Twigstats software sheds light on mysteries of Europe's old-school migrators

Software developed to investigate more detailed differences between ancient DNA data has proved its worth this week after a paper describing human population movements in Europe in the first millennium AD was published in the science journal Nature.

Speaking to The Register, mathematician and open source maintainer Leo Speidel said he hoped the project would go on to help reveal the human genetic history of Japan, Poland, and the UK.

Speidel, who recently started as group leader in interdisciplinary math and science at Japan's research institute RIKEN, was the first author of research, which sheds new light on population movements during the Roman occupation of Northern Europe, as well as Anglo-Saxon and Viking migration in the first millennium.

Hitting the headlines, the study promises new evidence for a much-debated period of European history.

Speidel said the software project powering the study – dubbed Twigstats – is built on an earlier tool that has been widely used in the field since 2019 to help date genetic mutations and build a genetic family tree. Mutations cascade through genetic history, so the more mutations one individual shares with another, the more closely related they are to each other.

Coded in C++ and employing the statistical language R, Twigstats allows researchers to focus on a specific period in history with more fine-grained analysis than previously possible. Demand for such analysis has grown since researchers developed techniques for extracting DNA from ancient bones over the last 15 years (see Tom Higham's excellent book, The World Before Us, for a description of these techniques). DNA from samples more than 100,000 – even a million – years old has been successfully extracted.

"We can now take these mutations, throw away old mutations, and focus on the more recent mutations," Speidel said. "Those more recent mutations should be more informative on events that happened quite recently. If you're interested in the first millennium AD in Europe, then you're interested in mutations that are closer to that time period than mutations that occurred hundreds of thousands of years ago. The idea was to take mutations, throw away the old ones, and focus on the more recent ones and that increases the signal-to-noise ratio manyfold and allows us to get a better sense of these mixtures."

• Tardigrade genes may hold secret to radiation treatments for humans
• Say hello to the epi-bit, a new approach to DNA data storage
• Heart of glass: Human genome stored for 'eternity' in 5D memory crystal
• In homage to Jurassic Park, researchers store DNA in amber-like polymer

The research published this week was led by the Francis Crick Institute, where Speidel was a postdoctoral researcher.

Pontus Skoglund, group leader of the Ancient Genomics Laboratory at the Crick and senior author, said: "The goal was a data analysis method that would provide a sharper lens for fine-scale genetic history. Questions that wouldn't have been possible to answer before are now within reach to us, so we now need to grow the record of ancient whole-genome sequences."

Speidel continues to maintain Twigstats, which is available as an open source project under the permissive MIT licence, and hopes other researchers will contribute.

He said other studies hoping to use the software are set to look into population movements in Anglo-Saxon England, Poland, and ancient Japan. But he said the technique could be equally applied to other species. "This isn't only applicable to humans; we can apply this to essentially any sort of sexually reproducing organism, like other mammals, and even plants and fish and all sorts of things," he said.

"What's really nice is that it's quite simple. It combines two ideas, starting off with mutations and these existing tests with ancient DNA will add. We were able to combine those two to make this new tool. The hope is that it's quite similar to what people are already doing and doesn't really need a big change in thinking from the user side." ®