Tony Capra will present his preFPO on Monday November 3 at 1PM in Room 401 (note room). The members of his committee are: Mona Singh, advisor; Tom Funkhouser and Olga Troyanskaya, readers; Fred Hughson (MOL) and Chris Floudas (CHE), nonreaders. Everyone is invited to attend his talk. His abstract follows below. --- Predicting functionally important residues in proteins from evolutionary conservation and 3D structure All residues in a protein are not equally important. Some are essential to maintain the proper structure and function of the protein, while others tolerate replacement by different amino acids. Identifying the functionally important sites in a protein is an important step towards characterizing its molecular function and has direct applications to drug design and protein engineering. In this talk, I will describe three projects that substantially improve our understanding of and ability to predict functionally important sites. First, I will present an information theoretic algorithm for predicting functionally important sites using evolutionary conservation. It acheives state-of-the-art performance and is several orders of magnitude faster than other competitive approaches. I will then describe a bioinformatics pipeline we developed to create the first large dataset of specificity determining positions (SDP). This dataset allows us to characterize properties of SDP and demonstrate that existing SDP prediction methods do not perform as well as a simple approach we developed. Last, I will introduce a method that combines our new conservation-based measure with 3D models of protein structures to signficantly improve the identification of ligand-binding sites.