Arsenic and an Old Lake

Felisa Wolfe-Simon collecting samples at Mono Lake

About a year and a half ago, NASA announced a press conference "to discuss an astrobiology finding that will impact the search for evidence of extraterrestrial life." This led to wild speculations on the internet that NASA might have discovered extraterrestrial life, that they would finally admit that they had known about aliens all along! However, what it was really all about was a discovery by geomicrobiologist Felisa Wolfe-Simon, about to be published in the prestigious journal Science. What she claimed was that she found a bacterium in the sediment of Mono Lake in California that could incorporate arsenic instead of phosphorous in its DNA. In theory, life would therefore not be completely dependent on phosphorous (as we always thought it was) but could also grow in its absence and thus in extraterrestrial environments that we would normally consider to be devoid of life. A very exciting discovery in its own right!

Sunset at Mono Lake (my own picture)

Mono Lake is very, very old -about 760,000 years- and, because water goes in but doesn't really go out anymore, it is also very salty. It is not, however, devoid of life-as-we-know-it; brine shrimp live in the lake and migratory birds visit the lake to feed of the shrimp. Arsenic concentrations in the lake are higher than usual, but not toxic. Besides being geologically very interesting, famous for it's tufa towers, Mono Lake is also a very beautiful place and certainly worth a visit if you're in the area.

Arsenic-loving bacteria from the Wolfe-Simon paper

Now, when Felisa's paper came out a few days after the press release from NASA, I was highly disappointed and many other scientists with me. First, this new bacterium, GFAJ-1, did not represent a 'completely new form of life' but rather an old form of life that had adapted to survive at high concentrations of arsenic, such as found in Mono Lake. Second, the authors did not, in any way, conclusively demonstrate that GFAJ-1 could survive and replicate in the complete absence of phosphorous or that it incorporated arsenate, instead of phosphate, in its DNA. Trace amount of phosphate the bacterium's growth medium could easily account for its growth and they did not exclude that the arsenic they found in samples of DNA isolated from GFAJ-1 (through rather crude methods!) could simply be due to contamination of their sample. In addition, they showed no direct evidence that the arsenate was actually incorporated in the DNA of GFAJ-1. Canadian microbiologist Rosie Redfield extensively discussed these and other flaws in her blog, almost immediately after the paper was released.

Rosie Redfield

Redfield has subsequently attempted to culture GFAJ-1, as obtained from Wolfe-Simon, herself and perform the control experiments that Wolfe-Simon neglected to do. She has consistently posted her progress on her blog, responding to suggestions from readers in the process. Earlier this year, this led her to conclude that the Wolfe-Simon paper was indeed flawed; GFAJ-1 does not grow in the complete absence of phosphate and its DNA, even at high concentrations of arsenic in the medium, does not incorporate any arsenate. These findings have now been published in Science as well (also freely available here), back to back with a paper from another group that found the same thing. Redfield's experiment with 'open science', publication of her constant progress on her blog, was an interesting one in it's own right. Hopefully one that will be repeated more often, especially where controversial issues involved.

Redfield's evidence; no phosphate, no growth

An extraordinary claim requires extraordinary evidence and Wolfe-Simon did not provide this evidence for her claim. It's therefore hardly surprising that the entire scientific community fell over Wolfe-Simon like an avalanche and that this caused poor Felisa more than a little bit of stress. She admitted that the data presented in the Science paper were somewhat preliminary and that she hoped the paper would get her the help she needed from specialists outside her direct field to really determine whether or not the DNA of GFAJ-1 really incorporated arsenate instead of phosphate. I find that defense a bit weak. Why would you even submit a paper to a prestigious journal like Science if you're not absolutely, positively sure about the validity of your claims? I certainly wouldn't. Plus, if you are aware of the flaws in your paper, then why don't you connect to the specialists you need to clear up those flaws? It's really no excuse at all and the paper is just full of obvious flaws (as Redfield pointed out in her blog).

Someone should have stopped her from staking such claims without the evidence to support them. In part, NASA is to blame for hyping their press release and making it sound like they had discovered more than they had. The journal Science is to blame as well for their obvious over-eagerness in accepting this study for publication. Qualified peer-reviewers would have never allowed this. I wouldn't have allowed it. I'm guessing that the editor decided to publish in disregard of the peer-reviewers' criticisms, hoping to generate publicity (and they certainly did!). This case would certainly argue for the co-publication on-line of the reviewer reports of every publication. Some journals already do that, but it still isn't common.

Wolfe-Simon has staked the hypothesis that arsenate might replace phosphate in some organisms before and her Science paper was supposed to be the proof of the pudding. She apparently wanted to prove her hypothesis so badly that she forgot about the fundamentals of science, got over-excited, forgot to falsify her findings and got blinded by the positive evidence. Scientists are only human in the end, we see what we want to see, especially when we get excited about our findings. In theory, that's where peer review should step in. In this case, peer review failed, probably as a consequence of the Science editor looking for sensation (although the reviewers might have been incompetent too). I wouldn't go so far as to call Wolfe-Simon a fraud. She merely suffered from over-confidence and a certain dose of naivety. I hope this experience has not embittered her and that she has come to see that she was wrong. It is very hard to admit when you're wrong, especially about something you really believe in, but I really hope she will admit her mistake and emerge from this experience that much wiser.

Most of all, I hope that other scientists will learn from this scenario that they should always, always, question their own data and their own hypothesis! If the data leaves any room for doubt, get back in the lab and eliminate that doubt! It has become a worrisome trend among scientists to focus on positive data, data that supports their pet hypothesis, instead of looking for ways to refute the hypothesis. Could the data be explained in any other way? That is the question that should always be asked. If you find another explanation, exclude it. That is the way science is supposed to work, and if you can't think of a better explanation for your data, ask a colleague. There's something to be said for making bold claims and building hypotheses, but the top-tier journals are not the place for that. Please don't contaminate the literature with those half-supported claims! In the end, there's nothing wrong with rejecting your hypothesis as false. Better to admit you're wrong than to stubbornly defend an indefensible position.