GLSDock – Drug Design Using Guided Local Search

The guided local search method has been successfully applied to a significant number of NP-hard optimization problems, producing results of similar caliber, if not better, compared to those obtained from algorithms specially designed for each singular optimization problem. Ranging from the familiar TSP and QAP to general function optimization problems, GLS sits atop many well-known algorithms such as Genetic Algorithm (GA), Simulated Annealing (SA) and Tabu Search (TS). With lesser parameters to adjust to, GLS is relatively simple to implement and apply in many problems. This paper focuses on the potential applications of GLS in ligand docking problems via drug design. Over the years, computer aided drug design (CADD) has spearheaded the drug design process, whereby much focus has been trained on efficient searching in de novo drug design. Previous and ongoing approaches of meta heuristic methods such as GA, SA & TS have proven feasible, but not without problems. Inspired by the huge success of Guided Local Search (GLS) in solving optimization problems, we incorporated it into the drug design problem in protein ligand docking and have found it to be effective.