Hi! I'm Mark, a research assistant professor at TTIC working on protein
and drug design algorithms. I've been interested in protein design since
the tenth grade, when I made a paper model of an enzyme (a ribonuclease) to see if I could
figure out what mutations would make it specific to digest viral RNA.
I realized soon that much more sophisticated
techniques were needed, and as a graduate student in Bruce Donald's lab at Duke University, I began working on such techniques.
The field has already made a lot of progress, and I
think I've made some significant contributions (see below).
But there is still a significant gap between the model of reality that we currently use in our
designs, and the actual reality dictated by the laws of physics that
determines how proteins and drugs behave in nature. Closing this gap would
let us design molecules like an architect designs buildings—we will know
ahead of time how they will work in real life. Builders building a
skyscraper don't expect their first several tries to fall down, and we
shouldn't have to expect that for drug design either. I think we
can learn to design proteins and drugs in a similarly systematic way, even for
fairly complicated functions.
Some algorithms I have developed toward this goal: EPIC (Energy as Polynomials in Internal Coordinates) DEEPer (Dead-End Elimination with Perturbations) COMETS (Constrained Optimization of Multistate Energies by Tree Search) LUTE (Local Unpruned Tuple Expansion) CATS (Coordinates of Atoms by Taylor Series) I have implemented all these algorithms as part of the OSPREY open-source protein design software package. |