Live-blogging Arsenic Life

By Carl Zimmer | July 8, 2012 4:39 pm

I’m spending the weekend in Ottawa, where a couple thousand scientists have gathered for the Joint Congress of Evolutionary Biology. I’m drowning in a torrent of fascinating talks, on everything from sexually cannibalistic crickets to the future of the Amazon’s biodiversity. In the evenings, the meeting features high-profile talks–Friday night, the science writer David Quammen spoke about his career, on the occasion of winning the Stephen Jay Gould Prize. I have a particular interest in tonight’s talk, so much so that I’m going to live-blog it. The speaker is one Rosie Redfield, and she’ll be talking about the endlessly intriguing case of Arsenic Life.

Before Redfield takes to the stage at 7:30 pm ET, I want to write a short preface. In December 2010, rumors swirled for a few days that NASA had discovered alien life. When they finally held a press conference, the world discovered that a team of scientists had found a species of bacteria at Mono Lake in California that appeared to be able to build DNA out of arsenic. If true, it would be unlike any known life on Earth.

Rosie Redfield, a microbiologist at the University of British Columbia, read the paper after the embargo at Science was lifted, and didn’t like it. After she posted her complaints on her blog, I got in touch with her, along with a number of other scientists to see what they thought. Most of them didn’t like the paper either. I wrote about their collective reaction in an article on Slate, which was called “‘This Paper Should Not Have Been Published.'” (That was a quote from one of the scientists I spoke to.)

The authors of the paper, who were willing to hold a big press conference, give TED talks, and so on, refused to provide me any comments. They declared that all discussion must be restricted to peer-reviewed channels. Six months later, when Science finally published the arsenic life paper in a print edition, they also published a number of comments from Redfield and other critics.

Redfield and others argued that the arsenic life study had been poorly carried out. To put a scientific claim to the test, critics can try to replicate it. But, as I wrote in the New York Times, that’s a pain in the neck, and a lot of researchers would rather spend their precious time doing something interesting, rather than cleaning up other scientists’ messes. But Redfield decided to replicate the arsenic life study anyway. She had another motivation for doing so: she’s a big fan of open science, and so she used her blog to chronicle her experiences, from receiving the bacteria from the original authors to failing to replicate their results to posting her paper on Arxiv to getting her paper accepted to Science, where it’s now in press.

As Redfield notes on her blog, tonight’s talk presents a certain complication to the traditional way that high-profile journals publish papers. Journals like Science keep their papers under strict embargo, meaning that people are supposed to keep their mouth shut until an appointed time. Science provides pre-publication copies of the papers to journalists, usually a week in advance, so they can write up articles that can appear once the embargo lifts. So…if Redfield talks about her research, what does that mean for her paper? And does the fact that you can download it already from Arxiv make all of this moot?

Stay tuned!

7:36: Getting started. Redfield’s getting introduced.

7:38: “Rosie Redfield led the charge that NASA got it wrong.”

7:41 Redfield is starting: this is a story about the process of science, that starts with a press conference.

7:42 Redfield glosses over the utterly insane week of speculation that NASA had found aliens. Hey, she’s a microbiologist.

7:43 Why was this done? “The researchers were looking for exactly what they found.” (Not always a good thing in science.)

7:44: Looking for alien life is very expensive. So NASA is looking for terrestrial life that can help the search.

7:44: 2009 paper in Astrobiology: arsenci life team pointed out that arsenic is similar to phosphorus. So maybe early life could have used arsenic instead of phosphorus (in DNA for example). And maybe they’re still around.

7:45: Origin of life. You can start forward from early Earth, or go back from life today to figure out what first life must have had.

7:46: Early earth: simple chemicals created complex ones. Also, early Earth was “stinky.”

7:47 Comets delivered dirty water, loaded with polycyclic hydrocarbons.

7:48 All life today is all related. All life descended from a common ancestor.

7:50 Common ancestor of all life today: DNA genome, cellular, lipid bilayer, protein synthesis, many modern biochemical pathways.

7:51 There’s a gap scientists have yet to bridge from prebiotic Earth to last common universal ancestor.

7:52 What properties are essential in living things? Reproduction, heredity, and heritable variation that affects survival or reproduction. Redfield: “That spells natural selection.”

7:53 Natural selection probably kicked in long before our last universal common ancestor. First life could have much simpler metabolism. May not have been based on DNA. Life could have started as RNA. (Check out this story I wrote for Discover.)

7:55 Today, proteins do much of the work of metabolism.

7:57 Creationists seized on this open question. Evolutionary biologists claimed evolution doesn’t include origin of life. But RNA world gets closer to the origin, via evolution.

7:58 Back to arsenic: Poisonous because cells sometimes take up arsenic, which screws things up.

7:59 “Shadow biosphere”–a cool name, but almost certainly wrong.

8:00 Shadow biosphere hypothesis: weird life exists even today, hiding from scientists. Redfield: “This is extremely cool.”

8:01 Redfield recaps arsenic life experiment. Mono Lake bacteria kept growing as arsenic was swapped for phosphorus in medium.

8:02 Once Felisa Wolf-Simon could grow bacteria, brought in other scientists. Sophisticated tests such as NanoSIMS to show arsenic in (or near) arsenic.

8:04 Weirdly, Science has just posted Redfield’s paper, and a second paper documenting a failure to replicate, on their web site without telling journalists.  

8:05 Redfield was suspicious. NASA’s track record was not good, given Martian meteorite debacle.

8:06 Whenever Redfield wants to think something through, she blogs.

8:07 Press release has just gone out on arsenic life. There are TWO  papers refuting it.

8:09 Redfield: Medium in arsenic life had some phosphate in it–same as ocean water.

8:11 Arsenic bacteria got fat. They couldn’t divide any more, so they swelled up.

8:13 Redfield argues that DNA of arsenic life bacteria that was already dirty with arsenic–arsenic not part of DNA.

8:15 Dan Vergano at USA Today is reporting on the new papers. Arsenic life paper lead author Felisa Wolf-Simon declares, “There is nothing in the data of these new papers that contradicts our published data.”

8:17 I forgot to link to my database of the responses from many critics. Here it is.

8:18 For some reason, some people made fun of Redfield for her dyed hair. It’s just gray at the moment.

8:19 Arsenic in DNA would fall apart very quickly. A big problem.

8:19 The bacterium in the arsenic life paper was not part of the shadow biosphere. It was just a member of a well-known genus called Halomonas.

8:21 The bacterium, called GFAJ-1 (Get Felisa Wolf-Simons A Job), is very, very, closely related to other Halomonas. Not much exotic evolution to become arsenic life.

8:22 No reason to think there would be natural selection for arsenic use, since Mono Lake has a fair amount of phosphorus (although a lot of arsenic too).

8:24 NASA suddenly switched from big press conference to shutting up. Felisa Wolf-Simon “had been been tweeting up a storm.”

8:25 Six months after the press conference, paper published, plus criticisms. Response of authors: “We think our research is fine.” –Redfield.

8:27 Redfield’s “selfish reason” for focusing on arsenic life: she blogs about her work, pursing open science. Important that the public see how science is done.

8:29 Redfield lacked all the expertise required for trying to replicate arsenic life. Blog post led to collaboration with experts who could do the job. (Actually, they volunteered their grad student!)

8:30 Redfield spent a month trying to get bacteria to grow, blogging all the way. Trouble with amino acids. Eventually, figured it out.

8:32 Redfield used better reagents with much less phosphate. Arsenic didn’t help growth. Arsenic life team was wrong.

8:33 Her DNA was stable for a couple months. Couldn’t have arsenic in it.

8:35 Redfield et al wrote up a paper, submitted to Science, and uploaded to Arxiv–because this is open science.

8:36 Science planned to publish on July 26.

8:36 Journals use embargoes to control how research is publicized. For more, read Embargo Watch.

8:38 Embargoes enforced by “vague threats.”

8:38 Weirdness of paper being on Arxiv for months and Science enforcing embargo. Redfield informed that embargo being lifted tonight.

8:39 I just told Redfield her paper has been up for half an hour. Applause breaks out.

8:40 Redfield: This is a story of serial failure. Lead author convinced of evidence without good research, senior authors didn’t provide supervision. Co-authors should have accepted responsibility. Reviewers failed, missed a lot of problems. Science failed in selecting reviewers.

8:42 “And finally, NASA failed big time.”

8:42 But the process of science did not fail.

8:45 Talk over. Question–will this episode change science? Redfield: If you blog about it, you mark that it’s your idea.

8:47 Peer review like democracy (as described by Churchill), is terrible, but the best we have.

8:48 Redfield: lots of “seriously flaky” stuff gets published, so we don’t have to worry about strangling science.

8:49 Idea of arsenic life “unlikely to be true,” but an “okay hypothesis.” You always fall in love with your own ideas. “It’s the ultimate high in science.” But you need self-discipline to test your hypothesis and see if it’s wrong.

8:52 That’s it, folks. Good night!



CATEGORIZED UNDER: Arsenic life, Writing Elsewhere

Comments (26)

  1. Jerry Hodge

    Your link to “her blog” actually takes you to the paper. Here’s the link to her blog: (Note: be sure to check out the comment thread)

  2. george

    Let’s give this a rest. A paper was wrong. People figured out it was wrong. Happens a lot. Life goes on.

    I think that much of the blame rests with the NASA publicity people who pushed this story too hard. The investigator may have gotten somewhat caught up in the rush.

    However, having read the original paper (I am a university biochemistry professor) I would say that the data suporting the claim that there was arsenic substituting for phosphorus in the DNA was indirect and weak and that the review process failed.

  3. fascinating to see the live-blogging of this talk synchronous with release of paper… some day this whole affair will be a classic case in a textbook on the evolution of science journalism/reporting.

  4. Is FWS serious? The Redfield paper completely shreds her data and conclusions.

  5. NotAnAstrobiologist

    People may often figure out that papers are wrong, but in those cases they are generally retracted. I am excited to hear more about things, so I am glad this article is up. Thanks Carl!

    People have to remember that this wasn’t a simple case of a mistake that was caught soon after publication. This was an audacious announcement that was met with virtually immediate criticism, delayed only by the credit one would give to such a widely publicized event from a legitimate organization. Those criticisms were met by strong denial from the authors (take a look as to how FWS – Steve Benner interaction during the original press conference) *and* officials (with, apparently, little scientific background).

    Personally, I don’t think this is (just) a case of scientists “seeing what they want to see”. I don’t have a better word for it than “TED Complex”…the idea to go from a poor graduate student to famous scientist who quotes Sagan, reads classic papers (not quite sure why FWS would read the Watson and Crick papers…doubt Bessel functions and layer lines were very useful for her), and effortlessly paints masterwork in front of us that reveals some deep secret of the natural world.

    Articles were written, news reports were broadcast and children produced classroom assignments (you can find them on youtube) extolling the GFAJ-1 discovery. Personally, I would feel bad if we let kids get all confused with this stuff. It is also important to let go of this idea of the Science Martyr. There are some good examples in the past of smart people getting it very wrong…this isn’t one of those times. Until the authors admit to the shenanigans, it is important for articles like this one to continue to straighten things out.

    There was (and is) just something so peculiar about this entire event. Somehow a bunch of people that should’ve had there thinking caps on decided to take them off (some seem to refuse to put theirs back on). Why?

  6. The methods for the paper from the Zurich group looks like they didn’t need the additional glutamate to make GFAJ-1 grow. Yet FWS still says “There is nothing in the data of these new papers that contradicts our published data, which is also consistent with our current results”?!?

  7. NotAnAstrobiologist

    Mat McFarlane,
    I can’t see the papers, but if I understand your comment correctly, the addition of glutamate was something Rosie found she needed to grow GFAJ-1 (not FWS). It’d be interesting to resolve the difference between Rosie and the Zurich group; I don’t see how it is anything other than a minor difference.

    The key contradiction to the FWS paper and these two new papers is the fact that GFAJ-1 does *not* grow with arsenate in limiting phosphate (although Rosie did notice a stimulatory effect of the rate of growth, but not final density). Without that observation, it isn’t clear why anyone would even try to prove that aresnate is incorporated in the DNA backbone (or anywhere else).

  8. One of the most interesting aspects of this is that it turns out to be a controlled test of “open science”! The second paper, Erb et al.– completely unexpected as far as I know — offers a chance to compare Redfield’s Open Science paper to the traditional Closed Science version. They reach essentially identical conclusions, were published at exactly the same time in the same place — the only significant difference is that Redfield described her work as it progressed, where Erb et al. took the standard approach.

    What are the predictions for the hypothesis that “Open Science is better”? (Or, state the hypothesis and make predictions.) Will there be a difference in formal citation rates? Will the impact on the respective first authors be different (job offers? Talk invitations?) Will one paper be more influential?

  9. I’m having a really hard time understanding this line in your article, Carl.

    “To put a scientific claim to the test, critics can try to replicate it. But, as I wrote in the New York Times, that’s a pain in the neck, and a lot of researchers would rather spend their precious time doing something interesting, rather than cleaning up other scientists’ messes.”

    As scientists, we replicate others experiments all the time. A lot of science is derivative, meaning one finding breeds a whole other set of findings. In the process of making these other findings, we replicate the original finding many times over because it usually serves as the positive control! Whenever a new cell line, plasmid, etc comes into my lab, I do quality control experiments to determine if its behaving as originally described. In the case of a cell line, I make sure it expresses all of the right markers, harbors the right virus, etc. A bacteria like GFAJ-1 would be no different. The work that was done here was not “a pain in the neck.” It’s already a built in error correcting function of the scientific process.

    [CZ: As I write in my linked essay, a number of scientists dismissed arsenic life but didn’t want to spend the time to try to replicate the results. This has proven to be the case in a number of other high-profile studies. Making matters worse, journals will sometimes refuse to publish failures to replicate, saying that they’re not presenting any new results.]

  10. There’s more than one journal to publish your results in. If The Journal of Personality and Social Psychology won’t take your article, then dump it off at PLoSONE and let altmetrics take over.

    Your presentation of the XMRV data as “still standing” is somewhat misleading. Every virologists knows that data was flawed. We KNOW her PCR reagents were contaminated and that’s what produced the errant data. The scientific community has rooted out these flaws through replication. My point is that the process of science is all about replication and the further research scientists do on a particular topic is constantly validating (or in some cases invalidating) previous findings. This IS the scientific process, and it’s not a pain in the neck (usually).

  11. zmil

    @Brian Krueger

    “As scientists, we replicate others experiments all the time. A lot of science is derivative, meaning one finding breeds a whole other set of findings. In the process of making these other findings, we replicate the original finding many times over because it usually serves as the positive control! ”

    This is of course true, but the key difference is that, as a rule, when you try to replicate results, you expect it to work. Trying to replicate results when you don’t believe the original paper is a whole different kettle of fish. First, proving a negative is boring, most of the time. Showing something doesn’t work isn’t very interesting, because most things don’t work. It’s the null hypothesis.

    Second, it’s hard. When you’re looking for a positive result, the details don’t matter so much, as long as you can get it to work. When you’re proving a negative, you have to do a lot more work to prove that your failure to replicate is legitimate, rather than a result of incompetence. Troubleshooting is flipping time consuming, even with experiments that should work. Wantonly inviting extra troubleshooting into my life is…no.

    And last but not least, unless you actively delight in destroying careers, disproving other people’s work is not entirely pleasant. Sure, it’s for the greater good, but you’re still basically trying to show that someone else is/was incompetent.

  12. zmil

    Actually the original CF+XMRV paper has been fully retracted. Though I would note that, in my view, the final nail in the coffin was discovering the laboratory origins of XMRV, rather than the bazillion papers that tried and failed to replicate the original results.

  13. @zmil
    And when the experiments don’t work, and the hypothesis was that you thought the original data was flawed, and you’re an astrobiologist…the process doesn’t seem so unnecessary or boring. On the flipside, say you thought GFAJ-1 actually incorporated arsenic. Even if you don’t do the quality control step, in the process of figuring how GFAJ-1 incorporates arsenic you’re going to discover that arsenic isn’t showing up in the right places. You’ll essentially be performing another replicate in that case, but using different techniques.

    Again, my point is that one of the cornerstones of science IS replication. It’s how we know whether our results are right. And it’s even better when other labs validate our results…through replication. The major finding from my graduate work was validated by four other labs, all using similar techniques and publishing within a year or two of one another so it’s definitely not impossible to get similar work published. The important theme is that most scientists are careful, thorough, and see replication as a very important part of the scientific process.

  14. Sara

    Was it really necessary to mention her hair? What does her hair or hair color have to do with the topic at hand?

  15. jonathan

    One of your links is wrong – the link “her blog”, referring to Ms. Redfield’s blog, goes to the same link as the “her paper on Archiv” does.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

The Loom

A blog about life, past and future. Written by DISCOVER contributing editor and columnist Carl Zimmer.

About Carl Zimmer

Carl Zimmer writes about science regularly for The New York Times and magazines such as DISCOVER, which also hosts his blog, The LoomHe is the author of 12 books, the most recent of which is Science Ink: Tattoos of the Science Obsessed.


See More

Collapse bottom bar