A genetic algorithm for the ligand-protein docking problem

Camila S De Magalhães,Hélio J.C Barbosa,Laurent E Dardenne

doi:10.1590/s1415-47572004000400022

Camila S De Magalhães, Hélio J.C Barbosa + Show 1 more

Open Access

https://doi.org/10.1590/s1415-47572004000400022

Copy DOI

Abstract

We analyzed the performance of a real coded "steady-state" genetic algorithm (SSGA) using a grid-based methodology in docking five HIV-1 protease-ligand complexes having known three-dimensional structures. All ligands tested are highly flexible, having more than 10 conformational degrees of freedom. The SSGA was tested for the rigid and flexible ligand docking cases. The implemented genetic algorithm was able to dock successfully rigid and flexible ligand molecules, but with a decreasing performance when the number of ligand conformational degrees of freedom increased. The docked lowest-energy structures have root mean square deviation (RMSD) with respect to the corresponding experimental crystallographic structure ranging from 0.037 Å to 0.090 Å in the rigid docking, and 0.420 Å to 1.943 Å in the flexible docking. We found that not only the number of ligand conformational degrees of freedom is an important aspect to the algorithm performance, but also that the more internal dihedral angles are critical. Furthermore, our results showed that the initial population distribution can be relevant for the algorithm performance.

Highlights

With the increasing amount of molecular biological structures available, docking approaches have been very important and useful tools in structure-based rational drug discovery and design (Gane and Dean, 2000)
For a protein/receptor with known three-dimensional structure, the ligand-protein docking problem basically consists in predicting the bound conformation of a ligand molecule within the protein active site
The algorithm success is measured by the root mean square deviation (RMSD) between the crystallographic conformation and the conformation found by the algorithm

Summary

Introduction

With the increasing amount of molecular biological structures available, docking approaches have been very important and useful tools in structure-based rational drug discovery and design (Gane and Dean, 2000). For a protein/receptor with known three-dimensional structure, the ligand-protein docking problem basically consists in predicting the bound conformation of a ligand molecule within the protein active site. The docking problem is a difficult optimization problem involving many degrees of freedom, and the development of efficient docking algorithms and methodologies would be of enormous benefit in the design of new drugs (Marrone et al, 1997). One of the major problems in molecular docking is how to treat the protein and the ligand flexibility, taking into account hundreds of thousands of degrees of freedom in the two molecules. Some docking programs treat the receptor and the ligand as

Methods

Results

Conclusion