In our research, we often deal with bond breaking and formation processes in macromolecules (such as catalysis in enzymes or ATP hydrolysis in motor proteins). Therefore, we need to describe both the electronic structure and nuclear dynamics for the system of interest at a satisfactory level.

The electronic structure problem usually requires a quantum mechanical description, which is prohibitively expensive for a large molecule like enzyme. Fortunately, for the specific question of interest, only a small part of the system (e.g. active site of the enzyme) needs to be described at the level of quantum mechanics while the rest can be dealt with by use of a more approximate theory, such as classical force field. This forms the basis of the hybrid quantum/classical approach. Our group is actively involved in the development of such combined Quantum Mechanical/Molecular Mechanical (QM/MM) methods. Substantial effort will also be paid to develope efficient and accurate methods that describe transition metals and electronically excited states in the context of biomolecules. A recent collaboration with a physicist group in Heidelberg/Harvard on a semi-empirical density functional theory revealed exciting possibilities.

For the nuclear motion, typically classical mechanics (i.e. the Newtonian or Hamiltionian equations of motion) is sufficient. In some cases, however, quantum effects can become important even at room temperature. An important example involves proton transfer reactions in enzyme, in which case zero-point energy and tunneling have been shown to be significant by both experimental and theoretical studies. Much effort, however, is required to obtain an accurate estimate for the effects of quantum mechanics in biomolecules (and more importantly, factors that govern such effects) . We are working towards this goal along the lines of the Feynman path-integral formalism and semi-classical mechanics.

Solvent plays an important role on the dynamics and functions of biomolecules. Modern statistical mechanics is applied to study both the equilibrium and dynamical (non-equilibrium) effects due to solvation. For example, non-linear optical spectroscopy (e.g. Photon Echo) can be used to probe the local dynamics of protein active sites; theoretical calculations can help to reveal the contribution from the solvent. Novel simulation techniques are also necessary to allow the efficient sampling of the configuration space of macromolecules, which is essential to our theoretical investigations.

In conclusion, students in the group will have the opportunity to be exposed to all areas of modern computational/theoretical chemistry: electronic structure theory, quantum nuclear dynamics and equilibrium (or non-equilibrium) statistical mechanics. Furthermore, students will be actively involved in programming (Unix, Fortran and/or C) , which is extremely valuable for their future careers.