Intellectual Merit Protein folding pathway prediction is a very important problem since protein misfolding has been identified as the cause of several diseases such as Creutzfeldt-Jacob disease, cystic fibrosis, hereditary emphysema and some cancers. Furthermore, knowledge of pathways for protein folding can give important insight into the structure of proteins. To make pathway based approaches to structure prediction a reality, plausible protein folding pathways need to be modeled and validated. Our novel approach outlined in this proposal is to start with a folded protein in its final state and learn how to unfold the protein in an approximately ordered sequence of steps, to its unfolded state. The reversal of such a sequence then represents a plausible protein folding pathway. Using the known structure and a fast, graph-based algorithm for recursive splitting of the structure along its most energetically labile contacts, we are able to develop a general model for topological unfolding. We propose to explore the effect on the folding pathway of mutations that are known to cause (or suppress) misfolding diseases, especially those associated with amyloid fiber formation. We will calculate the distribution of folding intermediates and then compare these distributions for natural mutatations that are known to cause amyloid accumulation in the cell. Broader Impact This proposal builds on a proven track record of success of the PIs, both in collaborative research and educational efforts. For instance, Bystroff and Zaki have team taught NSF Chautauqua Summer Course on New Directions in Bioinformatics and Biotechnology, held anually at RPI (1999-2004). The PIs have a track record of advising and promoting research activity in under-represented groups. For instance, both PIs jointly supervised a Native American M.S. student, and Zaki has supervised 7 femaleM.S. students. Currently he is supervising 2 female PhD students. Bystroff currently has three African-American undergraduate researchers. The proposed work will be made publically available as a web server, and also open-source software, that predicts the unfolding pathway. Two full-time graduate students will be trained on this grant. Undergraduates, including underrepresented groups, will continue to play a large part in our research efforts. This proposal will support a continuing inter-disciplinary research team, which has resulted in several joint courses and publications. Much of the preliminary work described in this proposal is the fruit of this collaboration. Finally this grant will support one of the nation's strongest teaching programs in bioinformatics.
