Collaborations are becoming increasingly important in biology because of the need to apply multiple technologies to tackle the most complex current problems. The U.S. National Institutes of Health recognizes this need, and has created the “multi-investigator” granting mechanism to facilitate this process. I have reviewed a number of proposals that utilize the multi-investigator mechanism and have generally found them to be superior to individual investigator grants. Setting up a good collaboration, however, can be extremely difficult.

First, it is probably necessary to say what a collaboration is not. I do not consider providing reagents or advice to your fellow scientist a collaboration. That is part of normal scientific citizenship. Every scientist should be willing to accommodate reasonable requests for assistance. No, a collaboration is when you work with another scientist with complementary expertise to solve a common problem.

Sharing the same scientific goal is essential for a collaboration, but it is also what makes collaborations so difficult to establish. Scientists with the same research goals tend to have the same expertise and are usually your competitors. Scientists with different expertise, unfortunately, are likely to be interested in totally different problems that require that expertise. The lack of overlapping interest in scientists with divergent expertise is a huge problem for creating productive collaborations.

I learned about the difficulty in starting a collaboration when I began searching for one early in my career. I was building mathematical models of receptor dynamics, but I was trained in biochemistry and cell biology, not mathematics. I was fortunate to have taken several advanced calculus courses in college, but this was only sufficient for some of my simplest mathematical models. As the models became more complicated, I started to struggle. Around this time, I met a mathematician while visiting Los Alamos, who was also interested in receptor dynamics. I thought that a collaboration with him would help us both, and he seemed to feel the same way.

Unfortunately, this collaboration did not get very far because we wanted to pursue quite different problems. He was interested in the mathematical representation of receptor interactions and I was interested in the consequence of those interactions. My mathematical needs were too simple for him and his experimental needs were too tedious for me. We both became frustrated with the situation.

Fortunately, he knew a chemical engineer from the University of Pennsylvania, Doug Lauffenburger, who was an outstanding mathematician and also shared my interest (coincidently) in the regulatory dynamics of receptors. He introduced us and we immediately hit it off. Doug and I did indeed share an interest in the same problems and we had complementary skill sets. The result was a collaboration that has lasted over 20 years and has resulted in over 30 joint papers.

I have had 3 other important collaborations during my career, and each has greatly expanded the breadth and depth of my research program. This is not to say that there were not problems associated with these collaborations. Receiving appropriate credit is always an issue and your partner can sometime spring unpleasant surprises on you, such as bringing in additional people who dilute the credit even more. There is also the well-documented bias against collaborations by academic tenure and promotion committees.

Although my collaborations have been very worthwhile, I still find them hard to establish and maintain. However, I know other scientists who excel at generating numerous and productive collaborations. These scientists also tend to be extremely creative and productive investigators who invariably ask important, long-term questions. It seems that the longer the timeline that you set to a problem, the more general it becomes, providing more opportunities to include the research interests of other investigators.

So how do you start a good collaboration? First, you must identify what you can offer, then you identify someone who could benefit. I have found most of my collaborators at small scientific meetings, where it is easiest to identify a scientist’s true interest and establish the trust that collaborations require. You must also be willing to share credit for your ideas. Like any relationship, collaborations take time and energy. Still, there is nothing that can accelerate your research faster or expand your intellectual horizons more. And it can lead to lifelong friendships.

Steven Wiley is a Pacific Northwest National Laboratory Fellow and director of PNNL’s Biomolecular Systems Initiative.

We were talking to an industry and we required that we talk to some 20-30 senior scientists of various backgrounds to create a network. The experience was an eye opener. We came across three groups of people. One dismissed the idea right away since the answer to the question is very clear in their minds and a collaboration would be a waste of time. It took less than half hour that we were dealing with individuals whose attention span is fully predetermined. We beat a hasty retreat. Then we came across the second group...who would collaborate in the sense that they would facilitate getting samples, etc and the problem per se was of very marginal interest. Since it was an industrial problem, the possibility of consultancy was always there. An aspect of the problem was of high social relevance without remuneration though very facsinating. People simply disappeared when there were no returns.
The last category consisted of close friends over decades who collaborated over many projects. These formed the solid backbone. They were willing to switch problems and take on new ones just because some one, whom they can trust, finds it very interesting.
The moral is quite simple and self-evident.
There are situations where you are denied collaborations. We thought we needed to microarrays for arabidopsis for problems like drought tolerance. Our request was turned down. We devised alternative physiological approaches that did not require us to do the microarrays. We solved the problem at the level of the whole plant by devising whole plant energetics. Had we been offered the required help, we would not have found the solution so fast.
Having actively collaborated and published with physicists, statisticians, mathematicians,botanists that include systematics, systems theorists, simulation and modeling and computational experts, chemical engineers? I find that there is no substitute for collaborations if the problem needs to be taken to dimensions that you have not concieved. That is the whole idea of collaborations...a true collaboration would perhaps have no more than 20-30 % of the finished work in the province of a single individual.That required trust above anything else.

Collaborations come in many different flavors. It is hard to tell in advance whether they will pan out. I agree with Steve that nowadays single scientists or groups rarely have the necessary expertise to tackle multidisciplinary problems that result in the most insightful and highest-impact science. I think that an assertive discussion of future credit allocation PRIOR to establishing a collaboration is key to minimizing the risk and maximizing the mutual benefit. A lot of bad karma will result if you start discussing authorship issues right before sending the paper out for review in a multi-investigator project.

Sure collaboration is a double edged sword. However, I think the point of the article is that as many of today's problems are revealing themselves to be complex and while a single investigator may have all the tools and skills necessary to tackle such problems, most do not. This is where a complementary set helps.

With respect to the first blogger's comment, sure there are uneven sacrifices, but if you trust each other and you spread that out, things even out...but this point was mentioned in the article.

With respect to the second blogger's comment, the collaboration wouldn't have been funded in the first place. I won't bother responding to the rest because those are as random as the first comment.

Besides, what's the alternative? Compete with 10, 100 other people doing the same thing you are? How about new postdocs? Are they going to just keep doing the same thing their advisor did?

And that's becoming increasingly difficult in an increasingly competitive world, in a global recession, no less. Collaboration often times requires some uneven compromises or exchanges among the involved parties or indivduals. Sometimes one has to be willing to give up more or risk more for the collaboration to work, without necessarily reaping more rewards - if the rewards are realized at all. Steve, I don't know what your personal definition of collaboration is really like, but in America where individualsm coupled with capitalism drive most endeavors, including science, collaboration is a form of socialism which is strongly abhored and does not form more naturally than in other nations. So, keep dreaming, doc!

Hi

I wonder if this article is not an argument against special funding for collaborations? How much more difficult would the separation from the first collaborator have been if you were jointly funded? Moreover, did you not achieve your ultimate productive collaboration without any special program? And won't productive collaborations (but not unproductive ones) benefit all parties when they apply for individual (or joint) funding through normal processes?

Take care
Jim