Extracting knowledge-based energy functions from protein structures by error rate minimization: Comparison of methods using lattice model

We describe a general framework for extracting knowledge-based energy funct ion from a set of native protein structures. In this scheme, the energy fun ction is optimal when there is least chance that a random structure has a l ower energy than the corresponding native structure. We first show that sub ject to certain approximations, most current database-derived energy functi ons fall within this framework, including mean-field potentials, Z-score op timization, and constraint satisfaction methods. We then propose a simple m ethod for energy function parametrization derived from our analysis. We go on to compare our method to other methods using a simple lattice model in t he context of three different energy function scenarios. We show that our m ethod, which is based on the most stringent criteria, performs best in all cases. The power and limitations of each method for deriving knowledge-base d energy function is examined. (C) 2000 American Institute of Physics. [S00 21-9606(00)51844-3].