I’m finishing up the primary research for my PhD, and as I start to go through my list of “fun projects” to try, I am terribly aware of the big picture. Part of growing up (as a researcher) is not just doing analyses for fun, but moving toward doing work that has clinical applicability. It also seems that the traditional role of the academic – as a PI at a large research university – is drastically changing. A PI is less of a research scientist, and more of a manager with skills in acquiring grants to fund the work of the lab, and acting as a spokesperson for those that provide the funding. A PI who can foster relationships with private companies that (actually) have the infrastructure to create meaningful medical applications or databsaes for mining (eg, Apple, Google) is a successful one. A PI who still has his or her nose in some small local data is likely not going to survive. The training that I am getting in Biomedical Informatics used to be to make an academic – a professor, a leader of a lab, and someone who is invested in the research. The training that I am getting is more to produce a expert in all things data analysis to either:
- Start / lead a company that will do big, great things
- Be employed at a company to do big, great things
- Become a PI to foster a relationship with said private entities
I think all of these options are great. But I still have about two years left. And as I finish up this current work mapping genes to brain to disorder, I’m haunted by the question of 1) my ability to truly demonstrate clinical applicability in the short scope of graduate school, and 2) what I should work on next that could be useful, and 3) if any work that can be done in the scope of a graduate school career can ever really be as meaningful as I’d want it to be (in cases where it does happen, maybe it is just luck?)
What am I supposed to be doing?
It’s my “job” to come up with a well scoped project, and follow through and complete it for a thesis. I guess that doesn’t seem so hard, I’m almost done. But with two years to go (the golden rule is that you can pretty much go as long as you have funding) I am looking critically at my own work, and craving challenge. I am pretty good at bringing together completely loony ideas, testing them out, and sometimes stumbling on a use of data or methods that is (mostly) useless, but minimally, cool, but that doesn’t feel like enough. I have (recently) been trying to think of “bigger” problems in medicine, specifically with imaging, since I am coined as an “imaging person.” The problem is that there is so much issue with the current infrastructure with regard to acquisition and sharing of the data that there is nothing to build on top of. Assuming easy access to data, the other problem is that imaging is just not as great as we want it to be. It’s another marker of phenotype (and thus important) but we aren’t going to find single biomarkers for some disorder in, for example, a brain scan. What we can find are interesting relationships between genes, behavior, and brain, but I’m not convinced that finding such a relationship, and publishing on it, is really so useful. Do we really need a highly expensive brainscan to tell us that someone is depressed? Just ask them. When we find genes that can explain some tiny percentage of risk for a disorder – is that really so useful?
Anyway, aside from my skepticism about this, as a graduate student I am pressed to ask myself “What am I supposed to be doing?” Aside from my well scoped thesis work, I (think) that I should be hammering my head against some unsolved problem in the field. So then I can ask myself – what is that problem? Again, I have ideas, but arguably this kind of “expert” insight should come from a PI. Here is where it gets challenging – neither of my PIs (I don’t think) can offer the specific expert insight that I would need. I’m interested in the integration of large scale imaging data with genes, behavior, and maybe even drugs. I can find PIs that are experts in either or, but not both. So I feel like the burden of figuring out “the big problems” is in my hands. But I’m not sure I have the experience to have that insight (yet).
So – what to do? What am I supposed to be doing? To step back, maybe as a graduate student I don’t (yet) need to be solving these kind of huge problems. Arguably, this phase of my life is about learning, and growing to fill the hole of the expert that I haven’t found myself. In that light, the little projects that I do over weekends, the playing with APIs and visualizations and data, is probably a really great thing, and I’m just contemplative and thoughtful, wondering about my own potential, and why I cannot do more.
A Plan of Action
The best thing to do is probably to finish up my thesis work, and continue reading everything, looking for data in unknown places, and thinking about gaps in methods and technology that would be fun to work on. I was terrible at publishing before graduate school, but I’m getting better. When I was 21 there wasn’t a computational bone in my body, and now at 28 I have quite a few. Before graduate school (but when I knew that I wanted to pursue a PhD) I imagined that my job was to join a lab, and do the bidding / vision of the PI, and then graduate and hope that someone would need the skill set associated with that bidding. I’ve now realized that my job is to be an independent, inspiring, and passionate data scientist. I am responsible for my own learning, for my own collaborations, and I have to be able to develop a vision and (sometimes single handedly) bring it to life. In fact, it is frustrating to have to rely on someone higher than you to move forward, and that should be minimized. It would be nice to have structured and concerned guidance and leadership, but I’ve realized over the years that no one will ever be concerned enough with my development or success to fit that role. It’s actually not fair to expect that of a person, given the change of the PI from data scientist to manager and link to private industry. I want to be the best, the greatest, at all things data analysis and visualization. I’m convinced that I can be. So I’ll keep working on that skill set for now, even if the “big picture” project or research aim is not yet clear.