BERKELEY - Scientific publishing has hit a digital divide. On one side of the chasm are the leading journals, such as Science, JAMA (the Journal of the American Medical Association) and The New England Journal of Medicine, which restrict access to their content to individual or institutional subscribers. On the other side are the majority of scientists and some not-for-profit journals, which believe that papers describing the results of publicly funded research should be available to anyone, anywhere, at no cost, soon after publication.

Longstanding reputations and healthy bankrolls may allow the most prestigious of the journals to fend off calls for free access in the near term. But what about their midlevel competitors: journals that are highly specialized, carry hefty subscription prices and draw small audiences? How will they stay afloat if new electronic distribution models become the norm — models like PubMed Central, a free, full-text biomedical archive founded by the National Institutes of Health (NIH), or the Public Library of Science, an advocacy group turned scientific publisher that will make all research articles it publishes immediately and freely available online.

UC Berkeley biochemist Nicholas Cozzarelli, editor-in-chief for seven years of the Proceedings of the National Academy of Sciences (PNAS), discussed some of the challenges and responsibilities facing scientific journals in the new world of electronic publishing. He was joined by his colleague, Michael Eisen, a molecular biologist at the Lawrence Berkeley National Laboratory, who helped spearhead the creation of the Public Library of Science.

Why is there so much resistance among scientific publishers to making articles available free of cost online?

Biochemist Nicholas Cozzarelli, editor-in-chief of the Proceedings of the National Academy of Sciences, in his Koshland Hall lab. Noah Berger photo

Cozzarelli: Scientific publishing can be extremely lucrative and many publishers are commercial outfits. People make hundreds of millions of dollars in the business, so there are enormous sums of money coming in to the commercial publishers. Many of the noncommercial publishers run by societies have used journals as cash cows. They don’t spend their profits on yachts, but the money is used for things like keeping the society going, running the annual meetings, maintaining a headquarters, and helping out the membership of that society. I don’t think that is a good idea in the long run. It seems to me that the cost of the journal should reflect what it costs to produce it and nothing more.

With my own journal, the Proceedings of the National Academy of Sciences, we can’t make a profit. We are a break-even operation, required every year to have a zero balance. If we make any profit, we have to give it back. I’m not suggesting that every journal be run this way, but I would love to see all journals run with a modicum of profit, so that they have enough flexibility to keep things going, but are spending most of their money solely on scientific publishing.

So the main concern among journals is economic survival, not making huge profits?

Cozzarelli: There’s more to it than that. The economic argument is overblown. You hear a lot about that. But there’s a lot of inertia in the system. The journals are large organizations, they have bureaucracies, and they’ve been operating a certain way for a long way. They don’t want to change.

Eisen: That’s true. There’s very little economically that argues for maintaining the current publication system, except inertia. The Public Library of Science proposes to change the way we pay to cover the costs involved in publishing. In the current economic model for scientific publishing, revenues come mainly from subscriptions purchased by individuals and institutions. This system necessitates restricted access, because who would pay to access something that is freely available? We believe that it is far more sensible for the costs involved in publishing the results of scientific research to be borne by the groups — universities, government agencies, foundations, et cetera — that funded the research itself, viewing the costs of publication as a final, integral cost of the research process. Money that is currently used to pay for journal subscriptions would simply be diverted to pay the costs of publishing at the time of publication — the costs of production, and even some profit, would be covered, but there would no longer be any need to restrict access to the published work.

Then what would the major source of revenue be for new journals, like the Public Library of Science, that aren’t operating on a subscription basis?

Cozzarelli: Really, it’s a three-fold issue. First of all, we have to have foundation support. This is a new venture, and there are some foundations that are interested in providing seed money to get a new concept journal going. The major source of revenue in the future will be author revenues. The authors will pay to have their work published, like a reverse royalty. The articles will still be peer-reviewed, and I’ve emphasized that it should be a very high-quality peer review. It should be a really first-rate journal. There will be a third source of revenue, which has yet to be worked out completely. There are ancillary sources of revenue in all of publishing, and the hope in the future is that some of these will become more important.

Eisen: You have to remember that electronic distribution costs close to nothing, so all the costs involved are incurred in producing the first copy [of the article]. A lot of money is already spent to support scientific publishing, about $5,000 to $10,000 per published article, and much of this money already comes from NIH and other agencies that support scientific research. If the costs of producing the original first copy of an article are included in the research grant, those costs would be minimized and this wouldn’t be such a controversial issue.

When did this issue really start heating up?

Cozzarelli: The Public Library of Science initiative got it rolling and it’s died down a bit since then. Two or three years ago, they issued a challenge to the scientific journals, saying that the journals were to give away their content within six months of publication or face a boycott. They got 30,000 signatories to support a boycott. As it turned out, it scared the publishers so much that some of them came into compliance or near compliance, and so, in that sense, it was wonderfully effective as a threat.

Eisen: We [the Public Library of Science] formed a pressure group to urge journals to adapt new policies for electronic publication of science results. We wanted journals to allow their content to be redistributed in electronic archives, and we got overwhelming support from scientists. But very few journals have warmly embraced this idea.

What role has Proceedings of the National Academy of Sciences played in recent years to make journal articles available free of charge on the Internet?

Cozzarelli: I felt that PNAS should take a leadership role in scientific publishing. And this has been done, and done so easily, because I’ve had the complete support of the academy. We were the first major journal to join PubMed Central, an initiative that was set up by Harold Varmus when he was head of the National Institutes of Health, and by Pat Brown [of Stanford University] and David Lipman [director of NIH’s Center for Biotechnology Information, which operates PubMed]. The PubMed Central initiative was aimed at making science available worldwide. The federal government, through the National Library of Medicine, is paying a lot of the costs. They release the science for you online and archive it, which is a difficult thing, but they’re obviously in the archiving business. I’m a member of the PubMed Advisory Board and I am very proud of the move of PNAS to PubMed Central, because I wanted to make a statement to the scientific community … about their responsibility to release results to the world. PNAS went online free to everyone, and we started out making articles available four weeks after the journal had been published. That was probably a little too gutsy on my part, so we changed that and went to six months after articles have appeared in print. We encouraged other journals to do the same, and some have.

Have you influenced PNAS in other ways during your tenure?

Cozzarelli: One procedural change that I instituted had to do with a nutty tradition that had developed in science, whereby structural biologists were allowed to publish papers on their work, but they didn’t have to tell anyone what the data were until a year later. The idea behind that policy was that they would protect themselves and their data … but PNAS and some other journals, and some granting agencies and a few leading structural biologists got together and told the structural biologists that they weren’t going to follow that practice anymore. If they were going to publish, they would have to publish their coordinates [scientific data]. … PNAS wasn’t the only journal to institute this change, but we were the major journal to do it. Now only the rarest of journals still uses that nutty system.

I also instituted a new track for articles. In the past there were only two ways to get an article in the journal: if you were an academy member, all you had to do was submit it … without a review … and if you were not an academy member, it was reviewed. I made two changes: the first is that articles by academy members must also be reviewed. The second change, track two, is that articles are sent to the PNAS office and then to the editorial board, and they decide whether the article will go on for further review. If it does, they find the appropriate members of the academy to conduct the review — scientists who specialize in that particular discipline — then it goes through a regular peer review process. So it’s much more like an ordinary journal. Scientists no longer feel like they have to ask a favor of someone in the academy to submit a paper.

How did interjournal linking come about?

Cozzarelli: We took a tough stance on the issue of interjournal linking. PNAS got together with the Journal of Biological Chemistry and a few others and decided that we would share interjournal links among ourselves, so that if anybody ever referred to an article that was from this initial group of publications, the reader would be able to click on that article and get it free. You didn’t have to have a subscription. It was an important move in taking advantage of the new possibilities of online publication and making science more easily accessible. It was another change that spread so well around the scientific community. Today interjournal linking is routine.

Is there much demand for access to science articles online among Third World countries?

Cozzarelli: Yes, there is some, and we give PNAS away to Third World countries. Once I calculated how much we would lose if we did that, and it was chicken feed in terms of the subscriptions for developing countries. For starters, there just aren’t that many scientists with paid subscriptions. It cost us a little money for electronic publication — it isn’t absolutely free — but that was only something like $20,000 a year. Our budget is around $8 million to $9 million a year, so what’s $20,000? Our president, Bruce Alberts, is very much interested in science being an important, democratizing influence in the developing world. If the National Academy of Sciences doesn’t do it, who will? If we aren’t for science, then who is?