We are used to radioactive substances having a half-life, but DNA? Now a study of bones from extinct birds suggests the double helix too has a measurable half-life and that we have underestimated its ability to survive in the fossil record.

"DNA degrades at a certain rate, and it therefore makes sense to talk about a half-life," says Morten Allentoft at Copenhagen University, Denmark, who together with Mike Bunce at Murdoch University in Perth, Australia, and colleagues, extracted DNA from the leg bones of 158 extinct flightless birds called moas.

Part of the reason a DNA half-life has been so elusive is that it is hard to find a large enough cache of samples that have been exposed to similar conditions. The moa bones were all between 600 and 8000 years old, and came from a 5-kilometre-wide area of New Zealand's South Island, key factors for helping identify a regular pattern of decay.

With an estimated burial temperature of 13 C, the DNA's half-life was 521 years almost 400 times longer than expected from lab experiments at similar temperatures, says Allentoft.

The oldest DNA to date belongs to insects and plants and was found in 450,000 to 800,000-year-old ice. Under subzero conditions, Allentoft and Bunce estimate that DNA's half-life can be up to 158,000 years, meaning the last remnants would disappear around the 6.8-million-year mark. Allentoft does say that is an optimistic assessment, and doesn't imply that samples of DNA large enough to measure could be extracted from such old bones.

Eva-Maria Geigl at the Jacques Monod institute in Paris, France, is still to be convinced by the half-life claims, which she says rest on statistically weak evidence. She points out, for example, that the correlation relies heavily on the moa bones older than 6000 years when fewer than 10 of the 158 bones are this ancient.

"Old fossils are rare and hence there will be less data in this part of the analysis," says Bunce. "There is nothing we can do about it other than present what we have at hand and clearly, the signal is present. The correlation is highly significant."

If DNA decays in a predictable way, can we calculate the chances of finding it at key sites? Ever since the Indonesian island of Flores yielded remains of the "hobbit", Homo floresiensis in 2004, speculation has been rife that some specimens might contain DNA that would help pin down its position in the human family tree. This notion has been spurred by evidence that the hobbits may have survived until as recently as 18,000 years ago.

Unfortunately, Bunce thinks the new calculations will be difficult to apply to specific sites. "A host of other factors come into play," he says, including the season the organism died. In fact, although the moa bones in the analysis had been buried in a similar environment, the age of the specimens could account for only about 40 per cent of the variation in DNA preservation in other words, the half-life signal is noisy.

Alan Cooper, director of the Australian Centre for Ancient DNA at the University of Adelaide, South Australia, agrees. "The rotting process after death is very seasonal and context dependent, and has a major impact on DNA survival."

Read more:
DNA's half-life identified using fossil bones