A necessary stage in protein structure determination by NMR is the assignmentof resonances to their originating sites in the polypeptide. Assignment of large proteins (even with the aid of +{13}C and +{15}N labeling) becomes difficult due to large overlap of the numerous resonances, exacerbated by line-broadening effects. Our NMR studies of glutamine-binding protein (molecular weight 25 kDa) motivate the development of an automated assignment method. Several 3D-NMR experiments, performed in the laboratory of Prof. Chie Ho, provide cross peaks which will become the inputs to an automated assignment program . The proposed method evaluates the overall probability of a tentative assignment, and maximizes this score by shuffling peak assignments, using a simulated annealing strategy. The program is tested using data on previously-assigned proteins. A starter grant will enable us to optimize this program, and gain familiarity with the PSC's structure determination software.
Showing the most recent 10 out of 292 publications