By Erika Check Hayden of Nature magazine

Most scientists would never pursue an experiment that was almost guaranteed to fail. But not only is microbiologist Rosie Redfield, of the University of British Columbia in Vancouver, attempting to replicate the disputed claim that a bacterium can incorporate arsenic into its DNA backbone, she is doing so in public view, detailing her work in an open lab notebook on her blog.

"It takes a ton of huevos on her part," says biophysicist Steven Koch at the University of New Mexico in Albuquerque. "She knows that everyone is watching her work day by day and there are probably a lot of people who want to see her fail."

Partly, Redfield wants to perform the detailed analyses of so-called 'arsenic life' that she thinks should have been in the original paper, published in Science last December, by Felisa Wolfe-Simon, a NASA Astrobiology Research Fellow based at the U.S. Geological Survey in Menlo Park, Calif., and 11 others. Redfield's findings are already contradicting aspects of the original work.

But she also wants to strike a blow for openness in research. "It is such a great opportunity to do open science," she says. "I've been doing all my research openly for a while, but nobody pays attention."

Other researchers say her work could be an important test case. "This is a great case study for open science, because it raises issues about peer review, it raises issues about sharing data and materials, and it raises issues about engaging the public and press more actively in science," says microbiologist Jonathan Eisen of the University of California, Davis.

Growing pains

Redfield had two major criticisms of the original paper: that the authors had not ruled out the possibility that the bacteria they described, dubbed GFAJ-1, were feeding on phosphorus contaminating their growth medium; and that the bacterial DNA was not properly purified, so that the arsenic detected might not actually have been in DNA.

On August 2, Redfield blogged about some initial results that, she said, contradicted the Wolfe-Simon paper. She reported that she could grow GFAJ-1 in a 3 micromolar solution of phosphorus, but that adding arsenic to this low-phosphorus medium killed the bacteria.

Wolfe-Simon's team, in contrast, claimed that the bacteria could not grow on such a low concentration of phosphorus, and that the bacteria could grow on arsenic in the absence of phosphorus

In the original study, however, the researchers grew the bacteria in media containing arsenic before transferring them to various concentrations of phosphorus and arsenic. Redfield is trying to replicate these conditions, but so far her bacteria have not survived in media containing arsenic.

Other researchers also working with GFAJ-1 laud Redfield's efforts, but agree with her that it is too early to conclude that she has debunked the original work.

To get GFAJ-1 to grow, Redfield and others have also had to give it one nutrient not used by the Science authors. This means that Wolfe-Simon's team will always be able to dismiss the work.

"The authors will not accept conflicting results in any case, but ascribe them to differing conditions, as we have changed the growth medium," says Simon Silver of the University of Illinois in Chicago, who is studying GFAJ-1 but is not trying to replicate Wolfe-Simon's work.

Science live

This is not the first time that scientists have kept open notebooks as they tried to replicate controversial findings. In 2009, chemist Paul Docherty, who now works at Akzo Nobel in Reading UK, and commenters on his blog, Totally Synthetic, live-blogged their work as they debunked an astonishing claim that a strong reducing agent, sodium hydride, could oxidize an alcohol.

Peer review by bloggers has also dissected a study on the genetics of ageing published online in Science that was retracted in July, and a solution to a major computer-science problem--the so-called 'P vs NP' problem--mooted by Hewlett-Packard scientist Vinay Deolalikar last year.

Advocates of open research say these cases show how open peer review is speeding the process of science, adding that the practice will become common as young scientists comfortable with Facebook and Twitter begin work.

Open notebooks seem natural to undergraduates, says Koch. "It's a natural tendency for 99 percent of us; it just hasn't been possible until recently."

So far, Redfield is the only scientist who has revealed that she is trying to replicate the arsenic-life research. Most scientists say that it is not worth redoing research that the scientific community has largely dismissed, and that they wouldn't be comfortable having students and postdocs spend time on work that probably would not be published. (See Will you take the arsenic life test?)

Doubters also say that researchers who publicly document their work risk getting scooped by competitors, killing the chances of publication. This does not concern Redfield. "The work is out there, and the credit is out there, so I don't worry about that," Redfield says.

Open-research advocates say that an open lab notebook might be protection against being scooped: "If somebody is going to plagiarize, they're going to do it from things that can't be found on Google," says chemist Jean-Claude Bradley at Drexel University in Philadelphia, Penn.

And, he says, if Wolfe-Simon's group had used an open notebook, other scientists might have pointed out the shortcomings in their methods, allowing them to correct the errors and, perhaps, publish a stronger paper. "If we had access to the lab notebooks, none of this would have happened," Bradley says.

This article is reproduced with permission from the magazine Nature. The article was first published on August 9, 2011.