How to start collaborations.

Someone asked about this in a comment, implying it was something that should be taught in grad school.

Here's how I think I learned:

1. Panhandling from a poor lab.

This was out of necessity. I used to be the kind of kid who was shy about answering the phone at home, so you can imagine I had to get over a lot of shyness to go door-to-door in my department asking for reagents and equipment (and sometimes for someone to show me how to use the equipment).

I met a lot of people this way and found out what was going on in all the labs.

Step 1: Know who's doing what.

2. Trial and error.

I did a couple of collaborations in grad school, set up by my advisor. One went pretty well; the other ended prematurely at my request. I've blogged about that one before.

Step 2: Know who knows what they're doing, and who is going to treat you like dirt.

3. The kindness of strangers

I set up collaborations as a postdoc primarily by asking for things, just like in grad school, only this time I did it by email. My best collaboration of all time was done almost exclusively via email and Fedex.

Step 3: Pick people who respond promptly, who communicate clearly, and who genuinely want their stuff to work for you.

4. Talk about publications up front.

If it's going to be your paper, say what you're doing, what your goals are, what your vision is, and what you need them to do for you. Make it clear what it's in it for them, and don't be afraid to make the boundaries clear, too. If you're going to be first author and you absolutely will not accept a co-first* author* scenario, make that clear, but also be aware of what kinds of expectation that sets up.

Most people will appreciate constant communication on this subect, with some wiggle room. Experiments are added during revision, which often means adding authors or changing the order. Most people will be okay with this provided everyone is kept in the loop and has a chance to voice their opinion (and maybe even influence the final decision).

Some people will turn down authorship and then avoid giving you information you need to finalize publication. Other people will be offended at the suggestion of discussing authorship before the project even begins. With those people you can talk about doing pilot experiments first, but nail them down on meeting to discuss nuts and bolts at a set time after that.

If you're trying to get on other people's papers, offer to do an experiment but make it clear that you're sticking to the standard rule: if you contribute a figure or a part of a figure, you deserve to be an author, too. This can help your CV, particularly when you're a junior grad student or postdoc (for example), but don't let yourself get too distracted with doing these kinds of favors unless you get something in return. For example, I did a couple of these kinds of figures as a way to get unpublished reagents that were being used in the course of the work. Everybody knew that was the arrangement, and I think it was win-win for everyone.

If your potential co-authors keep ducking the authorship issue in the beginning, they're likely to put up a fight about it later on. Avoid these types. They're usually insecure, bad communicators who have an agenda of their own, and they know that open communication makes it hard to sneak around. You can try to pin them down, but keep in mind they can always change their story later, when it's too late for you to argue (i.e. they can take your name off the paper without your permission) or find someone else (i.e. when they did critical experiments for you, and you can't publish the work without their permission).

Yes, as a postdoc there are some special pitfalls. In my opinion, it's not worth it to try to lodge a formal complaint if all you did was 1 figure and got left off the author list. Do what we all do, and most of us have had this happen once. But once is enough. Choose carefully who you want to help out.

LIVE AND LEARN.

5. Publish the damn work already.

I have several papers pending right now, and they're all collaborations. In some cases, it's my fault and/or my PI's fault. In the other cases, it's my collaborator's fault. There's only so much I can do about papers where I'm a middle author, but I resent knowing I did work for these people and it's just sitting around. I'm sure my co-authors feel the same way about my papers.

The key thing here is to keep people informed when it's your responsibility to get the thing published.

And if you're the middle author, don't be afraid to send an email every 6 months or so asking how it's going and whatever happened with that manuscript. It probably won't make much difference, but in one case I found out after the fact that the paper had been revised drastically and accepted at a different journal, and my collaborators never sent me the manuscript. So I would not have known to add it to my CV unless I had, for some narcissistic or competitive reason, Pubmedded myself. Totally inappropriate, but sometimes it's good to check in and make sure they're not (inadvertently or otherwise) screwing you over.

And that, my friends, or more or less all there is to it. Questions on this subject are welcome.