AAAI Publications, Workshops at the Twenty-Seventh AAAI Conference on Artificial Intelligence

Font Size: 
Pareto-Based Optimal Sampling Method and Its Applications in Protein Structural Conformation Sampling
Yaohang Li, Ashraf Yaseen

Last modified: 2013-06-29


Efficiently sampling the protein conformation space is acritical step in de novo protein structuremodeling. One of the important challenges in sampling is the inaccuracy ofavailable scoring functions, i.e., a scoring function is not alwayssufficiently accurate to distinguish the correct conformations from thealternatives and thereby exploring the very minimum of a scoring function doesnot necessary reveal correct conformations. In this paper, we present a Paretooptimal sampling (POS) method to address the inaccuracy problem of scoringfunctions. The POS method adopts a new computational sampling strategy byexploring diversified conformations on the Pareto optimal front in the functionspace consisted of multiple scoring functions, representing consensus withdifferent trade-offs among multiple scoring functions. Our computationalresults in protein loop structure sampling and protein backbone structuresampling have demonstrated the effectiveness of the POS method, wherenear-natives are found in the ensemble of Pareto-optimal conformations.


Protein Structure Modeling; Pareto Optimality

Full Text: PDF