ABSTRACT<br/>Proteins are large biomolecules that possess motion along specific modes and inherent flexibility. Protein flexibility is important to protein function and plays a key role in protein/protein interactions, especially those mediating cellular communication and anchorage to the extracellular matrix. Thus, to understand protein function, knowledge of not only structure, but also protein internal dynamics is important. As experimental and computational study of protein dynamics is oftentimes technically difficult and resource intensive, prediction of flexibility and motions along specific nodes, will play an increasingly key role in large scale predictions of protein function. While prediction of protein flexibility directly from sequence is an area of active research, many effective tools are available to solve the related problem of predicting the likelihood that any given part of a protein is disordered. We have developed a heuristic that uses predicted residue-wise likelihoods of disorder to add weights to predicted contact maps, rendering those maps suitable for contact modeling of Lipari-Szabo order parameters and for normal mode analysis. We report our progress in testing our heuristic using disorder and contact map predictions from the Scratch Protein Server. Preliminary results indicate that contact models using disorder-weighted predicted contact maps perform comparably to contact model predictions of backbone N-H Lipari-Szabo order parameters made using experimentally determined protein structures. For example for Murine Ets-1DN301, the RMSD and Pearson correlation coefficient between order parameters predicted using our method vs. experimental (BMRB ID 5991) order parameters are 0.121 and 0.782, respectively. These results compare favorably to the RMSD and correlation coefficient between order parameters obtained using a contact model (Zhang & Brüschweiler, J. Am. Chem. Soc. 2002, 124, 12654-12655) based on the high resolution structure (PDB ID 1R36) vs. experimental data, which are 0.113 and 0.483.

Heuristic of predicting protein flexibility and motions along specific nodes

251st National Meeting (2016)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.