Although academic institutions dominate the natural-science focused Nature Index, there is notable contribution from industry, which in 2014 represented nearly a quarter of institutions in the index, contributing to nearly 7% of articles. Analysis of collaborations in the index from 2012 to 2014 reveals several global hotspots in East Asia, Northwest Europe and the United States where there are multiple industry–academia collaborations. And although there are several factors behind this geographic clustering, ease of communication is eroding many of them.

“The barriers to global collaborations are very low right now,” says Edmund Graziani, head of bioconjugation chemistry and natural products discovery at Pfizer in Groton, Connecticut, who frequently collaborates with academics. New York-based Pfizer has more than half of its collaborations with institutions outside the United States (see 'Pfizer's partners') but Graziani says this is nothing out of the ordinary for a multi-national company. Pfizer is experienced at running long-distance projects, plus the company has several R&D centres across international locations that work together. “We don't typically see that as a barrier to collaboration,” he says.

It seems that location has little effect on the success of a collaboration.

Indeed, it seems that location has little effect on the success of a collaboration, according to a 2010 study1 by researchers from MIT, which interviewed 25 research-intensive multi-national companies. Study author, Julio Pertuzé, now at the Pontificia Universidad Católica in Chile, says that although geography has historically been considered integral to industry–university links, this is less true for collaborations between large firms and prestigious universities — those most often found in the Nature Index. There is a trend for clustering of businesses around top universities, though no established pattern for how these clusters start. “It is a little bit like the chicken and the egg in terms of how industry-university relationships evolve,” Pertuzé says. Such technology clusters can include university spin-off companies, larger corporate research centres and collaborative industry-university research initiatives, all attracted by the expertise the university can offer, and the networking opportunities within the cluster itself.

In Japan, Osaka is the second largest industry–academia collaboration cluster after Tokyo. The city is home to several renowned universities including Osaka University as well as many electronics and pharmaceutical firms, such as Takeda Pharmaceutical Company, Sumitomo Electric Industries and Shionogi & Co. Ichio Shimada, a physical chemist in the University of Tokyo, has worked with Shionogi for five years. Shimada says the company contacted him because of its interest in his published work and conference presentations on cell membrane protein receptors and their molecular interactions.

IBM researcher Stuart Parkin (right) with Stanford graduate student Roger Wang inspecting part of the thin-film deposition system that underpins spintronics technology at IBM's Almaden Research Center. Credit: IBM Research

Shionogi proposed a collaboration that closely paralleled his work under which experiments were carried out in Shimada's lab in Tokyo with company researchers who also provided reagents. The collaboration resulted in one paper2 in the Nature Index in 2012, on monitoring the interactions between drug molecules and protein receptors. Shimada says he met Shionogi management to discuss project progress every four or five months, making the 500-km journey between Osaka and Tokyo in only three and a half hours on a bullet train. “They collaborate where there is what they need,” he says, of Shionogi. “Location is not a big issue. Japan is not as large as the US.”

In France, university–industry collaboration often happens through 'mixed' laboratories. One such physics lab, Unite Mixte de Physique CNRS/Thales (UMPhy), involves electronic systems company, Thales Group (previously Thomson-CSF), which is France's leading company for academic collaborations in the Index (see 'Parisian firms'). The Paris-based lab was created in 1995 by France's National Scientific Research Centre (CNRS), Thales and the Université Paris-Sud. UMPhy grew out of a longstanding magnetic materials collaboration between Nobel Prize winner Albert Fert of Paris-Sud and Thales Central Research Lab. It is considered the birthplace of spintronics — an emerging branch of electronics that manipulates the spin of an electron in addition to its charge. Frédéric Petroff, deputy director of UMPhy, says that the lab “strongly favours interactions with Thales researchers and engineers”, having been set up to stimulate what he calls “the coffee machine effect” — where new ideas are conceived by colleagues chatting while taking a short break.

Javier Villegas, a physicist in UMPhy, has two collaborative publications in the index, on superconducting and magnetic materials3,4. Villegas says interaction with the 22% of industrial researchers in the lab is helpful for the academics. “It allows me to identify and imagine concrete ways in which my more basic, fundamental research could address current technological challenges,” he says. “Conversely, Thales researchers can find novel technological applications based on my research findings. This symbiosis is extremely profitable for everyone.” There are other benefits of industry–academia tie-ups, says Marko Erman, chief technology officer at Thales. ''The tax reduction mechanism in France (the so- called Crédit d'Impôt Recherche) makes such cooperation financially attractive for both partners.” The benefits of working in such close proximity are embraced by Thales, which has four other joint laboratories in the United Kingdom, the Netherlands, Singapore and Canada.

Push and pull

The Nature Index also contains many collaborations between companies that have based a research centre near a major university. New York-based IBM is the company with the second highest collaboration score in the index, after China's BGI. In the heart of Silicon Valley, IBM has its San Jose-based Almaden Research Center, which works closely with nearby Stanford University — its most frequent collaborator, with 25 jointly authored papers in the Index (See 'US hubs'). Former Stanford electrical engineering student and current IBM Almaden researcher, Timothy Phung, has experienced both sides of such collaborations. His 2013 index paper5 was the basis of his PhD work, and part of a continuing ten-year collaboration involving his Stanford supervisor, electrical engineer, James Harris, Stanford physicist, Shoucheng Zhang and the IBM team of Stuart Parkin. The group, known as the IBM–Stanford Spintronic Science and Applications Center, is developing magnetic materials for a new type of high density solid-state computer memory known as domain-wall or 'racetrack' memory. “Stanford is a very entrepreneurial place, and collaborations of this type with companies are very much welcomed,” says Phung.

IBM Almaden lab director, Jeff Welser, says setting up near university and innovation centres “has certainly contributed to our ability to collaborate effectively”. He says both Stanford and IBM staff attend the same symposia — leading perhaps to a 'coffee break', rather than coffee machine effect. “It is also very easy for our researchers to spend a day or more a week at the university[ ... ]facilitating improved knowledge transfer, more opportunities for co-invention, and joint facility usage,” says Welser.

figure 1

Digital Science To explore the graphic in more detail online, visit: go.nature.com/i2LVfb

All institutions in the index are shown here, sized by fractional count (FC) and coloured by geographical region. A pair of institutions is connected when they collaborate on a paper — the more papers, the stronger the connection and the thicker the connecting line. Clusters of institutions are determined algorithmically and depend on the size of each institution and the strength of its various connections (analogous to many charged balls that repel each other, connected by springs of different strengths). The diagram shown emphasizes certain country clusters, with industrial institutions picked out in black. The distribution of corporations within the clusters reveals information about their relationships with other non-commercial (mostly academic) institutions.

Sometimes, a collaborative push might come from academia. Harvard Medical School microbiologist Gerald Pier started his collaboration with French pharmaceutical firm Sanofi in 2009, after consultants helped him search for a commercial partner. As well as his academic post, Pier has a start-up company, Alopexx Pharmaceuticals, as a side project, and it was through Alopexx that he collaborated with Sanofi on an antibody treatment for serious bacterial infections as an alternative to antibiotics.

boxed-text

The partnership led to a 2013 index publication6, and in 2014 the collaboration successfully ended with Sanofi licensing the technology. Pier says the long distance between Boston, Massachusetts, and Toulouse, France, posed some problems such as time-zone differences, but essentially the issues were the same as with any collaboration, such as working out how to establish protocols and share data.

The greater Boston area, home to several leading universities and teaching hospitals as well as global medical technology firms, has the third largest industry–academia collaboration score in the United States. Pier says the local environment encourages academic interest in commercializing technologies. But he also suggests another reason why such collaborations seem to cluster in regions like Boston: he has seen a rise in the number of “two-career couples”, who will be looking to work at different organizations in the same location. Pier admits, however, that this effect is difficult to quantify. So does location matter? Julio Pertuzé says his work showed that it is communication, not distance, that matters. “Proximity is only important if firms act upon the opportunity for increased face-to-face interactions,” he says. But he adds that, in many collaborations, “there are other options for increasing these interactions that have little to do with geographic distance.” Pertuzé says these include relocation of industry personnel to work in a partner's lab, or facilitating the exchange and travel of researchers between sites. And while location plays a part, ultimately “it is a question of the specific technology or innovative approach that a given academic brings to bear,” Graziani explains, “the quality of science far outweighs location.” R.B.