THE GUTS OF ROSETTANow don't get me wrong; impressionism is my favorite art style, but somehow I am always going to be a little uncomfortable about a program that relies more on statistics than physics to simulate protein folding. I already have this hang up about models in general which I have articulated before. Although modeling reality is what models are supposed to do, ultimately you can still be in for a nasty surprise if you are not paying too much attention to the actual physics and chemistry behind the molecular interactions.
In the last two CASP meets, David Baker from the University of Washington, using his program Rosetta has come first by a hefty margin in the New Fold category. The success of Rosetta has electrified the protein-folding community.
Yet, there are theorists out there who feel slightly queasy when poking through the innards of Rosetta. Theorists such as Wilfred van Gunsteren, write programs such as GROMOS, which have the richness of 17th century Dutch paintings. Just as Vermeer was fetishistically obsessed with painting every detail of the Dutch bourgeoisie, right down to the hem-line of the chamber-maid's dress, GROMOS is obsessed with modeling every detail of 21st century atomic physics, right down to the quadruple expansion of the electron shells of polarizable atoms. The problem with programs like GROMOS is that they are lumbering giants, bloated programs that devour all the computing that you could ever offer, and still beg for more. Although GROMOS is used for many things, attempts to fold a protein have lurched to a stuttering halt, even after agos of computing time.
Programs like Rosetta, on the other hand, are more like Impressionists paintings, virtuoso dabs of paint that trick the eye into seeing a protein fold in no time at all. For instance, whereas GROMOS fastidiously models all 6 atoms in carbon rings attached to the protein and each atom in the ring is allowed to wobble, Rosetta models the carbon ring as one fat unmovable atom. Water molecules surrounding the protein? No problem, says Rosetta, we'll just ignore them. Rosetta also uses a clever trick by folding similar proteins from different species of animals, and then averaging all the structures to obtain a consensus structure. In reality, when proteins like hemoglobin fold inside your body, they don't get to watch how hemoglobin folds in rats or flies in order to come to a consensus.
As an aside, I have used Rosetta a little and it can be hideously user-unfriendly. Why the authors never sought to collaborate with a software company who would design a nice GUI for it is something I have never understood. Now in spite of the above rants let me not be misleading here; I think Rosetta is a fantastic program that has achieved some spectacular results reported in places like Nature and Science; perhaps its most stunning achievement was designing an enzyme from scratch that would catalyze a Kemp elimination reaction, a reaction that no other enzyme in nature is known to catalyze. It's just that I think that using it, at least for people who are not members of David Baker's group, might be like flying a highly sophisticated spaceship whose workings are somewhat mysterious. It could be a problem when those ill-understood cumulonimbus (or Romulans) start looming on the horizon.
I like the analogy between coarse-grained models and impressionist art. However, one statement you made struck me as a bit strange: "I am always going to be a little uncomfortable about a program that relies more on statistics than physics to simulate protein folding."
ReplyDeleteI don't think the aim of Rosetta is to "simulate protein folding", as in the dynamic process of folding. The idea is to solve for the structure of the folded protein, with no regard for the actual dynamics of folding. Obviously, a fully atomistic folding simulation would be ideal, but that is so far from feasible that other approaches have to be employed. Baker and co-workers have come up with very clever tricks that seem to work much, much better than anything else out there.
Regarding the user-unfriendliness of Rosetta, I think this is a common flaw of programs written for academic research. I suspect that the origin is two-fold:
1. when starting to write a program like that, I imagine the original thought is that no one will use it outside of the research group. Once it becomes clear that the program has real advantages over competitors then the idea of distributing to the "public" comes to the fore, but by that point a complete redesign is unfeasible (though, as an aside, I think Rosetta has been completely written by former Baker group members, though I'm not sure about the user-friendliness of that version)
2. Most programs for academic research are distributed for free, so there is no real incentive to spend the money it would cost to build a user-friendly UI. Plus most research group members are going to be experts in whatever field they are working in, which in general will not be GUI development.