The present paper attempts to employ hybrid genetic algorithms (GAs) to solve the flexible-ligand docking problem i.e. predicting the binding conformation of a flexible ligand molecule into a rigid protein. Our hybrid GA scheme uses the concept of Lamarckian genetics to perform a local search about an individual, followed by replacing it with a better solution found in its neighborhood. Two local search schemes have been investigated and their performance relative to the standard GA have been compared. Preliminary results obtained on a set of three protein-ligand complexes are encouraging.
展开▼
