preLights talks to John Inglis

John Inglis is Executive Director and Publisher of Cold Spring Harbor Laboratory Press (CSHL Press) and co-founder and PI of bioRxiv and medRxiv. Prior to moving to CSHL in 1987, he worked as an editor at The Lancet before founding the monthly review journal Immunology Today (now Trends in Immunology). We were delighted to speak with John about his career, the development of bioRxiv, and the changes that preprints continue to make across the science publishing landscape.

We always ask interviewees how they became interested in science – could you tell us where that interest came from and how you came to pursue a PhD in Immunology?

I grew up in Scotland at a time when the liberal education system gave us the qualifications at 18 to study more or less anything at university. I was drawn to journalism and medicine but with some hesitation I embarked on a science degree. I hadn’t really done much biology, but I discovered that I really liked it. And in the course of that, I discovered the immune system, which I thought was the most amazing thing that evolution has created. In the final year of my zoology degree, I got the chance to do an immunology research project with someone in the medical school, which went well. I wasn’t sure about pursuing a PhD, so I worked as a research associate in the veterinary school in Edinburgh, then decided to register for a PhD in the medical school with a clinician who was a pioneer of human autoimmunity.

And after your PhD you moved into an editorial role?

My PhD went well but at the end of the process I wasn’t convinced that doing more research was for me. Immunology was exploding at the time, and there were so many interesting things going on that the idea of narrowing my focus for a postdoc didn’t really appeal. I discovered an advertisement for an assistant editor position at The Lancet, which I applied for and was offered, and that pitched me into the world of science publishing and communication. It was a fantastic experience, working in a very small staff of amazingly knowledgeable people, and I learnt a lot in the three years I was there.

Then in another piece of good fortune I got the chance to invent a monthly review journal called Immunology Today (now rebranded as Trends in Immunology). Again, that was a fantastic experience, an opportunity to capture all the theories and model building and ideas and new technologies of the time. I started another journal, in parasitology, and had another one in the pipeline, when I was contacted by James Watson who invited to come to CSHL to build its small but prestigious book publishing department into a kind of University Press. I agreed, thinking it would just be for a couple of years, but it’s been a continuously interesting and evolving process: we launched CSHL Press in 1988, and now we have nine journals and have published hundreds of books, and we’re continuously growing our enterprise and expanding our interests. So, here I still am, after 35 years!

Which brings us to bioRxiv, the preprint server you co-founded at CSHL in 2013. Where did the idea for bioRxiv come from?

For over a century, CSHL has been uniquely both a research institution and an important locus of scientific communication and data sharing. That started with conferences, then courses, then publishing books and manuals and journals. So a preprint server from CSHL is really part of that continuum of activities devoted to sharing science, and we had tremendous and instant support from the President of the lab, Bruce Stillman, to initiate bioRxiv.

Of course, the reason it’s called bioRxiv is because of arXiv, which began in 1991 and has become central to communication in the physics community. I had never heard of arXiv until publishers started putting journals online in the mid-90s. This came with huge challenges, so we got together as never before to share experiences. Through the physics publishers I learnt about arXiv, and then at a meeting held at CSHL I met Paul Ginsparg, its founder. The meeting was organised by genome scientists who even early in The Human Genome Project were already thinking about open data and open science. When Paul spoke about arXiv I thought the concept was fantastic but had no chance of working in biology. That was in 1998, but in the years following, there were various attempts to coopt the arXiv idea – one at NIH, one at Nature, one at the BMJ. They didn’t take root. But by 2012 we had a hunch that it was time for another go, and that CSHL would be a great and appropriate place to launch a not-for-profit preprint platform for biomedical research. The sharing culture of genomics was gaining traction. And being assisted by younger scientists  –  digital natives who were now postdocs and PIs and didn’t see any difference between sharing their science online as they shared other aspects of their lives. So, while we had ambitions to make it a big project, we felt it would still be a useful contribution even if it didn’t go beyond genetics and genomics. And with that, we launched in November 2013.

How important was community support at that stage? Did you come up against a lot of resistance from researchers?

An important part of the launch was talking to lots of scientists in different fields. One group in particular, a new group on the faculty here at CSHL, was the quantitative biologists, people with backgrounds in maths and physics who were asking what the bio-equivalent of arXiv was for the projects they were now working on.

There were certainly concerns, too, about the idea of distributing scientific information before it had been peer reviewed. But these concerns were field, age, and demographic dependent. On the whole, younger people seemed to have very little problem with it, but more senior folks were more hesitant. Another important group we spoke to were people steeped in journals  – officers of societies that published journals and their editors. We wanted them to understand that we felt what we were doing was complementary to journal publishing, not an attack on it. There were some high emotions in certain quarters, but I must say that society publishers as a whole were hugely supportive. It was important for us to be able to reassure authors that journals were supportive of preprints so their papers could still be published after being posted as preprints. And in the past 8 years we’ve seen how most biomedical research journals have changed their policies and made it explicit that they will consider manuscripts posted on bioRxiv.

Was launching medRxiv a similar process?

We briefly considered whether bioRxiv should have the scope to embrace clinical medicine but quickly decided that it shouldn’t because of anxieties around medical information. But in 2016, there was an op-ed published in The New York Times by two high-profile clinicians saying there should be something equivalent to bioRxiv for medicine. We began a collaboration with experienced clinicians at Yale and experts in medical publishing at BMJ to devise ways of bringing preprints to medicine responsibly, and we launched medRiv in June 2019. But of course the pandemic began 6 months later and medRxiv immediately became a major source of new information about COVID-19.  So instead of having a slow on-ramp, medRxiv burst into the community’s awareness and I feel that suddenness has tended to prod people into pro or con positions about clinical preprints. The pandemic really robbed us of a period of slow acculturation. But many medical journals including the New England Journal of Medicine and JAMA have become perfectly ok with manuscripts that have been preprinted. I think it has helped that we’ve been careful to build in screening procedures and author requirements that are generally seen as responsible when distributing medically relevant information before peer review.

Which leads me to ask about your bioRxiv and medRxiv affiliates. How did you originally recruit affiliates, and how do you maintain and update your board of affiliates as bioRxiv and medRxiv grow?

This is an ongoing initiative. We decided right from the beginning that we would do in-house screening of manuscripts, and we have a fantastic group of former postdocs as our content team, but we felt that the last word on a manuscript should come from an independent principal investigator. The first affiliate recruits came from the scientists we spoke to in the preliminary stages of creating bioRxiv, but as the volume has grown, so has the number and now we’re up to about 180. If we feel we need more we ask the current group to recommend other people, and then there’s an interview to make sure everyone is on the same page about the process concerned. This has been very effective, and we’re immensely grateful to this increasingly large group. I think we need to be more diverse, geographically and in other ways, but they cover a wide range of research expertise and they have been super helpful with advice on manuscripts, particularly related to the pandemic.

You mentioned digital natives and the idea of freely sharing information. Do you view bioRxiv and Twitter as being quite synergistic? Did you plan for the use of social media in sharing preprints?

In 2013 there were not that many scientists on Twitter. The scientific Twitterverse has grown at the same time as bioRxiv and there’s certainly been synergy that has been extremely helpful. We auto-tweet every paper that is posted to bioRxiv and medRxiv, and some authors do amazing tweetorials to explain the work they’ve posted. All kinds of conversations take place around preprints and yes, Twitter has been important – no doubt about that.  We have a  dashboard on bioRxiv designed to bring together all the conversations around a particular preprint, including tweets but also reviews and commentary.

The pandemic saw a huge acceleration in the uptake of preprinting. How did you manage the influx of preprints posted on bioRxiv and medRxiv?

We received generous support from CZI for bioRxiv in 2017. We launched medRxiv with funds from the Laboratory in June 2019  and medRxiv was expected to be a modest amount of additional work for the content team, which was quite small at that point. Suddenly, in the second half of January 2020, 30 papers were posted to bioRxiv about the novel coronavirus. And then things went crazy – medRxiv began getting an enormous number of pandemic-related submissions and in May we posted 2,000 manuscripts, with a 40% rejection rate. That gives you some sense of the volume of material we were receiving and the pressure the content team was under at that time. Fortunately, CZI reached out to offer assistance and we were very fortunate to get additional funding for medRxiv that allowed us to expand the team. There has been a decline in the intensity of pandemic-related preprint submissions since then, but we’re still posting hundreds of them on both servers.

And that’s led to a shift in the way the research community views preprints, so now we’re seeing a bigger push for preprint reviews and commenting. Do you think that’s something we should encourage in day-to-day science?

I think so. Research is a constant process of assessment: looking at literature, critical thinking, and constant re-evaluation, that’s all baked into the everyday of science. Preprints are still relatively new, so people have to expand their information habits to include getting information this way. And with the volume= of preprints on bioRxiv and medRxiv, it’s hard just to keep up, without thinking about having to contribute a response. But as well as the time factor there are also cultural issues: we don’t have a community where we can completely openly and comfortably share critical thinking about a given piece of work. We may do it with our friends, in a research group, or in our department, but doing so in a broader context is something there’s still anxiety about. We hear this particularly from younger scientists who feel they might have something to lose if they are openly and identifiably critical of work published by other groups – especially groups run by senior scientists. The same anxiety attaches to signed peer reviews, so there’s frequent discussion about the benefits and downsides of signed public peer reviews versus the problems associated with anonymity. I don’t know what the correct answer to that is. But I do think that science will benefit if we foster more conversations around preprints, whether they are anonymous or not, whether it’s formal peer review or a news and views type discussion as it is with preLights. All of that is valuable. We’re still in the early days of preprint review projects, so we don’t know what will emerge or what will be successful, but it’s inspiring to see these experiments being done with sincere commitment and great intentions.

Expanding on the culture of preprints – we still see that topic-specific uptake and geographical uptake varies greatly. What can we do to expand preprints into the Global South?

arXiv began in high energy physics, and gradually percolated through other branches of physics – and that’s exactly what we saw with bioRxiv. As we predicted, there was quick uptake in genomics and bioinformatics, but now the biggest category is neuroscience. Other biomedical disciplines are still catching up and we are planning to increase our advocacy efforts in those fields.

Both bioRxiv and medRxiv have preprints from 150 countries but there’s no question there is less uptake in the Global South than in other areas of the globe. That might be linked to concerns about scooping, as there’s still a sharp focus on peer reviewed literature as a means of career advancement in many communities. Another thing that hasn’t changed as much in local communities is the perception of double publishing when a journal discovers that a manuscript is already posted as a preprint. We still sometimes hear from anguished authors who say they can’t get their manuscript published in a local journal because they’ve posted it as a preprint on bioRxiv. We always try to engage with the journal concerned to explain what we are doing and why, and often that’s successful. But we could do more, and should do more, to make the case in these communities, and I think one way to do this in the future is by expanding our affiliate pools to include leaders from communities where preprinting is less prevalent. They could to help argue for the benefits. I think this situation will improve with time but we also have to put more energy and resources into accelerating uptake in the Global South.

And to finish on a similar note. The preLights community is predominantly made up of early-career researchers (ECRs) who advocate for preprinting. What advice would you give to students or postdocs who would like to persuade their PIs or advisors to preprint their work?

If a postdoc is getting resistance from a PI, it’s going to take conversation and evidence and data to support the case for a preprint.  There are broad general arguments about why preprints are helpful, especially to ECRs as evidence of productivity and about their role in accelerating the pace of research. But it might help to point out to an anxious PI that competing labs are likely posting to bioRxiv. So there’s a community-based argument, but also a one based on self-interest. Judging from my interactions with faculty here at CSHL, PIs are generally very supportive of what their postdocs or students want to do with their work, so if you provide facts and rationale, I believe that will have a beneficial and persuasive effect.

John was interviewed by Helen Robertson, Community Manager for preLights. This interview has been approved by the interviewee.