So, you are starting a new graduate program, or postdoc or even your very own shiny new lab. The world of science is at your fingertips. Pools of knowledge await to be discovered. Clouds in the horizon look puffy and white, reminding you of all the wonderful papers that you will publish. The lab benches gleam with rainbow-colored rays. Unicorns dance around your Pipetmen. Music from the Mojo Banjo fills the air. You're ready to kick some science butt!!
But…what the fuck are you going to work on?! Sure, you picked a lab or wrote a grant that more or less indicates the area you are going to study. Your PI might suggest some stuff, or you have some stuff left over from your postdoc work. But this is YOUR project, how do you find ideas for new projects?
When I started graduate school, I joined a lab in which my advisor did work that was very interesting to me, and stayed since we got along well. When it was time to pick a project, he said "Well, I was just reading this somewhat crappy paper from an even crappier journal which, if true, would be a very interesting finding. Why don't you replicate this properly and use this as the basis for your PhD." Naively I said, "Sure!". Of course this turned to be a rather obscure brain nucleus with very few papers showing how to do electrophysiological recordings from it, and most of the labs that had worked on it had since moved on to something else. When I asked my PI for help he said, "Well your guess is as good as mine, I've never worked with this part of the brain." So I had to ask an older, bearded neuroanatomist upstairs to show me on his dusty old brain slides how to locate this nucleus in rodents and how to best prepare brain tissue for this. And it took me the better part of a year to figure out the experimental preparation and to do experiments reliably. But then after that, I managed to properly replicate the findings of the original paper, and expand them onto several high-profile papers which stemmed from the original experiments. Not only that, but I was one of a handful of people who could record record from this nucleus, so we didn't have to worry about much competition. At least at first.
This approach of reading even the lowliest of papers in your field, I think is a good way to get ideas. Not because you are stealing someone else's (they are published after all), but really to use them as a basis for coming up with better and creative ways to re-do experiments that were interesting, yet perhaps not conclusive, or sloppy. Also, sometimes in a paper there will be one obscure figure, that is somewhat beside the point of the main paper but shows an interesting finding. These sort of side findings are also often sources of good ideas. That is why its important to really read widely in your field and beyond. Not just focus on papers in fancy journals, because you may be missing out on a treasure trove of good ideas, even from crappyish papers.
How about you, where do you get your ideas for new projects?