Energy landscape paving with local search for global optimization of the BLN off-lattice model

Abstract

The optimization problem for finding the global minimum energy structure is one of the main problems of protein structure prediction and is known to be an NP-hard problem in computational molecular biology. The low-energy conformational search problem in the hydrophobic-hydrophilic-neutral (BLN) off-lattice model is studied. We convert the problem into an unconstrained optimization problem by introducing the penalty function. By putting forward a new updating mechanism of the histogram function in the energy landscape paving (ELP) method and incorporating heuristic conformation update strategies into the ELP method, we obtain an improved ELP (IELP) method. Subsequently, by combining the IELP method with the local search (LS) based on the gradient descent method, we propose a hybrid algorithm, denoted by IELP-LS, for the conformational search of the off-lattice BLN model. Simulation results indicate that IELP-LS can find lower-energy states than other methods in the literature, showing that the proposed method is an effective tool for global optimization in the BLN off-lattice protein model.