Title: "Predicting Protein Function by Surface Comparison"
Speaker: Sael Lee, Department of Biology/Computer Science, Purdue University
Place: HORT 117; February 15, 2011; Tuesday, 4:30pm


Prediction of functions for proteins that the structures are known but functions are not determined is one of the tasks in structural bioinformatics. The informatics approach of function prediction deals with inferring the function of a protein from similar proteins that the functions are already known. There are several features that can be compared for the similarity analysis including the amino-acid sequence and the protein tertiary structure. Sequence comparison can provide fast and accurate function predictions when the similarities of the proteins are high (higher than 30% sequence similarity), however for proteins of remote homologues, structural information can provides better bases for predictions. Protein surface is one of the structural features that provide more direct information about the interaction region to other biomolecules which in turn can be compared to predict the function.

Our group has been working on structure based function prediction using the surface properties such as shape, hydrophobicity, and electrostatic potential which determine the recognition process of other proteins, ligands, DNA, or other molecules with which proteins interact. To compare the protein surface, 3D Zernike descriptor, one of the object abstraction methods used in the graphics field, is used to efficiently represent and compare protein surfaces. Discussion will be made on the properties of 3D Zernike descriptor and selected results on high throughput global structural search and binding ligand prediction.

Associated Reading:

[1] Sael, Lee, Bin Li, David La, Yi Fang, Karthik Ramani, and Raif Rustamov, Kih. "Fast protein tertiary structure retrieval based on global surface shape similarity." Proteins: Structure, Function, and Bioinformatics 72, no. 4 (2008): 1259-1273.

[2] Sael, Lee, and Daisuke Kihara. "Binding Ligand Prediction for Proteins Using Partial Matching of Local Surface Patches." International Journal of Molecular Sciences 11, no. 12 (December 2010): 5009-5026.

Click here for a full schedule of BIOINFORMATICS SEMINARS, past and present.