Friday, August 26, 2011

Random notes for 26 August 2011

I don't have anything individually notable happening lately but I thought I'd talk about a few different things. Yesterday lunchtime I talked to some folks at Harvard Medical School who are cooking up a great little open source project, and I'm hoping to contribute to it. Their idea is to use Blender to make it simple and easy and quick for medical researchers to put together animations involving multiple proteins interacting. How things have been so far is that making an animation is a colossal hassle (I know this from experience) and usually the researcher is way too busy doing science, so any animation would be done by public relations folks as a means to communicate the research to the non-scientific public.

If animation were so quick and easy and painless that the researcher could do it himself or herself, then researchers could share animations with one another, modify another researcher's animation, append commentary or publication references or other metadata (I am envisioning something like a git repository with forks), and animations would become more than just a PR tool. They would become a part of the active scientific literature, and they would make it possible for researchers to dig deeper into problems, to have more detailed and nuanced communications with one another, and ultimately for better science to be done.

They've decided to attack the problem of simulating and animating several proteins (or other similar sized structures like cell membranes) interacting simultaneously. Most software tools in this area are designed to deal with only one large molecule at a time. But disease processes often involve interactions. Think about viral self-assembly where a bunch of pieces come together to form a protein shell around some RNA. Think about ligands binding with receptors. These involve two or more large molecules, and you need to keep track of their position, their orientation, their mechanical properties, their electrostatic properties, their relative linear and angular momenta. It's a yummy area of inquiry, and done correctly, this could be a significant advance to science, in an area that might easily be dismissed as "pretty pictures".

So yeah, I'm pretty psyched about their project. But there is other news. I'm saddened to learn that Steve Jobs is stepping down as Apple's CEO, presumably for health reasons. It's a sad sad thing that we don't have a better handle on cancer, AIDS and other big diseases. As an engineer in a society that considers itself advanced, I find it a little embarrassing that we've done so poorly. I'm also embarrassed that our economy distorts the motivations for dealing with these -- pharmaceutical companies make bigger profits "treating" diseases than curing them. And don't get me started with medical insurance, or the FDA approval process.

On to happier topics. At my job we are starting to use Flask for some things, which is a sort of simplified Django. Neat stuff. And we're tinkering with Mongo and Redis, both of them fascinating NoSQL databases. Very very cool. I need to think of some interesting home projects for these.