High Resolution Protein Structure Prediction

I'm currently working with several teams of students, to (1) develop a new software platform (SHARPEN) for protein structure refinement, as well as protein design applications and (2) start running these calculations on the Folding@Home distributed computing network.

Cytochrome P450 Structure Determination and Library Design

I've been doing a bit of crystallography this last year, crystallizing the heme domain of P450 variants. Also, I've been working on P450 library design for various directed evolution efforts in the group.

Cellulase Library Design

A new project that is getting started: we are collaborating with DNA2.0 to synthesize a library of designed cellulase variants. Stabilized enzymes in this family are relevant for biofuel production.
I made some explanatory movies:
  • explaining the basics of our target enzyme: cellobiohydrolase II
  • Enzyme chimeras via recombination

    Graduate Work

    Distributed Protein Folding Kinetics

    As detailed in the Publication page, my graduate work involves folding simulations for several peptides and mini-proteins. The concept is simple - using large ensembles of molecular dynamics simulations we can observe transitions that are typical at very slow timescales.
    One favorite protein is the N-terminal domain of the ribosomal protein L9. I have been running simulations using the Folding@Home distributed computing cluster on this system for over two years. We have generated an unprecedented aggregate simulation time, with tens of thousands of personal computers simulating the molecular dynamics of this molecule each day.

    After the Ph.D., I wrote an essay Convergence of Experiment and Simulation summing up the thesis work. And here is my thesis

    Undergraduate projects

    I've now been in this field for 10 years(!) since I was a freshman at MIT. Below are a few graphics from projects along the way. . .
    One of the great aspects of computational biology is the limitless variety that nature has provided in the way of biological macromolecules. How they move, how they assemble, how they interact with other molecules including ligands or drugs. The long term goal is to develop our models and methodology to the point where we may routinely design new proteins, new enzymes, new protein-protein interaction partners, and molecular machines.

    Orotidine-5-Monophospate Decarboxylase: An example of an extraordinarily effective enzyme, with a controversial mechanism.

    Inhibitor Design against Dihydroorotate Dehydrogenase from the Malaria-causing Parasite Plasmodium Falciparum. This diagram was generated during an unusual undergraduate course, which combined the efforts of modelers, organic chemists, and molecular biologists to design novel combinatorial malaria inhibitors.


    My introduction to computational biology came through the study of protein electrostatics. I was lucky to work with Bruce Tidor and collaborate with Dan Raleigh as an undergraduate. At Stanford, I was fortunate to collaborate with Ian Suydam from the Boxer lab and renew my interest in electrostatics for that comparison of experiment and modeling.