Thursday, June 15, 2006

Open Access: Stage Two

Public Library of Science (PLoS), the non-profit open access publisher, has announced the launch later this year of a next-generation publishing system. The new service will allow researchers in biology and medicine to publish peer-reviewed papers in an open access online repository called PLoS ONE. Chris Surridge, the UK-based managing editor of PLoS ONE, talks to Richard Poynder about the new initiative, and explains why he quit one of the world's top science publishers to work for a small upstart with a still unproven business model.

A 2011 follow-up article on PLoS ONE is available here

RP: Can you begin by saying something about your background? You started out as a biophysicist?

CS: Yes. I did an undergraduate degree in biophysics at Leeds University and then a PhD at Imperial College.

RP: What was your area of research?

CS: I was looking at microtubule assembly, mainly in vitro.

RP: You'll forgive me, I'm not a scientist. What do biophysicists do?

CS: Biophysics is a component of molecular biology. What you are looking at are the molecules that make up organisms and how they interact. So the focus is not on the chemical properties of those molecules but on their physical properties.

RP: Their structure rather than their chemical interactions?

CS: Well, structural studies are very important within biophysics, but the point is that we use physical techniques rather than chemical techniques to explore molecules.

RP: Your web page says that your specialism was "the dynamics of microtubule assembly". What is a microtubule?

CS: Microtubules are part of the cytoskeleton, which is the structure that holds a cell in its shape, and allows it to do the things it does. So in the way that an organism has a skeleton — whether an internal one consisting of bones like you and I have, or an external one like those that insects or, say, lobsters have — cells have a structure that holds them in shape and gives them mechanical stability. Microtubules also help cells to divide.

Sounded like fun

RP: You subsequently went into science publishing and worked as an editor at Nature. Why publishing?

CS: The opportunity arose at the point where I had done a couple of years as a postdoc, trying to get the experiments that didn’t quite work for my PhD to work, and generally tidying the research up. My grant money was running out so I started exploring a number of different avenues and I came across an editorial job at Nature Structural Biology. I decided to take a punt, and applied for the job — as much to find out what an editor actually did as anything. It also sounded like it would be fun!

RP: No regrets at leaving research?

CS: Not really. While I enjoyed doing experimental research I could see that it could become wearing. I have always liked the breadth of science, which is something you can easily lose track of when you are doing research. You have to concentrate quite hard on the problems that you are specifically handling, especially when you are a postdoc.

RP: You also edited Nature's Brief Communication?

CS: Right. I was an assistant editor on Nature Structural Biology for eighteen months. Then an opportunity came up to work on Nature itself, again looking after primary research in the biological field. I did that for about a year, and then I looked after the Brief Communication section of Nature half time for about nine months. I also spent some time as Nature's web editor.

Just another publishing job?

RP: Then last year you gave up traditional publishing and moved to the open access publisher PLoS. You underwent a conversion to OA then?

CS: You make it sound like moving to PLoS was some kind of road to Damascus thing! It wasn’t quite like that.

RP: So is working at PLoS just another publishing job to you?

CS: No, it's not just another publishing job. It was more that moving to PLoS to look after this new project that we are calling PLoS ONE was too exciting an opportunity to miss.

That said, I have always had sympathy for open access publishing, which I think is a very, very good idea. Even while I was at Nature the logic of it seemed compelling. It wasn’t so much a conversion to OA, therefore, but an opportunity to do something that I think is very important for scientific publishing.

RP: How were you recruited? Did you answer a job advertisement?

CS: No. A lot of the editors within PLoS are old colleagues of mine — at least four of them have passed through the Nature editorial office in their careers. They knew the kind of things I was interested in, so I was rung up and told about the PLoS ONE project, and asked whether I would be interested in being involved.

RP: Clearly you were interested. But tell me, why does the world need PLoS ONE?

CS: Because the system of disseminating scientific research has become extremely inefficient, and the concept of the journal has been eroded by the Internet.

RP: How do you mean?

CS: The Internet is providing us with all sorts of new tools for communicating science, so simply publishing journals, and distributing them electronically isn’t using the full potential of the Web to make the dissemination of scientific information efficient and effective. It just isn't the best way to do it anymore.


RP: When I interviewed PLoS co-founder Harold Varmus he described PLoS ONE as "a very large compendium of papers that have been vetted for scientific quality, but which will not be confined in terms of their likely importance." Is PLoS a journal?

CS: I haven’t got a better word for it, but it's not really a journal: it's somewhat different.

RP: In what way?

CS: If you think of a journal you immediately start asking questions like "So what area of science will this journal cover?" But we can’t think of a good reason to sub-divide PLoS ONE.

RP: Why?

CS: Because if you are looking to provide an extremely efficient and cost-effective way of disseminating the results of scientific research, and you start to think about dividing things up into biology and medicine, and specialities like biophysics, you quickly run up against a boundary question: where, say, does biophysics end and biochemistry begins? You find yourself starting to have to make lots of unnecessary decisions about borders.

RP: The way that science has been divided into different disciplines says something about the way that it has developed presumably?

CS: Actually, I think science subjects developed in a pretty arbitrary way. The boundaries have effectively been imposed as a consequence of the way that journals work, and the way that universities are structured.

RP: So there is no need for such rigid boundaries?

CS: No. At least, not when publishing. Consider, for instance, if you were to have a study that compared the genomes of man and the great apes, looking at a cluster of genes that controls the development of the brain cells associated with language. Now what subject is that? Is it neuroscience? Is it evolution? Is it genetics? Is it genomics? Actually, it is all of these things, so you could classify it in any one of these subject areas.

RP: Why the name PLoS ONE? What does "one" imply here?

CS: One implies inclusivity. What we are saying is that we offer one place to publish; that if you can go to PLoS ONE you don’t need to go elsewhere.

RP: So your message to the scientific community is, "If you have a paper you want to publish we are happy to take a look at it?"

CS: We are happy to look at your research.

RP: You must have some boundaries in mind though. After all, if someone came to you with, say, some research on nuclear physics you would presumably say that that is not what PLoS ONE is about?

CS: Actually, I wouldn’t say that at all. What we are about is science, and science is a way of answering questions about the world.

So if someone came to PLoS ONE with nuclear physics we would welcome them with open arms, and go and find some editors who knew something about nuclear physics who could make a decision about whether the paper was worth publishing. And if they felt it was worth publishing we would then publish it. That's what I mean when I say we are about inclusivity.

That said, we don’t anticipate people with nuclear physics coming to us to begin with because that is not where PLoS' reputation currently is.

The idealised general reader

RP: Right. PLoS' reputation is in biology and medicine. So PLoS ONE is not so much a journal as an online repository — or to use Harold Varmus' word, compendium — of peer-reviewed papers in biology and medicine?

CS: Right. The other point to make is that if you think of a journal you start imagining selecting papers for presentation to a particular target audience. The problem is that the decisions surrounding this selection are generally done in a subjective way.

RP: Can you explain what you mean?

CS: What happens is that journals like Nature, Science, Cell and even PLoS Biology and PLoS Medicine work on the principle that they want to take the most interesting and best science that they can lay their hands on, and provide a selection that their audience will want to read.

Yet the audience that these journals have in mind is in the main a convenient fiction, since they are thinking about some idealised general reader for every journal, and assuming that that general reader will be interested in pretty much everything that the journal publishes. No journal editor believes that such a ‘general reader’ exists but that is the mindset behind the journal model.

RP: And as you say, that is an outdated mindset in the age of the Web. One could perhaps also argue that PLoS ONE harks back to the idea of "Proceedings" or "Transactions"?

CS: In a sense, yes. Certainly we are making it much easier for authors because they won't have to ask questions like: "So what journal am I supposed to submit my work to?" They won’t have to worry about establishing the subject remit of a journal when publishing with PLoS ONE.

RP: So talk me through how PLoS ONE works in practice, and how that differs from the traditional journal model?

CS: OK. Researchers will submit papers to PLoS ONE and the editors — of which there will be a large number in order to cover the breadth we expect PLoS ONE to have — will make a decision whether to publish the paper. In this regard we will not be changing the system hugely.

RP: The criteria for accepting a paper, however, will be different will it not? The PLoS ONE press release, for instance, says that "subjective considerations like 'likely impact,' 'degree of advance,' or 'interest to a general reader' will not play a role in deciding whether an article should be published or not."

CS: That's right. Traditionally a lot of the work that goes into peer reviewing consists of asking questions like: "How significant is this? How surprising are the conclusions?" Essentially, these are subjective questions. A more objective question to ask would be: "Is this properly done science".

Simpler question

RP: So what question will peer reviewers be expected to answer when considering whether a paper submitted to PLoS ONE should be accepted for publication?

CS: They will be asked to answer a simpler question than has traditionally been asked. Essentially, that question will be: "Has the science in this paper been done well enough to warrant it being entered into the scientific literature as a whole?"

What is also different about PLoS ONE, by the way, is that we do not see peer review ending on publication of the paper.

RP: Can you expand on that?

CS: Sure. We believe that the more subjective questions about how a paper relates to other work, and where it fits into the whole corpus of scientific literature are still important questions — but we feel that these can be better answered via an open peer review process that takes place after the paper has been published.

RP: So the intention is to go through the traditional peer-review process (although asking a simpler question when considering whether a paper should be published), but then to have additional reviewing take place after the paper is published?

CS: Exactly.

RP: How will this open peer review take place?

CS: Every paper in PLoS ONE will have a discussion thread attached to it. We are also developing ways to allow people to directly annotate the papers themselves.

RP: What else will PLoS ONE do as part of the publication process?

CS: We will also be tagging the papers so that references to them can be linked into the major databases. And we will be assigning DOIs [Digital Object Identifiers] to them, and to their component parts.

In addition, we will undertake archival maintenance of the papers, and we will also submit them to PubMed Central for archiving, as we do with the other PLoS journals.

RP: Do you expect people to find the papers primarily by searching PLoS ONE, or will they come via PubMed Central, or even perhaps Google?

CS: People will come through Google Scholar, and through PubMed Central, but we hope that they will also come to PLoS ONE. After all, with such a broad spectrum of papers we are offering more than a single journal. We also hope to make the experience of coming to the PLoS ONE site a very personalised one.

RP: Can you elaborate on that?

CS: What we plan to do is to give users of PLoS ONE an identity. This will allow us to develop a number of personalised features for them. When they come to the site, for instance, we will already know the sort of things that they are interested in, and so will be able to suggest recently published papers that they are likely to be interested in.

If they tell us that they are only interested in biochemistry papers, for instance, we will show them the biochemistry papers first. And instead of traditional table of contents alerts we will tailor alerts for them that will signal when papers in the subject area that they are interested in have become available on PLoS ONE.

Semantic Web

RP: Presumably these will be based on automated keyword searching. Do you expect PLoS ONE to also embrace the Semantic Web in the future? I'm thinking, for instance, of a scenario where, instead of scientists having to input a search query into the PLoS ONE database, or indeed into Google Scholar or PubMed Central, software agents will be sent out to discover and interrogate papers. They would then analyse and manipulate the data in them, and deliver the results to researchers in the form of a computer-generated report; alerts if you like, but a step beyond today's notion of an alert.

CS: The Semantic Web makes a lot of sense in the context of PLoS ONE, and we are having lots of discussions about how to exploit the tools of the Semantic Web in the service.

Certainly we want to make PLoS ONE papers accessible by those sorts of routes. We also think it is important to try to establish links between papers. Scientific papers don’t exist in isolation but have a host of other papers related to them both closely and more distantly, by virtue of their content. We want to make those connections discoverable. So we are looking to provide lists of articles related to the article you are reading in PLoS ONE.

RP: This would be similar to the way that Amazon makes recommendations to its users would it — telling them what books other people have bought?

CS: That too. In addition to showing papers that are related because of the topic they are covering or the techniques they employ we want to be able to say "Other people who have read this paper also read these other papers." So we plan to incorporate a lot of different algorithms to help uncover the links and connections between papers.

Open Access 2.0

RP: PLoS ONE is very different to anything else that PLoS has done to date isn't it? While they utilise an open access business model, for instance, PLoS' journals are still subject-based publications.

CS: That's true.

RP: Indeed, PLoS ONE seems closer to the original concept of the Public Library of Science. The PLoS Open Letter published in 2001, for instance, stated: "We support the establishment of an online public library that would provide the full contents of the published record of research and scholarly discourse in medicine and the life sciences in a freely accessible, fully searchable, interlinked form." Does that not sound more like PLoS ONE than the PLoS journals?

CS: Of course. Personally, I don’t see the journals as an end point in themselves. In a sense, the journals are a demonstration that the open access model can work. Certainly PLoS ONE is an idea that we have been kicking around from the start, and it underlies the core reason for the Public Library of Science to exist — which is to get as much of the literature into an open access environment as possible.

RP: Perhaps we could say that PLoS ONE is stage two in the development of the Public Library of Science, and indeed of Open Access?

CS: In some ways. Certainly we see PLoS ONE as an example of what we are calling Open Access 2.0.

RP: This is a reference to Web 2.0?

CS: Yes. There is a lot of buzz around Web 2.0 applications; and that is the way we are looking at PLoS ONE. Essentially, we are providing a new synthesis of the things you can do on the Web to improve the dissemination of primary research.

RP: Web 2.0 implies greater interactivity and a more community-driven service I guess?

CS: Indeed. So when we talk about PLoS ONE as Open Access 2.0 we are saying that we see it very much as a community-up approach to science.

RP: Can you give me an example of how that might work?

CS: Absolutely. We have, for instance, been looking at tagging as a way of organising the literature — in the manner of sites like Flickr, which is creating a vast resource of pictures where users themselves apply tags, and provide structure to the information.

RP: Folksonomies.

CS: Yes. So we see folksonomies as a way of letting the community organise the papers, and we are looking at having the community of PLoS ONE users rate the papers. Consequently, rather than having to guess how good a paper is (I don’t like words like good but I will resort to it here) people will find that others have given them a status.

A distraction

RP: At the moment scientists and administrators tend to judge how good a paper is in terms of which journal it has been published in — by reference to the so-called Impact Factor. Might new methods like folksonomies eventually replace today's notion of the Impact Factor?

CS: Maybe. But the question of impact is in many ways a distraction. As you say, when scientists talk about the impact of a journal they tend to be thinking of Impact Factor, which is effectively the average number of citations that a paper in that journal gets. That is a useful measure for comparing journals with exactly the same scope, but it is a useless way to assess the importance of individual papers. The distribution of number of citations to individual papers is too broad and it isn't normal in shape.

RP: Scientists and their institutions, however, still attach considerable value to the Impact Factor. Given this, why would an author choose to publish in PLoS ONE when publishing in a traditional journal like Nature, Science, or Cell, provides much great kudos?

CS: The problem with the Impact Factor is that you are judging a paper by how good its neighbours are, and it tells you very little about the quality of an individual paper, since the Impact Factor is applied to the entire journal, not individual papers. It's strange that this is the only tool for assessing quality we have at the moment. If you think about it, it is like deciding whether someone is likely to be a bank robber on the basis of who their neighbours are!

PLoS Medicine has, by the way, just published an editorial explaining why we need to re-think the Impact Factor.

The point is that with PLoS ONE we want to maximise the true impact of every paper. By that we mean making them available to everyone who will be interested in their content, while providing efficient tools to allow readers to find the papers and for the papers to find them.

RP: Nevertheless is it not likely that — at least until people feel more confident about new ways of measuring the quality of a paper — PLoS ONE will be viewed as a second rate place to publish; the place to go if you cannot get published anywhere else?

CS: We definitely don't see PLoS ONE as a journal of last resort. It is an alternative to conventional journals, and so stands outside the hierarchy of journals that we have got so hung up on.

RP: So the appeal of PLoS ONE to authors will be what?

CS: PLoS ONE will be a great place to publish work that is so hot that everyone wants to talk about it, because that is exactly what can happen. Everyone can talk about it within PLoS ONE, and listen to what is being said.

It is also a great venue for the papers that only a handful of people wish to read. To that handful of people — those doing highly related work — such "minority interest" papers can be vitally important. Currently, however, their access is inhibited because the papers take too long to publish, and appear in obscure journals — many of which the interested reader's institution may well not have a subscription to in any case.

RP: So who exactly should be looking to publish in PLoS ONE?

CS: We want everyone to use PLoS ONE. We aren't looking at this as a competitor for "conventional" journals, but rather as an alternative to them.

So the issue for researchers is this: if you want to spend months revising and re-revising your paper to satisfy the demands of two or three experts who may or may not be the most appropriate people to judge your work (and so face a small chance of eventually being published), and if you are happy for your paper to be inaccessible to many of the readers who might wish to read it, then conventional journals are for you.

If, on the other hand, you want to share your results — as soon as possible — with the whole scientific community in a way designed to stimulate discourse and so scientific advance, then PLoS ONE will be your journal of choice.

Rates will be lower

RP: The business model utilised by PLoS journals requires authors to pay to publish their papers. Will PLoS ONE authors also be charged to publish?

CS: Yes. PLoS is a not-for-profit organisation but we still have to be self-sustaining.

RP: Will the article processing charges for PLoS ONE be the same as those applying to the journals, which I understand have just risen from $1,500, to between $2,000 and $2,500 per paper?

CS: Hopefully the rate can be lower. One of the driving forces of PLoS ONE is that we want to be able to publish lots of papers. To that end we are setting up the system in a completely scaleable way so that we can cope with as many papers as people want to publish with us. One of the advantages of doing so is that we can start getting economies of scale, and this will keep author fees as low as possible.

RP: So you expect author fees to be lower than the journals?

CS: We haven’t worked out a figure yet, but I am almost able to guarantee that the author fees will be less than the PLoS journals. This is because we expect PLoS ONE to be much less expensive to manage than the PLoS journals. And so it is fair that the journals should cost more.

RP: Why are the journals more expensive to manage?

CS: Because the PLoS journals require quite a lot of intensive editorial work in order to answer those subjective questions I talked about: establishing exactly how exciting a paper is, for instance, or how much of an immediate impact it will have. Essentially their task is to select the cream, although I hate to use that term. That means that both PLoS Biology and PLoS Medicine have a very low acceptance rate, and the selection process associated with that requires a lot of time-consuming work.

Faster and more efficient

RP: When will PLoS ONE be launched?

CS: We hope to launch at the end of the year. We have just announced the service, and we plan to start accepting submissions at the beginning of August.

RP: In conclusion then, what will PLoS ONE offer to science, and to researchers, that is not currently available?

CS: A faster and more efficient way of publishing papers. As I said, we are looking at an easier decision on whether to publish, and researchers won't have to find the right journal to publish in, since PLoS ONE has sufficient scope that it will be able to publish any paper that is worth publishing. This will also give us sufficient economies of scale to make it a far less expensive way to publish. Moreover, since we hope it will stop the problem of reviewers being asked to review the same paper for a number of different journals it should free up researchers' time.

The point is that the whole hierarchy of journals we have today is wasting huge amounts of scientists' time, whether they are authors, referees or readers. It is an inefficient system that needs replacing. PLoS ONE provides an alternative; one that will make science more efficient, and will see much more of the literature becoming available within an open access environment.

RP: OK, thank you for your time; and good luck with PLoS ONE.

Further information about PLoS ONE can be found here. A blog is also available here.

A more recent (March 2011) article about PLoS ONE is available here.

Monday, June 05, 2006

Interview with Harold Varmus

Today I am publishing an interview with Harold Varmus, Nobel laureate, president of the Memorial Sloan-Kettering Cancer Center, former director of the US National Institutes of Health, and co-founder of the open access publisher Public Library of Science.
This is number eight of The Basement Interviews, the introduction of which I am publishing on my blog. a The full interview (including introduction) is available as a downloadable PDF file (see below for details). It is being published under a Creative Commons licence.

Freeing the scientific literature

Harold Varmus, Nobel laureate, former director of the US National Institutes of Health, and co-founder of open access publisher Public Library of Science, talks to Richard Poynder.

Harold Varmus was born in 1939, on the south shore of Long Island, New York, a product, as he put it in an autobiographical note he wrote years later, of "the early twentieth century emigration of Eastern European Jewry to New York City and its environs."

His paternal grandfather, Jacob Varmus, had arrived in the New World from a small village near Warsaw just after the turn of the century, settling first as a farmer in Newburgh, New York, and subsequently working as a hatter in Newark, New Jersey. His paternal grandmother, Eleanor, was a victim of the influenza epidemic of 1918, dying when his father was eleven. Varmus' maternal grandparents — Harry and Regina Barasch — came originally from farming villages around Linz, Austria. Settling in Freeport, New York, they ran a children's clothing store.

Three years before Varmus' birth, his parents also settled in Freeport, where his father — who had trained as a doctor — established a general medical practice. From there Varmus' mother, Beatrice, commuted to a social services job in New York City.

A lake with alligators

When the United States entered the War, Varmus' father was assigned to an air force hospital near Winter Park, Florida, where the family resided until 1946. "My first memories," Varmus recalled later, "were to be of long beaches, and bass fishing on a lake with alligators." It was in Florida that Varmus' only sibling, Ellen Jane, was born.

After the family returned to New York, Varmus attended local public schools in Freeport; schools, he later characterised as being "dominated by athletics and rarely inspiring intellectually".

In 1957, after graduating from Freeport High School, Varmus entered Amherst College to prepare for medical school. His experience at Amherst, however led to a change of plan. As he put it: "The evident intensity and pleasure of academic life there challenged my presumptions about my future as a physician, and my course of study drifted from science to philosophy and finally to English literature."

He also became active in politics and journalism, and served as editor of the Amherst college newspaper. In 1961, after receiving an English degree — graduating magna cum laude — Varmus went to Harvard, courtesy of a Woodrow Wilson Fellowship.

Midway through his first year as an English literature postgraduate, however, Varmus dreamed that he had become an English professor and that he had missed a day of lecturing due to illness. Rather than being disappointed, however, his students were jubilant, pleased that their class had been cancelled. When he woke, it occurred to Varmus that if he were a doctor no one would be happy if he didn't show up for work — a thought that inspired him to return to his medical studies.

Varmus applied to study medicine at Harvard twice, but was rejected on both occasions. Concluding that he was not yet mature enough, the admissions committee recommended he first do two years army service. Undeterred, however, Varmus enrolled at Columbia College of Physicians and Surgeons (P&S).

Happy accident

Varmus began medical school with strong interests in psychiatry and international health. After serving an apprenticeship in a mission hospital in Bareilly, India, however, he drifted towards basic medical sciences.

In preparation for a career in academic medicine, Varmus worked as a medical house officer at Columbia-Presbyterian Hospital. In 1968, however, his plans for an academic career were interrupted by the Vietnam War.

In part to avoid the draft, Varmus applied for a research training post at the National Institutes of Health (NIH), where he was accepted for a clinical associate position in the laboratory of molecular biologist Ira Pastan.

At the time of Varmus' interview, Pastan was working on the hormonal influences on the thyroid gland. By the time Varmus turned up at the lab, however, Pastan's focus had switched to the study of the genetics of E.coli — an area with which Varmus was unfamiliar.

While Varmus was dispirited by this, it turned out to be a happy accident, since the change symbolised a revolution that was taking place in medical research. As James Fallows explains in a 1999 New Yorker profile of Varmus, the thrust of medical research was at that time beginning to shift "from the operations of organs, or the disorders of whole organisms, like human beings or mice, to the mechanics of cells, and within them, specific genes."

In other words, rather than trying to target individual illnesses, researchers were adopting a new gene-level approach to medical research. As Varmus explained to journalist Susan Stamberg, in a US National Public Radio interview in 1999, the future of medical research now lies in "the development of a notion of the gene as a physical entity that we can understand, manipulate, dissect and use to advance the great themes in medicine."

As luck would have it, Varmus' new post at NIH put him at the centre of this new approach just fifteen years after Watson and Crick had discovered the structure of DNA. And by taking advantage of the evening courses offered to incipient physician-scientists at NIH, Varmus was able to undertake further postdoctoral training in molecular biology. He later developed an interest in tumour viruses.

In 1969 Varmus moved to the University of California San Francisco (UCSF), where he began collaborating with microbiologist and immunologist Michael Bishop — conducting research on bacterial gene expression and tumour virology. Specifically, he studied the behaviour of retroviruses, including aspects of their unusual life cycle, the nature and origin of their transforming genes, and their potential to cause genetic change, not least their ability to cause cancer.

At the time, many scientists thought that retroviruses caused cancer by injecting their genes into the host's own genome. In 1975, however, Varmus and Bishop found that these viral genes had in fact been "captured" from the host animals in the first place.

Their discovery demonstrated that cancer genes (oncogenes) can evolve from normal cellular genes, called proto-oncogenes. In short, retroviruses can transform normal cells into cancerous ones.

Nobel Prize

This new insight meant that many of the cellular genes involved in cancer could be isolated. "It's been known for a long time that cancer was in some sense a genetic disease," Varmus explained to the Boston Globe in 1989. "The importance of our findings is that one can identify explicitly the genes that play a role in cancer."

The groundbreaking work done by Varmus and Bishop was recognised as a major contribution to current understanding, and enabled the scientific community to conduct an aggressive and successful search for the genetic origins of cancer. In 1989, therefore, they were jointly awarded a Nobel Prize "for their discovery of the cellular origin of retroviral oncogenes".

As a Nobel Laureate, Varmus began to feel a growing expectation that he become an ambassador for science, not least from former UCSF colleague Marc Kirschner — who recruited Varmus to serve on advisory groups about the future of research. In 1992 Varmus also joined a group of scientists supporting the Clinton-Gore ticket, and the following year he was confirmed as the new director of NIH.

Having never chaired a college department, or indeed run anything bigger than a 36-person laboratory, Varmus faced a significant challenge in his new job. He was now in charge of 16,000 people and a large bureaucratic organisation composed of 24 quasi-independent internal institutes. Being director of NIH also required winning financial support from Congress.

In the event, he proved highly successful, getting NIH funding up more substantially than any other category of federal spending — from $11bn to $16bn. He also succeeded in persuading Congress of the importance of basic research — particularly studying cellular-level functions — rather than simply granting money for crusades against high-profile diseases like prostate and breast cancer.

At NIH, argues Fallows, Varmus "achieved a series of political victories that will affect scientific policy for many years to come and, at least by implication, may change the entire understanding of health, disease, and the limits of the human life span."

When, in June 1999, Fallows asked him about the secrets of his political success Varmus replied by e-mail: "I have the sense that you wish my life had more drama. I don't. As I have said before, my strategy has been to try to minimise it."


Ironically at the very time he e-mailed Fallows, Varmus had just embarked on a project that would maximise the drama in his life, and spark a long-running controversy that has still to run its course. In short, his new project was to prove the catalyst for the creation of the Open Access Movement.

A few months earlier Varmus had been sitting in a San Francisco coffee house with Pat Brown, a Stanford University biochemist who had at one time collaborated with the Varmus Lab as a trainee, and remained a close colleague. Brown began telling Varmus about, the web site where physicists place pre-published versions of their papers for review, usually prior to publishing them in prestigious journals. This allowed physicists to speed up the research process, and maximise the impact of research papers by making them freely available to everyone over the Internet. Brown posed the question to Varmus: could not a similar initiative in biology provide the same benefits?

The question resonated for Varmus, and he spent the next few months pondering on it. The end result was E-BIOMED: "a proposal for electronic publications in the biomedical sciences", which Varmus distributed to the research community in May 1999. In effect, E-BIOMED envisaged a fundamental and very radical change to the way biomedical research is communicated.

The problem, however, was that Varmus' proposal implied a diminished role for science publishers, and the likelihood, therefore, that scholarly publishing would become a somewhat less profitable business. As such, E-BIOMED was greeted with a firestorm of protest and outrage — a response that NIH colleagues felt Varmus should have anticipated. "Varmus wrote the proposal himself and just sent it out to a bunch of people as a rough draft," comments one still puzzled NIH insider. "For us it was like 'Oh, my God: he clearly doesn’t know what he is getting into'.

"Sure enough," he adds, "the response from publishers was explosive. It caused all kinds of problems."

As Varmus now concedes, he certainly didn’t realise what he was getting into. Consequently, he was startled by the vehemence of the reaction from publishers, and surprised and angry when he discovered that, instead of engaging with him directly on the issue, many publishers had gone straight to Congress and complained directly to his appropriators.

Over the next few months there was a heated public (and private) debate. In the end, however, the strength of opposition from publishers was so great that when, eight months later, E-BIOMED was launched (re-branded as PubMed Central) it was a pale shadow of the revolutionary new "electronic publications" system that Varmus had envisioned.

Significantly, he had had to concede that publishers would have the final say on whether papers they published were placed in PubMed Central. And publisher response was derisory: even today, papers from only 251 of a possible 6,000 biomedical journals are made available on PubMed Central, and the content of some of these journals is only released after an embargo of between six and twelve months.

In retrospect, Varmus agrees that he was naïve not to have anticipated the furore. "I must have known that I was not going to be at NIH for much longer," he joked to New Scientist in 2003, "because this caused a tremendous political argument: what the hell was I trying to do to destroy the publication industry."

Indeed, Varmus left NIH within months of publishing the proposal, taking over as president of the Memorial Sloan-Kettering Cancer Center. Ironically, he says, his last public act at NIH was to sign the press release announcing the launch of PubMed Central.

Public Library of Science

By now, however, Varmus was firmly committed to the cause of Open Access (OA) — the more so, it seems, in light of the virulence of publisher opposition to E-BIOMED. In November 2000, therefore, Varmus co-founded — with Pat Brown and Mike Eisen — the Public Library of Science (PLoS).

Initially an advocacy group, PLoS was founded on the principle that if publishers were not prepared to act in the best interests of science, then the research community would have to twist their arms. And the first act of the founders was to invite fellow scientists to sign an open letter in which they pledged to boycott any journal publisher that did not make the papers it published freely available online within at least six months of publication.

PLoS attracted nearly 34,000 signatures from scientists in 180 countries; but while a small handful of publishers complied, most blithely ignored the PLoS letter. Worse, most of the scientist signatories were happy to disregard their own ultimatum, and continued publishing in the very journals that had turned a deaf ear to their request.

Unprepared to give up the cause of freeing the scientific literature, in 2001 Varmus and his colleagues reinvented PLoS as a not-for-profit open access publisher — inviting scientists to pay to publish their research in new open access journals, on the basis that in doing so they could ensure that their research was made freely available on the Web from the date of publication. A number of PLoS journals were subsequently launched, and more are planned.

To Varmus' gratification, the journals have attracted some first-rate submissions, and quickly acquired a reputation as high-quality publications. What is less certain, however, is whether the author-pays publishing model they utilise is sustainable in the long term.

In many ways, Varmus has taken his role as an ambassador for science to its logical conclusion. After all, if the effectiveness of scientific research can be maximised by disseminating it in the most efficient way possible, then any ambassador worth his salt would surely do everything he could to make that happen?

Varmus' ambassadorial activities have not been confined to the US either. In March 2004, for instance, he appeared as a witness for the UK Select Committee enquiry into scientific publishing, which subsequently recommended that the UK government mandate all publicly-funded researchers to make their papers available on the Web.

Undoubtedly, Varmus will also have been involved in the behind-the-scenes lobbying that, in May 2005, saw the introduction of the NIH OA Policy, and the introduction into the US Senate in December 2005 of the CURES Act of 2005. No doubt he played a part in lobbying for the recently announced Federal Research Public Access Act of 2006 (FRPAA) too.


What sort of man is Varmus? Describing him as a "lean, energetic, and intense presence", Fallows pointed out that — for someone in such a public position — Varmus managed to maintain a surprisingly low profile at NIH. Indeed, his colleagues at the time often referred to him as "the invisible administrator."

So successfully did he maintain public invisibility in fact that when, in 1996, Varmus gave the commencement address at Harvard, the university newspaper, Crimson, reported widespread disappointment that an apparent nobody had been invited to speak.

Why then seek to play a leading role in such a controversial — and increasingly high-profile — movement as Open Access? "I believe that science is one of those activities that improves the state of the world," replies Varmus, "and once you realise how important publication is in the series of acts that constitutes the doing of science, and once you understand the incredible transformation of that publication process that the Internet, and software, and the whole digital world, now promises it is hard not to be pretty passionate about trying to make that part of the scientific universe work more effectively."

How significant a role has Varmus played in the Open Access Movement then? Critics often claim that his naivety has hindered more than helped. For instance, they argue, the E-BIOMED proposal only encouraged publishers to dig their heels in and resist change.

Varmus' supporters, however, contend that, like the music and movie companies, science publishers would inevitably have done everything they could to resist the impact of the Internet — since the new distribution models it permits pose a significant threat to their profits. Moreover, they add, as a Nobel Laureate and former NIH director, Varmus has provided the movement with an authority and credibility that it would otherwise have seriously lacked.

Importantly, comments the NIH insider, once alerted to the issue Varmus was immediately able to appreciate what was required, and then took personal responsibility for the task of making it happen. "Harold deserves credit for seeing that this was a real problem that needed to be solved, because most institute directors, and most scientists, just don’t have the insight."

Certainly Varmus is now widely viewed as a key figure in the OA Movement. In 2004 he and PLoS co-founders Eisen and Brown were given a Wired Rave award "For cracking the spine of the science cartel".

And in a recent profile of him, the magazine described Varmus as being "the prophet of open access." A prophet, moreover, who has become locked into a real-life struggle "that has turned this icon of the scientific establishment into a powerful subversive."

For all that, Varmus has a somewhat parochial view about Open Access, and makes no claims to understand the wider free knowledge movement, or even to particularly care about the role of Open Access within that wider movement, although he does support the principles of Creative Commons, whose licences PLoS utilises for its journals.

This disinterest in the bigger picture is a little surprising perhaps, because there is undoubtedly a commonality in origin, and in purpose, between the Open Access Movement and the various other free and open movements. If nothing else, they are all creative responses to the possibilities of the Internet, not least its ability to enable much greater collaboration and information sharing.

What is clear is that, in the same way that large and complex software projects now cry out for an open source approach to development, gene-level medical research requires an increasing degree of openness to be effective. The greater need for "open data" was graphically demonstrated, after all, by the huge efforts that NIH put into competing with privately-held Celera to sequence the human genome (by means of the Human Genome Project, or HGP). The fear was that vital biological information would be privatised if Celera became the only game in town, an outcome that would have obstructed scientific research.

In an online environment research papers will also increasingly be viewed as inseparable from the data on which their conclusions are based. And if both the data and the evaluation of that data need to be more widely and freely accessible, then Open Access will have to be treated as part of a larger and broader development.

Likewise, the new-style licensing models being developed by organisations like Creative Commons will become an ever more important component of the infrastructure that will be necessary if the various free and open movements are to flourish; and many of the new software tools being developed to support these movements will undoubtedly need to be based on Open Source Software.

In short, Open Access, HGP, Open Data, the Free and Open Source Software movements, Creative Commons, and the many other free and open movements, are all logical adaptations to the realities of today's increasingly complex knowledge environment. More importantly, they all face a common threat: the threat that, rather than being made more freely available, information could be increasingly privatised.

Excessively draconian use of intellectual property laws, for instance, could significantly hamper these movements, and hinder the improved collaboration, more rapid development, and greater creativity — be it artistic or scientific — that today's digital world promises. The ability of scholarly publishers to resist OA, after all, is based on their insistance that, as a condition of publication, they are given copyright in the papers they publish, thereby acquiring ownership of them. By working together the various free and open movements could therefore surely achieve their separate aims more effectively.

Interest in the arts

But it would be wrong to suggest that Varmus' vision of the world is in any way narrow. Despite his evident passion for science, Varmus retains a strong interest in the arts. In the talk he gave at his Nobel Prize ceremony, for instance, he included lines from Beowulf, which he recited in Anglo-Saxon. He has also written film reviews for the New York Times, and invested in Broadway plays. And when he visits foreign cities he often proposes meeting up with associates and friends in art galleries.

Varmus is known for his casual clothes — often khakis and rumbled open-necked shirts — and is a keen sportsman who loves outdoor sports, particularly bicycling, running, backpacking, skiing, and fishing.

In 1969, he married Constance Louise Casey, then a reporter for Congressional Quarterly in her home town of Washington, DC. They have two sons: Jacob and Christopher. Varmus' younger sister now works as a genetic counsellor in California, and his brother-in-law is the novelist John Casey.


Although this interview was conducted by telephone, I had seen Varmus in the flesh when he gave evidence to the UK Science & Technology Select Committee. A tall, slim man — with, as Jamie Shreeve put it in his Wired profile of Varmus, "sparse, reddish brown hair" — he sat in a slightly hunched position throughout, leaning forward on to the desk as he explained the merits of Open Access to British politicians.

Indeed, sitting next to BioMed Central founder Vitek Tracz, Varmus presented a striking contrast. A small restless man, Tracz spoke quickly, and at some length, causing the chairman at one point to ask him "to keep the answers a bit shorter and answer the question". Varmus, by contrast, delivered his views briefly and precisely.

And where Tracz was full of nervous energy, Varmus exuded a quiet confidence and unflappability. He also revealed a consummate skill for flattering politicians, a skill no doubt honed at NIH. His final words to the Select Committee were: "It has been a pleasure for me to see this high-minded conversation. I appreciate it greatly."

While entirely courteous, Varmus was less ingratiating with me. When I e-mailed him to ask for an interview I heard nothing for several weeks. Eventually his secretary invited me to suggest some dates, and asked how much time I needed for the interview. I requested two to three hours: I was offered 60 to 90 minutes.

The week before the interview was scheduled I was asked to send a list of questions so that Varmus could review them over the weekend. On the day of the interview, however, he declined to answer any questions not directly related to Open Access, arguing that in the time he had allotted for our conversation he was not able to address any other topic.

And even though we finished our discussion before our scheduled time was up, Varmus still declined to answer questions on any topic other than Open Access. He suggested that if I wanted biographical information I could use what was already in the public domain.

He also appeared unusally controlling about the text of the interview, asking several times to have sight of it before publication (When however I later e-mailed the text — to both Varmus and his secretary — I received no feedback at all). At one point in the interview he complained that I was talking over him.

All in all, when I put down the phone I knew very little more about the kind of person Varmus is than before I picked it up — other perhaps than that he seems a very focused person, and apparently keen to protect his privacy.

So let me leave the last words to the NIH insider, who spoke to me on condition of anonymity. "Harold Varmus", he said, "is a visionary who has very little feel for how people are going to react to what he says and does. But he isn’t generally concerned about that. He is also a man without guile."


If you wish to read this interview in its entirety please click on the link below. I am publishing it under a Creative Commons licence, so you are free to copy and distribute it as you wish, so long as you credit me as the author, do not alter or transform the text, and do not use it for any commercial purpose.

If after reading it you feel it is well done you might like to consider making a small contribution to my PayPal account. I have in mind a figure of $8, but whatever anyone felt inspired to contribute would be fine by me. Payment can be made quite simply by quoting the e-mail account: It is not necessary to have a PayPal account to make a payment.

What I would ask is that if you point anyone else to the article then you consider directing them to this post, rather than directly to the PDF file itself.

If you would like to republish the article on a commercial basis, or have any comments on it, please email me at

I would like to acknowledge the help of the
Open Society Institute, which provided a small upfront grant to enable me to get started on The Basement Interviews project. Further information about The Basement Interviews can be found at the Open and Shut? site.

To read the interview in its entirety (as a PDF file) click