Saturday, May 21, 2005

First-year recap: research interests, lab choice

I started out the year interested in a few rather different areas: immunology, malaria and synthetic biology, and a vaguely-defined notion of working on problems involving interacting "networks" of biological entities. Unfortunately, I ended up having to give up on two of these: the professor I talked to about immunology just ... disappeared, under circumstances nobody wants to talk about. And I couldn't find anybody at MIT interested in collaborating with Dyann Wirth, the Harvard professor heading up the malaria effort.

First lesson: if you're really interested in working on disease-related research, go somewhere that has a medical school.


As part of the CSBi PhD program, I was also required to do several rotations. The rotations were o
riginally supposed to be three month-long rotations, but the program was restructured to give us four two month-long rotations, and then I snuck in one more short rotation. The people I rotated with were:

- Drew Endy, synthetic biology's most visible proponent. Got a bit of wet-work experience, and actually got something to work.
- Doug Lauffenburger, one of the first people to start building mathematical models of signaling pathways. Learned a bit about the challenges of systems biology.
- Tom Knight, the other pillar of synthetic biology work at MIT. Got the chance to spend some time noodling about an idea I had to engineer cell-to-cell communication, as well as getting a good view at what it looks like when a good computer scientist becomes a good biologist.
- David Gifford, whose group works on machine learning algorithms that help shed light on genetic regulatory networks. Learned what large amounts of genomic data look like, and various ways to look at and try to interpret the data.
- David Bartel, one of the top folks in the currently hot field of microRNAs. More experience with evaluating large amounts of genomic data.

Originally, I was upset when the rotation program was expanded because I just wanted to select a lab and get going. With the benefit of hindsight, doing more, longer rotations and waiting an extra semester to pick a lab was a Good Thing. I got the chance to see more labs, and work in a few different areas. That said, one area I didn't really explore at all was cell biology, which deals with phenomena like how cells grow, move, divide etc -- basically, the next level up from the genome.

Second lesson: All other things being equal, pick a school/program that has a rotation system over one that doesn't.

Over the course of the year, a couple of new interests came up:

- RNA: As I've mentioned before, RNA used to be thought of as mostly just an intermediate messenger between DNA and proteins, but it's becoming clear that it's a lot more than that. What I think is neat about RNA is that it can act in both "digital" and "analog" ways: in its digital form, as messenger RNA, it can be viewed as a message that specifies how to make a particular protein ie it contains "pure" information that can be read by a "machine" and used to make a protein. However, it can also act in an analog fashion, by folding up into three-dimensional shapes that then serve to catalyze reactions. In this case, what's important about it is not the information it contains, but the actual shape it adopts ie it acts like a machine itself. It's kind of like the difference between the instructions on how to make a shovel versus the shovel itself -- RNA can act as the instructions, or as the shovel. [Sometimes even at the same time, like in self-splicing RNA.]

- Development: I find the notion that a single cell can give rise to an animal with 50 million million cells [in humans], of hundreds of different types, arranged in a precisely-defined three-dimensional pattern [to give you 2 legs, 2 arms etc] mindboggling. Talk about the ultimate algorithm.

I also went from the fuzzy "networks are cool" notion to being interested in genetic regulatory networks specifically. I think of genetic networks, determining when genes are turned on and off, in what order etc as being the source code for life -- when you run it, you get living things. And that, of course, appeals to my computer science side. You could certainly make the argument that the networks of interacting proteins, the end result of these genetic networks, are as important, if not more so, but I'm just more drawn to the "lower-level" genetic networks. I guess I'll always remain a systems-level [in computer science terms] geek at heart :-)

So, in terms of biology, I'm interested in synthetic biology, RNA, development and genetic regulatory networks. On the computational side, I'm still interested in pretty much all ways of modeling biological systems, whether it's via differential equations, Bayesian networks or stochastic approaches etc. That list of topics is, of course, non-exhaustive -- I can think of a bunch of other areas I'd like to work in [like malaria, immunology, protein motors, digital oscillations in cells etc] but it represents the subset that I plan to explore in my thesis.

I spent a lot of time thinking about lab choice, and broke it down into

- Science: can I work on stuff I'm interested in ?

- What I want to learn the most, computational or experimental methods:
whether I want to become a computer scientist who works on biological problems or a biologist who uses computational methods to work on biological problems, as well as where I need the most help/education ie the stuff I'm less likely to be able to pick up on my own.

In terms of science, think I could have satisfied my interests in several labs, whether by joining Drew or Tom's lab and building synthetic networks that mimic some aspect of development, by joining Dave Gifford's lab and working on embryonic stem cell differentiation or in Dave Bartel's lab, working on the impact of microRNAs on development. What really ended up being the deciding factor was the second question, the expertise I most want to gain -- I came down on the "biologist who uses computational methods" side, in the sense that, right now, I think it's more useful for me to be in a lab that has a lot of experimental expertise that I pick up, than in a lab focused on computation.

With that in mind, I ended up choosing to join Dave Bartel's lab, at the Whitehead Institute [which conveniently has a picture of Dave right on its homepage at the moment :-)]. From what I've seen, Dave is a hands-on guy who is very much up-to-date on the experimental work being done by his students and can give concrete advice when experiments are going wrong, and there are lots of folks in his lab who have oodles of benchwork experience. Both of these are good pre-conditions for somebody like me who has no clue about labwork, especially when coupled to the fact that the Whitehead is one of the best-known biological research institutions in the country and so hopefully I'll pick up more knowledge about biology just by osmosis.

That said, my plan is actually to end up being a joint student between Dave Bartel and somebody with a computational focus, but we haven't yet figured out who that second person is. The idea is for me to spend the summer thinking about a thesis project and then use that to decide whose computational expertise to tap into.

So, the first academic year is over and I've picked an advisor. Now all I have to do is the hard part of getting a PhD, the actual research ;-)

1 Comments:

Anonymous philipj said...

Hrm. For no obvious reason, I always assumed MIT had a med school. It certainly seems more "tech" related than, say, linguistics. :)

I'm Canadian, and we're often a little less on the ball in terms of getting new departments (like a Systems Biology virtual department) in hot research areas. Rotations are also not common at all (at least as far as I know), so I'm jealous about both of those things. The rotations in particular sound like a really great way to try before you buy. In these interdisciplinary areas, I'm finding my interests are changing in big ways as I learn more about the biology side of things, and how quantitative ideas can be applied to take new approaches to well known biological phenomena.

One thing I wonder, too, is if you are hoping to end up as a prof at a university, or to go to industry? And, how does that change the skills you want to pick up?

Cheers,
Philip

1:22 AM  

Post a Comment

<< Home