Coarse- and fine-grained models for proteins: evaluation by decoy discrimination

Chris Kauffman and George Karypis
Proteins, 2013
Download Paper
Abstract
Coarse-grained models for protein structure are increasingly utilized in simulations and structural bioinformatics. In this study we evaluated the e_ectiveness of three granularities of protein representation based on their ability to discriminate between correctly folded native structures and incorrectly folded decoy structures. The three levels of representation used one bead peramino acid (coarse), two beads per amino acid (medium), and all atoms (_ne). Multiple structure features were compared at each representation level including 2-body interactions, 3-body interactions, solvent exposure, contact numbers, and angle bending. In most cases, the all-atom level was most successful at discriminating decoys, but the two-bead level provided a good com promise between the number of model parameters which must be estimated and the accuracy achieved. The most e_ective feature type appeared to be 2-body interactions. Considering 3-body interactions increased accuracy only marginally when all atoms were used and not at all in medium and coarse representations. Though 2-body interactions were most e_ective for the coarse representations, the accuracy loss for using only solvent exposure or contact number was proportionally less at these levels than in the all-atom representation. We propose an optimization method capable of selecting bead types of di_erent granularities to create a mixed representation of the protein. We illustrate its behavior on decoy discrimination and discuss implications for data-driven protein model selection
Research topics: Bioinformatics | Data mining