BIOINFORMATICS (funded by DARPA)

• Protein annotation. Can we predict the structure and function of a sequence of a unknown protein? We develop a combination of scores from sequence alignment, threading, secondary structure prediction, atomically detailed modeling and more to annotate proteins. A server and software are available.
• Protein evolution. It is known that many sequences fit one fold, but how many?. Are different folds connected by a single point mutation? What are the relationships between sequence evolution and the structural templates in which they evolve? (with Prof. Jon Kleinberg)
• Protein folding potentials. Mathematical programming techniques are used to find optimal sets of parameters that are used in structure prediction calculations. (with Profs Jaroslaw Meller and Michael Wagner)
• Alignments. Can we learn an optimal alignment? Can we parameterize the (usually assumed ad-hoc) gaps and the insertion parameters? (with Prof. Thorsten Joachims)