infancy. Even if workable techniques existed, conservators were unlikely to allow destructive sampling of irreplaceable cultural artefacts.

Nearly two decades later, that curiosity has helped to give rise to a new field. The development of non-destructive sampling methods, alongside advances in genomics and proteomics, have made it possible to extract biological information from ancient parchments without visibly damaging them. The emerging discipline — known as biocodicology¹ — combines molecular biology with codicology, the study of books as material objects.

The results are transforming how scholars understand human history. By analysing parchment, researchers are uncovering evidence of trade networks, animal husbandry, medical and ritual practices, climate change, epidemics and floods.

In the process, they have found that ancient parchments preserve more than just words.

A biological archive

During medieval times, parchment was Europe’s dominant writing material, used for everything from legal records to sacred texts. It was made by soaking animal skins in lime, stretching them on frames and scraping them thin as they dried. Even after hundreds of years, parchment bears subtle traces of that process: follicle patterns on the hair side, smoother textures on the flesh side and variations that experienced scholars can read almost intuitively. Its durability has long made medieval manuscripts prized historical objects.

So you want to do biocodicology? A field guide to the biological analysis of parchment

In a 2009 article, Stinson argued that parchment manuscripts represent a year-by-year record of animal life and human–animal interactions spanning a millennium². Why, he asked, were zooarchaeologists focused on excavating bones when a vast, precisely dated, faunal archive has been sitting on library shelves for centuries?

The idea caught the attention of Matthew Collins, a biomolecular archaeologist jointly based at the University of Copenhagen and the University of Cambridge, UK. Collins had pioneered a technique known as zooarchaeology by mass spectrometry (ZooMS) to identify the animal species of old bones³. ZooMS works by analysing fragments of type I collagen, the predominant structural protein in skin, teeth and bone. Species-specific variations in collagen produce distinctive molecular ‘fingerprints’ when measured in a mass spectrometer.

Collins recalls one excavation project in Scotland for which his team analysed more than 1,000 bone fragments. After three years, they could confidently identify just 29 individual animals. “That was a really disappointing project,” he says. When he realized parchment was made from a similar collagen-rich material — and that manuscripts usually announce when and where they were made — Collins was eager to explore its scientific potential.

Without a trace

Sarah Fiddyment was finishing a PhD in cardiovascular proteomics at the University of Zaragoza in Spain when a chance lecture about applying scientific techniques to cultural heritage inspired her to pursue a postdoc with Collins. Collins asked her to develop a method for identifying the animal species in parchment. Fiddyment planned to obtain samples by shaving a thin strip from the manuscript’s edge. But when she arrived at the Borthwick Institute for Archives in York, UK, the conservators refused to let her bring a knife near their documents. “I was effectively faced with a two-year project that was not going to happen.”

The impasse reflected a long-standing divide between the sciences and the humanities — what British novelist and physicist C. P. Snow called the two cultures problem⁴. Scientists are accustomed to drilling into fossil cores or snipping feathers, whereas scholars typically consider even the smallest injury to a medieval page anathema. Any method of sampling the biological material in parchment would therefore have to clear an unusually high bar: its effects would need to be effectively invisible, even under a microscope.

Collins recalls that tense moment in the Borthwick archives as a turning point. “‘No’ is a really powerful word for scientists,” he says, “because it makes you kind of think around corners.” Fiddyment spent a month at the archives observing the conservators. She noticed that they routinely cleaned parchment using blocky white erasers, the kind that grace many a primary-school desk. So, she asked whether she could have the eraser crumbs. “Those little fragments you generate that you blow away, those are the bits I collected, and we found that that worked beautifully.”

The crumbs, which they later called ‘erdu’ for eraser dust, turned out to be molecular gold. When a polyvinyl chloride eraser is pushed across parchment, static electricity lifts microscopic particles from the surface, including collagen and traces of DNA. Fiddyment analysed the crumbs she’d collected using a version of the ZooMS protocol she called eZooMS⁵.

Fiddyment tested her approach on thirteenth-century ‘pocket Bibles’, the tissue-thin pages of which had long been thought to be derived from the skins of animals such as squirrels and rabbits. Her analysis showed otherwise. The parchment was made from the usual suspects: calf, goat or sheep skins. This finding highlighted not that unusual materials were used, but that extraordinary craftsmanship was involved⁵.

But other studies have raised more questions than answers. Stinson recalls the first book he worked on with Fiddyment and Collins: a glossy twelfth-century copy of the Gospel of St Luke. To his practised eye, the manuscript seemed to be made entirely of calfskin. “When the results came back, it blew everyone’s mind,” he says. Testing revealed a deliberate alternation between calfskin and sheepskin⁶. Goatskin was also present, but only immediately after the parable of the prodigal son, which includes the text’s lone mention of a goat kid. “Now, it could be a coincidence, we don’t know,” Stinson says. “But this book is deeply weird.”

Reading residues

Although effective, the method is laborious. It involves rubbing the same patch of parchment until enough crumbs pile up to fill the bottom of a microcentrifuge tube. In the rare-books library at Duke University in Durham, North Carolina, Stinson spent days sampling a single volume. “Honestly,” he says, “it’s like tennis elbow after two days of that.”

While looking for less-punishing alternatives, Stinson partnered with his colleague Kelly Meiklejohn, a forensic scientist whose background includes a postdoc at the FBI Laboratory in Quantico, Virginia. There, she developed methods to identify toxic plants and fungi that had been used as potential biological weapons. These were often in a powdered form and stripped of obvious identifying features.

First known map of night sky found hidden in Medieval parchment

The team tried a range of non-destructive methods on old manuscripts purchased online. Some ideas were ruled out quickly: the dull edge of a butter knife, forensic fibre-lifting tools used at crime scenes, and even gecko tape, which has microscopic bumps that enables it to adhere to surfaces without the use of chemical adhesives. Although technically non-destructive, the tape kept sticking to the laboratory tweezers and tubes, and contained traces of cow DNA, presumably from the manufacturing process.

Ultimately, the researchers zeroed in on two non-destructive approaches: erasers and soft cytology brushes, the disposable tools used for cervical-screening tests. Comparisons showed that the brushes were easier to use and recovered DNA as effectively as the erasers did⁷.

DNA extracted from parchment is typically fragmented into tiny pieces and present in amounts that are too low to be detected using standard assays. But “we proceed with every sample”, Meiklejohn says, because her lab uses a forensic-style workflow designed for such genetic material⁸.

Her team converts the DNA into sequencing libraries and uses a technique known as hybridization capture to fish out animal sequences of interest. Magnetic RNA ‘baits’, designed to match the mitochondrial genomes of species commonly used in parchment, bind to the target DNA, even when sequences differ by as much as 20% from modern genome references. The enriched material is then sequenced and mapped against a panel of 16 reference genomes, including those of human, dog, pig and various species of deer.

On a computer screen, the results appear as a dense, laddered stack of brightly coloured horizontal bars — short stretches of ancient DNA aligning imperfectly but convincingly with modern references. In repeated tests, results using the brush method matched known species identifications and often exceeded expectations.

However, the approach has its logistical quirks. When Meiklejohn had trouble sourcing the right cytology brushes before a planned research trip to the United Kingdom, she took advantage of the opportune timing of her annual gynaecological exam to ask where they were purchased. The clinic offered to provide her with a few bags, but another supplier eventually came through.

Beyond species

In a collaboration with Duke, the team applied its cytology-brush technique to documents across a vast range of time and space, sampling parchments from the eighth to the twentieth century and originating from Europe, North Africa and the Middle East. The results, yet to be published, draw on 351 samples taken from 91 manuscripts. The researchers identified the source species in 58% of cases. Most samples were from sheep, followed by cattle and goats, with a single curious sample indicative of pigskin. They found that species choice mostly tracked regional patterns; for instance, sheep were the main species used in England and goats in Mediterranean regions.