Pixar Meets Molecular Biology

December 15, 2009 6:51 AM

The crowded, complex environment inside living cells makes understanding spatial relationships difficult for biologists. Now, 3D animation software Maya is being used not just for illustration, but to see how our intuition holds up.

E.g. an animation of the process of endocytosis (Quicktime movie) by Janet Iwasa and Tomás Kirchhausen.

posted by jjray (13 comments total)

Look really cool; unfortunately I can't stream the video on this computer but will do so as soon as I can get to a machine that hasn't been crippled by our IT department.
posted by TedW at 7:11 AM on December 15, 2009

I did this watered down about a million times using Excel a while back. It's interesting how big the discrepancy is between your looking at simple phenomenon and then trying to scale to complex phenomenon.

It's also interesting how tightly some people hold onto their preconcieved notions.
posted by Kid Charlemagne at 7:29 AM on December 15, 2009

I'm not sure how useful Maya will be to molecular biologists. Sure, the resulting animations look good, but you still need a specialist to create them.
posted by demiurge at 7:49 AM on December 15, 2009

Wow! Pretty!
This looks like somewhere in between molecular modelling and scripted animations. The Nature article mentions that this approach is used to gain information on the level of the whole by loading the information of individual parts into Maya. That would mean that Maya is a physics simulator, as well as an scripted animation renderer.

Two questions come to mind:
1) What if the parts are not static/rigid objects, but can have conformation changes as well? Define smaller and smaller sub-particles, until we end up with rigid parts?
2) What´s the scale of simulation we can manage using this technique? Simulation on the level of cells rather than a sub domain of a bi-layer would be awesome, but I suspect that it is far beyond what we can manage a the moment. This movie illustrates what I mean, but looks to me like a scripted (not modeled) animation.
posted by Zigurana at 7:57 AM on December 15, 2009 [1 favorite]

A specialist? Like someone who knows both molecular biology and how to use Maya? Oh, where will we ever find a geek that's geeky enough to be interested in two areas of study?
posted by scrowdid at 8:05 AM on December 15, 2009 [1 favorite]

wait, is it called Maya again? I thought they switched it to Alias. FUCK, I can never keep up with this shit.
posted by shmegegge at 8:15 AM on December 15, 2009

That would mean that Maya is a physics simulator, as well as an scripted animation renderer.

Maya does do some basic physics simulation involving gravity, inertia, and the like. It does some cool things with water droplets and hair that demonstrate that pretty well. If you've seen Madagascar then you have some basic idea of what Maya does. All the hair in that movie was actually a pretty interesting step forward in realistic animation and rendering using their embedded physics simulation software.
posted by This Guy at 8:20 AM on December 15, 2009

scrowdid: I'm not saying there's no one who knows molecular biology and can use Maya. I'm saying that there are a lot of molecular biologists out there and only a handful of them are going to be able to make these kind of movies using Maya.

I'm a proponent of simpler, interactive visualization software rather than pre-rendered, scripted movies, but that's what I do, so I'm biased.
posted by demiurge at 8:28 AM on December 15, 2009

I'd just like to point out that these animations gloss over a great deal. Simulating the folding of even one protein by itself is a complicated thing, but simulating the interactions between huge numbers of proteins is stupendously difficult, and I think out of reach of the budgets of most scientists. You can see it in these kinds of animations. In the clathrin example, the first one seems to bumble around randomly until it hooks onto its target, but the subsequent clathrin proteins are drawn towards it like a magnet. Once the vesicle has pinched off from the cell membrane (I'm not 100% sure, but I think other proteins need to get involved for that), the proteins necessary to disassemble the clatherin are drawn straight towards the vesicle, again, magnet style. This is because depicting the action of diffusion would be both boring and confusing (too much stuff on the screen). The point is not that this process as depicted is somehow wrong (from what little I know, it seems right on), but that these animations don't represent some sort of honest simulation of a model of protein dynamics. They are an illustration of what one or more scientists think may be happening. As such, their purpose is for teaching or describing, not discovering. Any physics simulator components are purely to lend believability to the apparent motion of virtual objects in the movie.
posted by Humanzee at 8:40 AM on December 15, 2009 [4 favorites]

Keep in mind also that molecular reactions and interactions take place on a femtosecond scale -- and our study of biology at this speed is incomplete, to say the least -- so there's that level of artifice in any visualization of this kind.
posted by Blazecock Pileon at 10:51 AM on December 15, 2009

I can't check out the animations right now, but I do want to say that Mayas limits re really only defined by the programmer making the Mel script. Maya is animation software, sure. But it is also a highly customizable framework for automating and manipulating polygons. It wouldn't surprise me in the least if someone could code a Maya-based folding@home client, for instance.
posted by hellphish at 11:23 AM on December 15, 2009

I am really interested in this. I can't wait to dig into it after work.
posted by bunnycup at 11:57 AM on December 15, 2009

A specialist? Like someone who knows both molecular biology and how to use Maya? Oh, where will we ever find a geek that's only geeky enough to be interested in just two areas of study?

FTFY
posted by Kid Charlemagne at 2:03 PM on December 15, 2009 [1 favorite]