COMP 242 |
| Protonation states of ionizable residues play an important role in protein structure and function. We have observed that alternative protonation states due to local shifts in pKa can affect the accuracy of computational side chain and loop prediction methods. We have addressed this issue by developing a new algorithm to optimize the protonation state of titratable groups. This new algorithm is integrated with our hierarchical approach to all-atom protein loop prediction such that protonation states are evaluated simultaneously with conformational sampling using the OPLS-AA force field and the Surface Generalized Born solvent model. We have selected high resolution X-ray structures with ionizable residues embedded in short loops (up to 7 residues) that were calculated to have shifted pKa values. Preliminary results show improvements in side chain and loop prediction accuracy when the correct protonation state is accounted for. |
|
Poster Session
6:00 PM-8:00 PM, Tuesday, 12 September 2006 Moscone Center -- Hall D, Poster
Sci-Mix
Division of Computers in Chemistry |