Abstract
The chapter reviews various approximate and rigorous force-field-based methods that can fit into such a hierarchical approach. In practical drug design, it should be beneficial to use a hierarchical approach in which simple but fast methods are first used for preliminary screening of real or virtual compound libraries and for suggesting potentially useful drug candidates to make/screen, followed by further evaluations by models with increasing levels of sophistication. With the growing availability of inexpensive computer clusters, it should also be increasingly easier to carry out free energy perturbation calculations on at least a small number of potential drug candidates that deserve more thorough investigation. Intermediate models varying between the extremes of elaborate explicit solvent models and simple fixed-conformation molecular mechanics models have also been developed to fit into the hierarchical approach to drug design. These models include those that aim at deciphering the determinants of molecular recognition and using them to construct rules to aid drug design (e.g., sensitivity analysis), those that focus on first exploring a smaller chemical subspace (e.g., single-step Zwanzig perturbation method), those that mix empirical parameters with simulation data (e.g., semi-empirical linear response theories), and those that use a combination of explicit and implicit solvent models (e.g., Molecular Mechanic/Poisson Boltsmann method). Force-field methods have also begun to play a larger role in accounting for protein flexibility in virtual screening (e.g., the dynamic pharmacophore method and the relaxed complex methods).
Original language | American English |
---|---|
Journal | Advances in Protein Chemistry |
Volume | 66 |
DOIs | |
State | Published - Jan 1 2003 |
Externally published | Yes |
Disciplines
- Biology
- Bioinformatics
- Molecular Biology
- Artificial Intelligence and Robotics