There are lots of reasons to team up with other researchers. Doing so can help you access scientific expertise and equipment; get more publications; write stronger grant proposals; pursue new interests; and, hopefully, have some fun. But it also comes with potential risks and challenges. What initially looked like a good idea may lead down a blind alley, and often collaborators need to start doing their part before they know whether the whole project is viable. Collaborations are also interpersonal endeavors, requiring trust, communication, and open-mindedness.
There is no single recipe for making collaborations work. But these examples of collaborative efforts illustrate some effective approaches, and some pitfalls to avoid, for junior researchers looking to harness collaboration to develop their careers.
Collaborate for support and bold advances
Neuroscientist Robbert Havekes initiated his longest-standing collaboration 14 years ago, when he was a Ph.D. student studying memory, and it is still ongoing. It started at his department’s coffee machine, where he often bumped into Peter Meerlo, a senior postdoc in his lab who was investigating sleep. As they sipped coffee, the two started talking more and more about how they could bring their interests together, and the resulting investigations are now an important facet of both scientists’ research portfolios.
Collaborating with someone who has complementary expertise is “the perfect way to join forces,” says Havekes, who is now an assistant professor at the Groningen Institute for Evolutionary Life Sciences in the Netherlands, because it ensures that you have access to help as you venture into a new area. Interacting with people with different perspectives or approaches also “prevents you from getting a tunnel vision” about your research. Working with friends is a bonus, he notes. “It doesn’t mean that we always agree … but we always manage to figure [it] out.”
Early in his career, Havekes also learned the hard way how important collaborations can be. After spending 6 months as a postdoc trying to establish transgenic mouse lines, it became clear that the experimental model wouldn’t work and his entire project was killed. Luckily, his principal investigator (PI) was a fervent collaborator, and Havekes was able to contribute to many of these ongoing projects, which provided “a nice safety net,” Havekes says. During this time, “everything I did for collaborators seemed to fall into place.” Postdocs “have to get some first-author papers out, but keep in mind that through collaborations you may be able to land some co-authorships and recommendation letters that will really help your CV as well.”
As he started his own lab in 2015, Havekes followed his former PI’s advice to apply for a Young Investigator Grant from the Human Frontier Science Program (HFSP). The 3-year grants are designed to support teams of junior group leaders from different countries and disciplines as they address “questions that could not be answered by individual laboratories,” the website states.
When applying to a program like this, “the first thing is to get a group together that’s bold and strong,” Havekes says. He went to neuroscientist Sara Aton of the University of Michigan in Ann Arbor, whom he had met when the two of them were postdocs. Back then, “we talked a little bit every now and then about ‘it would be fun to do some projects together,’ and she got involved with two of my papers,” Havekes says. Now that they were both group leaders, they were excited to join forces for a grant application.
As they decided to study the role of circadian clock genes in neural plasticity and memory, they determined that they needed to find a synthetic biologist and a theoretical biologist to complete the team. Their network led them to Jae Kyoung Kim from the Korea Advanced Institute of Science and Technology in Daejeon, South Korea, and an internet search identified Matias Zurbriggen of the University of Düsseldorf in Germany. Both agreed to join the team. Collaborating with people you don’t know can be “a bit of a risk,” Havekes acknowledges. But this is “a nice group … of young people who really want to make something happen and are excited about working together.”
It took the team members a couple of months of regular Skype meetings and working on shared documents to put their proposal together. Incorporating scientific aspects all the way from modeling to behavior was challenging, and it would probably have been too bold for most funding programs, Havekes says. But it “was the right approach” for HFSP. The team was awarded $1.35 million last March, and the researchers will all meet in person for the first time at an awardee conference in July.
In addition to the excitement of working on a bold project with new people, the HFSP grant is helping Havekes make strides toward tenure by demonstrating his ability to raise funding. “With collaborative work, it might sometimes be easier to get grants, as you can work with experts from other niches,” says Havekes, who also recently secured a smaller grant on his own. One important thing for junior PIs to bear in mind, however, is the need “to make sure that the collaborative projects are reasonably related to the main research line of the lab.” With this mindset, he can ensure that his collaborations contribute to “further strengthening the scientific identity of my group.”
The whole is greater than the parts
At a conference in 2013, cell biologist William Earnshaw heard an exciting talk—and was convinced that his lab could help the researchers “push the work even further,” he recalls. Systems biologist Job Dekker and computational biologist Leonid Mirny were presenting their two labs’ work using DNA interaction maps and computational analysis to predict that mitotic chromosomes contain loops similar to those that Earnshaw had long observed under the microscope. By using new cell lines and DNA samples that Earnshaw’s lab had the expertise to prepare, the three research groups would be set to produce even more detailed models. “The best collaboration is when each lab can do something that the other labs can’t do, but you all have a common goal,” says Earnshaw, who is a PI at the Wellcome Centre for Cell Biology in the United Kingdom.
Because he was joining an existing collaboration, “it was very important that they were willing to welcome us as equal members,” Earnshaw says. Discussing what everyone wants to get out of a collaboration and agreeing on a set of rules at the beginning helps prevent potential conflicts, Earnshaw says, even though it may feel a little awkward. For example, the members of the new collaborative team agreed that any resulting papers would have three lead corresponding authors—postdocs Kumiko Samejima and Johan Gibcus and doctoral student Anton Goloborodko, who did the bulk of the work in each lab—and that each PI would be listed as a senior author. Then you try to stick with your original agreement while also remaining flexible as the project may evolve in unexpected ways, Earnshaw adds.
“Collaborations very often start slowly,” Earnshaw continues. Researchers typically have ongoing projects and interests vying for their time and attention, and because there is often more uncertainty about whether collaborative projects will work, it can be hard to make the work a priority, he explains. In the case of his work with the Dekker and Mirny labs—the first installment of which was published earlier this month—it took 2 years of “jiggling things until they fit together” and the project really got off the ground. After that, Earnshaw and his collaborators established a timeline for the work to be done, and the core team of six authors met via Skype whenever they needed to discuss new data or make key decisions.
Earnshaw hopes that his collaboration will keep going for as long as possible, but “collaborations will usually have a life time,” he says. “If we stop producing samples that are as interesting as Job and Leonid can get from other people, then they may decide it’s better for them to move on and collaborate with others. And I will simply have to understand that.”
Earnshaw, who has been involved in collaborations since the beginning of his career, also warns more junior group leaders to be particularly careful about power imbalances. “If you are collaborating with famous people, you have to make sure that the author list makes it clear what your contribution is so that you’re not regarded as a secondary person in the collaboration by the people who will evaluate your position, your grant, or your fellowship,” he says. There also is a risk of being scooped. “If you are a small lab who has few projects and you tell your secrets to a big lab and the big lab cannot fully control their appetite, they could really damage you” by becoming your competitors, Earnshaw says. You are safer if you have other projects or ongoing collaborations, he adds.
But the key is to choose your collaborators well to start with. It’s a good idea to get to know them in advance and get a feel for their reputation by asking people in their community, Earnshaw recommends. “A lot of science is based on trust,” he says, especially when you do not have the expertise or equipment to check other labs’ work. To guard against potential research integrity issues, do your due diligence and find out whether prospective collaborators are “really known as being very trustworthy, and then all you can do is keep your eyes open. And if you see anything that doesn’t seem right, then you ask.”
The most beneficial collaborations are when “everyone is able to answer questions that none could address fully on their own,” Earnshaw says. “The icing on the cake is if you are able to do this with a group of people who you both like and respect, so that for the life of the collaboration everyone feels that they are part of a winning team.”
Riding a wave of opportunity
When Hurricane Irene swept across the Caribbean and the eastern coast of North America in August 2011, it was a destructive force. But for Kathleen Weathers, a senior scientist at the Cary Institute of Ecosystem Studies in Millbrook, New York, the tropical cyclone helped create one of her most productive collaborations.
Just a couple of months after the storm, she was scheduled to chair a meeting of the Global Lake Ecological Observatory Network (GLEON), which provides a collaborative forum for scientists to share and analyze sensor data to examine lake function in the changing global environment. Hurricane Irene provided a wealth of data for researchers to dive into, and the meeting offered a perfect avenue to join forces. Altogether, 11 scientists from eight institutions along the coast and representing disciplines as diverse as engineering, biology, limnology, and ecosystems ecology engaged in a short-lived but intense collaborative push to glean everything they could from the serendipitous data, Weathers says.
From day one, they all agreed on a clear, common objective: studying the disturbance and recovery of lake ecosystems as a hurricane passed through the region. This helped guarantee not only that everyone was pushing in the same direction, but that every collaborator was in an equal position to contribute their perspective on how the research question should be tackled.
The collaboration that emerged “was truly interdisciplinary,” Weathers says. Bringing all disciplines to bear on a research question is rewarding, but it also is “a hard thing to do,” she adds. For example, at first there was disagreement about the best approach for analyzing some of the data because different disciplines had different ways of doing it, she says. Ultimately, the group decided that a hybrid approach would be the best option. Figuring out how to blend disciplines in new ways takes trust, respect, and “a lot of time at a whiteboard, crunching data, and doing reality checks,” Weathers says.
Another success of the collaboration is that it was “very egalitarian,” she continues. Typically, collaborations are led by a senior scientist who decides the roles of the other researchers, she notes. In the case of this GLEON collaboration, on the other hand, “at our first meeting we went around the room and discussed where everyone was in terms of career stage and how much time she or he might have to commit to leading this study to fruition,” Weathers says. “One of us, Jennifer Klug, was coming up for promotion to full professor and had the time, interest, and skills,” as well as the career incentive, to lead the project, with Weathers and two other team members working with her.
Everyone else’s contributions were also discussed at that time, and these conversations continued during a follow-up face-to-face meeting and regular Skype meetings. This approach provided “an effective way” to start talking about responsibilities and authorship, and a mechanism to update the plans if necessary, Weathers adds. In the resulting paper, published just 1 year after the initial meeting, Klug was first and corresponding author, Weathers was last, and the author information detailed everyone else’s role. “You can see that it really was a team effort,” Weathers says.
Ultimately, the purpose of collaborations is to enhance “the quality and the creativity of the science product,” Weathers says. “The gift of my mid-to-later career has been to be able to work in dynamic, truly collaborative teams with creative, talented, and enthusiastic people from all career stages.”