The research in the group is broadly concerned with many-body phenomena in biology, chemistry and physics. A major theme is understanding systems where a large diversity of long lived states is involved, necessitating the use of a statistical characterization of an energy or attractor landscape. The most notable examples are glasses, liquids, biomolecules and biomolecular regulatory networks.
In the area of protein folding we are interested both in describing folding kinetics in the laboratory and the development of bioinformatically based schemes for predicting structure from sequence using computer simulation. A key concept is that the energy landscape of a foldable protein looks like a rugged funnel. This idea guides the development of both simple folding kinetics models and structure prediction algorithms. Our prediction algorithms have already shown success in folding smaller proteins whose structure was previously unknown.
Similar issues of attractor landscapes also arise in higher order biological processes, such as gene recognition and genetic network regulation, which we are beginning to study. The energy landscapes of supercooled liquids and glasses also present interesting problems. We have shown how a new approach based on "random first order transitions" explains many quantitative relations found empirically both in liquids and under cryogenic conditions where quantum effects play a role. The same ideas show promise in the study of systems as different as high temperature superconductors, polymer assemblies, and microemulsions. They may also be useful for describing the three dimensional structure and dynamics of the interior of living cells.