Abstract

Chemistry, as the central science, utilizes models in virtually every aspect of the discipline. Integral to the progress of chemistry has been its ability to draw from physics, mathematics, statistics, and computer science to develop new subdisciplines, such as computational chemistry. As computing hardware has become faster and more accessible, so to have techniques to perform modeling and simulations of molecular systems. Software systems today assist researchers in the study of molecular systems and provide mechanisms for deriving a rigorous and consistent explanation for the chemical or biological behavior observed or help the researcher to develop a model for predictions.

To appreciate the origins of molecular modeling, we examine how computational chemistry came to be developed. From its roots in modern quantum theory, in the early twentieth century, through the evolution of empirical, semiempirical, and nonempirical (or ab initio) methods, computational chemistry or molecular modeling techniques have grown from being the province of an esoteric specialist to a “must have” tool in the chemical, pharmaceutical, and polymer industries today.

This article covers the major techniques of molecular modeling, including semiempirical and ab initio methods (eg, MNDO,AM1,PM3, SPARTAN), empirical force field or molecular mechanics techniques, molecular dynamics, and Monte Carlo simulation techniques (eg, MM3/4, MMFF, AMBER). Also reviewed are basic methods involved in the development of quantitative structure–activity relationships (QSAR). These include principal components analysis (PCA), partial least squares (PLS), comparative molecular field analysis (CoMFA), and 3D database searching. The article closes with a table listing representative software systems for computer-assisted molecular modeling (CAMM).