Four years of chemistry preprints

In summer 2017, the world of chemistry publishing was quietly revolutionised – by the launch of two preprint servers. Preprints are early versions of scientific papers that haven’t yet been through the peer review process or been published in a traditional journal, and preprint servers are the free, online archives (or repositories) that host them.

Preprint servers are not a new concept. The first, arXiv, was established in 1991, for theoretical physicists. Three years later, SSRN launched for the social sciences. Over the decades, both preprint servers have significantly broadened their scope and many other servers have launched. Some of these have chemistry in their scope, and there was an even a chemistry-focused preprint server for a while in the early 2000s.

Then, in August 2017, two things happened. First, the SSRN preprint server (originally for social sciences and currently run by publishing giant Elsevier) launched an offshoot called the Chemistry Research Network (ChemRN). Then, a week or so later, ChemRxiv was launched by the American Chemical Society (ACS) – it is now run as a collaboration between the ACS, the Royal Society of Chemistry and the chemical societies of China, Japan and Germany.

Everybody realised that preprints are like a moving train, you’ve got to get onboard or you’ll be left behind

Donna Blackmond, Scripps Research Institute

Neither was initially welcomed with wholly open arms into the chemistry community. An insistence – especially from some journal editors – that depositing work caused it to lose the required novelty to warrant publication in their journals was the biggest issue, explains Donna Blackmond, a professor of chemistry at the Scripps Research Institute in California, US, and a member of the ChemRxiv scientific advisory board.

At the start, some chemistry journals accepted submissions that had been preprinted, some did not and many others left authors guessing by not having a preprint policy at all. It took about year for all the chemistry journals to accept preprints, says Blackmond. The ACS flagship journal, the Journal of the American Chemical Society, held out the longest. ‘Everybody realised that preprints are like a moving train, you’ve got to get onboard or you’ll be left behind,’ she adds.

Once journal editors had given them the green light, confidence in the use of preprint servers amongst chemists soared. ‘The journals in chemistry accepting manuscript submissions that were already posted as preprint caused a big [mindset] change,’ says François-Xavier Coudert, a computational chemist at the French National Centre for Scientific Research in Paris and a member of the ChemRxiv scientific advisory board.

Four years after their launch, the preprint servers are now finding their feet in the chemistry community. In 2020, 5137 preprints were posted on ChemRxiv and 3538 on ChemRN. The same year, ChemRxiv preprints were accessed a total of 16,120,921 times and ChemRN pre-prints downloaded 499,553 times.

Rapid dissemination

During the pandemic, preprints have served a vital role in rapidly and openly disseminating research results. Before Covid-19 was even detected in Europe, the London-based Wellcome Trust had coordinated an open statement calling for all research findings and data relevant to the Covid-19 outbreak to be ‘shared rapidly and openly to inform the public health response and help save lives’. The statement was signed by dozens of funders, societies, publishers and institutions, and preprint servers were listed as a way to facilitate this.

Chemistry preprint servers saw less of a pandemic submissions boost than those in biology and medicine, but coronavirus-related content still takes eight of the top ten places in the ChemRxiv’s most accessed content of all time list.

When I watch the news and see someone talk about a Covid preprint, I’m often thinking “Come on, you know this could be bullshit”

Hosea Nelson, California Institute of Technology

The most accessed preprint in May 2021 (and the second most accessed preprint in April) was Vittorio Saggiomo’s protocol for turning an aluminium coffee pod into a miniature chemical reactor that can host a reaction that reports the presence of Sars-CoV-2 by colour change in less than 30 minutes. Saggiomo, an assistant professor in the bionanotechnology group at Wageningen University and Research in the Netherlands, doesn’t know if the ‘CoronaEspresso’ protocol has been used for its intended purpose yet. Regardless, it was a ChemRxiv hit. By the end of July 2021, the preprint had been accessed nearly 21,000 times. The preprint was also highlighted in 19 news outlets, on four blogs and was tweeted about 1044 times, according to Altmetric data.

The CoronaEspresso is just one of hundreds, if not thousands, of coronavirus-related preprints to make the news since the start of the pandemic. This hasn’t been without its problems. Not every news presenter or journalist treats preprints with the extra grain of salt required compared to peer-reviewed content. ‘When I watch the news and see someone talk about a Covid preprint, I’m often thinking “Come on, you know this could be bullshit”,’ says Hosea Nelson, an organic chemistry professor at the California Institute of Technology in the US.

Peer-review is certainly not perfect, but more often than not it acts as an effective gatekeeping process for keeping nonsense out of the scientific literature. Most preprint servers have basic screening processes in place to try and catch the worst of the submitted contents to stop the spread of misinformation. ChemRxiv sends all submission to a team of curating academics for a quick check before they are posted online. ‘It’s a basic evaluation of the appropriateness of the content to be posted with us. Our team checks that each submission fits our scope and doesn’t include anything inflammatory or potentially hazardous,’ says Ben Mudrak, senior product manager for ChemRxiv.

ChemRN does similar, running all preprints past Elsevier journal editors for a sense-check before posting. ‘Somebody quickly reviews each of the preprints that goes out to make sure nothing crazy has slipped through the cracks,’ says Gregory Gordon, SSRN managing director.

Rapid feedback

Another frequently expressed reason for posting preprints is to get early community input into research. ‘Preprints are a great way to get early feedback on ideas and when we talk to authors that’s one of the biggest ways that they see the utility in preprints,’ says Mudrak.

After feedback on his Covid-19 testing preprint, Saggiomo revised the text to clarify some points before submitting it to a traditional peer-reviewed journal. This updated text is also available on ChemRxiv. ‘When we are working on the same projects for months or years, adding the fresh eyes of not only two referees but hundreds of people reading the preprint I find extremely useful,’ Saggiomo explains.

Naresh Patwari, a professor of physical chemistry at the Indian Institute of Technology Bombay, has also benefited from community feedback on a preprint. In October 2020, he posted a preprint detailing a physical chemistry study into how van der Waals complexes behave after photoexcitation in arXiv. ‘I had lot of people, including big experts in the field, writing to me and offering comments and constructive criticism,’ says Patwari. ‘We had some very good discussions.’ He is currently revising the paper to incorporate this peer input, before submitting it to a journal.

Saggiomo was thrilled when a researcher went one step further than just mentally considering another of his preprints, and tested the procedure it detailed out for themselves. Posted on ChemRxiv in June 2021, this preprint described repurposing a cheap 3D printer into three (normally very expensive) syringe pumps. Less than three weeks later, a researcher at the University of Oklahoma in the US sent a photo of pumps made in her lab using the protocol. ‘I didn’t give them any input, so this proved the protocol is repeatable just by reading and following the instructions,’ Saggiomo explains. Since then four more groups have also told him that they’ve successfully copied the design. ‘One group even adapted it for a new printer and we’ve added this new design to our paper,’ he adds.

Neither of the chemistry preprint servers, however, allows readers to post comments on their sites. ‘We had [comments on SSRN] early on, probably 15 years ago, and it was a bunch of noise and a waste of everybody’s time to weed through a hundred comments to get to one or two that were thoughtful,’ explains Gordon. Instead, conversations happen via email or social media. ‘The Twitter space is where most of the interesting conversation around preprinted research is happening,’ he adds.

Some preprint servers for other subjects do, however, host comments on their sites. BioRxiv and MedRxiv also allow peer-review comments and correspondence between the journal editors and the authors to be posted alongside preprints.

Claiming novelty

Science is a competitive field, and the ability to stamp a name more quickly onto a novel idea is another recurring reason given for using preprints. The journal review process can take many months, and preprints offer a turnaround of typically 48 hours from submission to posting, explains Blackmond. ‘It’s a way to stake a claim,’ says Mudrak. ‘If you’ve got some new results that you’re very excited about, a preprint helps you get your name out there first.’

Preprint servers can also help you share the limelight if you haven’t been quite fast enough, explains Nelson. In October 2018, a group based in Sweden published a paper in the journal Angewandte Chemie containing similar research to a manuscript he had just submitted to a different journal. The papers both described the use of electron crystallography as a rapid alternative to x-ray crystallography for determining the 3D structure of small molecules. ‘The journal rejected our paper because of the other paper and we decided to put it on ChemRxiv so we could also get credit for it,’ says Nelson. His preprint was posted the day after the Angewandte Chemie paper was published, he tweeted about it and the groups shared the praise from the community.

The Nelson group also immediately re-submitted their paper to ACS Central Science where it was reviewed and published in less than three weeks. By then the preprint had already received ‘tens of thousands of readers’, he explains. It is currently the third most accessed preprint of all time on ChemRxiv, and by the end of July 2021 it had been accessed nearly 79,000 times.

Prior to being scooped, Nelson says he was indifferent to the idea of preprinting his work but now – provided his co-authors agree – he puts all his group’s papers on preprint servers. ‘With preprints there’ll be a record of you putting your idea out there before the publication process,’ he explains.

The discussion on whether a preprint can really be used to establish precedence is still ongoing, however. ‘My feeling is that once a paper that was on ChemRxiv is peer reviewed and in the literature, then the ChemRxiv timestamp should hold,’ says Blackmond. But that is not a view held by everyone, she adds.

Open conversations

Preprint servers can also be used to have open conversations. ‘We’ve seen some really interesting things in finance [on SSRN], especially during some of the financial crises over the years, where we’ve actually had really thoughtful debate with people, in some cases Nobel prize winners, discussing ideas, back and forth. One posts a preprint, another posts a different opinion,’ says Gordon. He draws similarities between these and the letters from the early years of the Royal Society used by scientists such as Isaac Newton and Robert Boyle to debate theories.

Using preprints to debate research ideas is something chemists should be doing more of, says Lee Cronin, a chemistry professor at the University of Glasgow, UK. ‘If you look at physics arXiv or bioRxiv, there’s lots of new ideas that people are working on that might take a few years to work through the system [and get published]. Preprints that I put on ChemRxiv are meant to open the conversation.’

Blackmond is currently using ChemRxiv to openly discuss a complicated situation with Mathias Christmann at the Free University of Berlin in Germany. Both groups are studying the enamine catalysis mechanism, and in March 2021 Christmann posted a preprint on ChemRxiv that challenged a large body of published work from Blackmond’s lab. Two months later, Blackmond posted a response on the same preprint server. ‘There have been Twitter comments marvelling at how quickly such a scientific discussion got underway compared to what it would take in the peer-reviewed literature,’ says Blackmond.

Her group’s response contains new data – a requirement for ChemRxiv. ChemRN, however, welcomes anything that contributes to the scientific discussion, says Gordon. It allows manuscripts that detail research still in its early stages, for example, and also literature reviews.

Career bonus

Early-career chemists likely have the most to gain from using reprint servers with many scientific funding bodies now allowing preprints to be included in grant applications. This includes UKRI (UK Research and Innovation), the European Research Council and the US National Science Foundation.

We actively look at ChemRxiv for commissioning leads

May Copsey, executive editor of Chemical Science

They have also proved useful in job applications. ‘I really like it when applications from early career researchers or lecturers have a preprint or two on their CV that you can read instead of it just saying “I have three papers being submitted to Nature”’, explains Coudert. ‘If there is a link to the preprint, I can go look it up and [assess the quality of the work for myself]. It helps the candidates when they do this.’

It easier also for students to know what of their work they are allowed to discuss at interview if it’s been preprinted, says Nelson. ‘They can go to job interviews and know its fine to talk about those results because it’s already on the record.’

It can also help researchers determine what journals are best suited to their work: some chemistry journals have recently started looking out for tweets about preprints and scouting around the preprint servers looking for articles that would suit their journals. ‘With more than half our preprints we will get associate editors reaching out to us asking if we want to submit to their journal,’ says Nelson. ‘We actively look at ChemRxiv for commissioning leads,’ says May Copsey, executive editor of the RSC journal Chemical Science.

The future

After a bumpy start for the chemistry preprints, chemists have increasingly embraced the preprints servers over the past four years. The pandemic highlighted their utility as a tool for the rapid and open dissemination of research results for use by others, but chemists are finding a myriad of other uses for them too. These include obtaining rapid feedback from the community to improve papers before submission to journals, stamping a name on novel finding and having open debates with other scientists.

Journals themselves should encourage authors to put papers in an archive

Naresh Patwari, Indian Institute of Technology Bombay

Still, the use of preprints in the chemistry community remains far from universal.  Coudert has previously estimated that roughly 500,000 papers in chemistry and chemistry-related fields are published in journals each year. By comparison, in the four years following their launch ChemRxiv posted 12,092 preprints in total and ChemRN just over 16,000.

Saggiomo believes their adoption will continue slowly but steadily. ‘If we teach the new generation that it’s more useful to have comments from a lot of peers rather than two anonymous referees I think [scholarly publishing] will change in one or two generations,’ he says. Patwari, however, has a suggestion for speeding things up: ‘Journals themselves should encourage authors to put papers in an archive for, let’s say, 15 days first before they’ll send it out to peer review.’

Coudert also predicts their use will grow as researchers increasingly shun conference travel for climate change reasons. ‘At conferences your talk about your latest finished or almost finished work, and you get feedback from people at the coffee break,’ he explains. ‘Preprints serve a role that is parallel to that of the conference communication.’

Nina Notman is a science writer based in Salisbury, UK

A graphic in this article was updated on 11 October 2021 to clarify the number of ChemRxiv preprints that have been published in Angew. Chem.