Plant biologists have withdrawn a study on Arabidopsis thaliana evolution published in a 2004 issue of Science, saying one of its conclusions was marred by contamination, according to a retraction appearing today (Apr. 10) in the journal.

The original paper, authored by then North Carolina State University genomicist Michael Purugganan and a team of international colleagues, suggested self-pollination had evolved in A. thaliana after a period of rapid evolution - or a selective sweep - after the most recent ice age. It has been cited more than 40 times, according to ISI.

The team pulled the paper because of "spurious PCR amplification by the primers and/or by DNA contaminations," according to the published retraction. Purugganan told The Scientist that in four out of the 21 A. thaliana samples the team genotyped, DNA contamination caused PCR primers to amplify a gene allele (ΨSCR1) that in actuality was absent from a particular haplotype. "We thought we were getting something when in fact we were amplifying some contaminated DNA," Purugganan said.

After the publication of the Science paper in December 2004, Purugganan and his postdoc at the time, Kentaro Shimizu, began hearing "second-hand from colleagues" that they were not finding the ΨSCR1 in A. thaliana samples they were genotyping. "When we heard this," Purugganan recalls, "we said, 'We've got to figure this out.'"

Purugganan and Shimizu realized their error after a few months of reanalyzing the data, and concluded that the extent of the DNA contamination was likely minor but enough to change the conclusion that self-pollination, or selfing, evolved only once throughout the entire A. thaliana species. "It was probably minute amounts [of DNA] that were in the buffer or something," Purugganan said. "We still believe there was a selective sweep, but it was not species wide."

Shimizu, Purugganan and colleagues published an updated study this January in an issue of Molecular Ecology, in which they reanalyzed the botched data and concluded that only the European population of Arabidopsis experienced the selective sweep. "We realized that the data were showing us that there was not just one origin of selfing in Arabidopsis," Purugganan said. "The picture that is emerging now is that selfing has evolved multiple times. It becomes a really fascinating story now."

Purugganan, who is now at New York University, said that discovering the errors in his 2004 Science paper led to a new view of the evolution of self-pollination in A. thaliana, in which the phenomenon evolved once in European and at least one separate time in African and Asian populations.

"By looking at something and knowing there was an error, we were not only able to correct it," said Purugganan, "it's actually more interesting than the original story."