Researchers have uncovered an error within the Science paper reporting the first ancient African genome, the New York Times reports.
The paper from the University of Cambridge's Andrea Manica and his colleagues reported on the genome of a man dubbed Mota who lived in Ethiopia some 4,500 years ago. Based on further analyses, they reported that Mota appeared to be related to the Ari — a contemporary Ethiopian highland population — and that there appeared to be extensive genetic backflow from Europe into Africa about 3,000 years ago that then spread throughout Africa.
After the paper was published, Harvard Medical School's David Reich and Pontus Skoglund asked the authors to send them the data, as they wanted to fold it into their own studies. But when Reich and Skoglund re-analyzed it, they found that their conclusions didn't match those of the original authors. As Zimmer notes at the Times, they were unable to find evidence of that backflow beyond East Africa.
It turns out, Zimmer says, that the bioinformatics approach Manica and his colleagues used requires a formatting change — a step they missed. "It was clear human error," Manica tells him. "It's just something that should have been done that didn't get done."
That error meant that they missed regions of Mota's genome that were similar to those of Eurasians and that their analyses then erroneously suggested that Africans from outside East Africa shared DNA with Eurasians that Mota did not.
Reich notes that other findings in the study hold up such as Mota's genome sequence and that the Ari seem to be closely related to Mota. And that Eurasian backflow did occur, though on a smaller scale. "At least the general story still stands," Manica says.
Manica and his colleagues have asked Science to issue an erratum, and the journal says it is considering it, the Times adds.