How has molecular modeling technology evolved?

About methods for studying chemical reactions, designing new materials and quantum chemistry

Molecular modeling is used in two reference planes. Firstly, this is the use of various theoretical methods to rationalize the available experimental phenomena and data on the structure and reactions of chemical substances, as well as the energies of the processes associated with the association or dissociation of atoms, ions and molecules. A detailed comparison of calculated data with experimental data provides an objective critical assessment of the theoretical tool used. Secondly, the calculation methods certified in this way are used to predict the molecular and  crystal structures of chemical compounds, which serves as the basis for the design and synthesis of new biologically active compounds and materials.

Milestones in the development of molecular modeling

a) Quantum-chemical methods (non-empirical methods, ab initio). The creation of quantum chemistry based on quantum mechanics is one of the milestones in the science of the twentieth century. At present, quantum mechanics is the most general and well-founded theory of the structure of chemical compounds. Applications of quantum mechanics to molecular calculations (quantum chemical calculations) is based on the solution of the Schrödinger equation (Nobel Prize in Physics 1933), which he derived in 1925. The key component of the equation is the wave function, which describes the state of the atom. The exact solution of the Schrödinger equation is possible only for the hydrogen atom. For other systems, the equation is solved approximately. The main method used for the approximate solution of this equation is the Hartree-Fock-Roothan self-consistent field method. The next important stage in the development of quantum chemical calculations is the creation of the density functional theory by Kohn and Popl in 1964 (Nobel Prize in Chemistry 1998). The authors of the method theoretically substantiated the possibility of calculating the electronic structure of molecules without solving the complete Schrödinger equation and showed that that the total energy of a system can be expressed in terms of the distribution of its electron density. Despite the fact that modern methods of quantum chemistry are recognized as a tool equal to experimental or analytical methods, due to their resource intensity, today they can only be strictly applied to small systems.

b) Semi-empirical methods. The development of these methods has been going on since the 1950s and is associated primarily with the high resource intensity of quantum chemical calculations. The main contribution to the development of semi-empirical methods was made by such scientists as Parr, Popple and Dewar. Semi-empirical methods contain significant simplifications in the quantum-chemical interpretation of the interactions of atomic nuclei and electrons, and also include parameters obtained experimentally – for example, from spectroscopy or when determining the ionization potential from various electronic states. The results of applying semi-empirical methods strongly depend on the molecular system under study.

c) Molecular mechanics (MM) and molecular dynamics (MD). Allinger began to develop the method of molecular mechanics from the 1960s. In MM calculations, a molecule is considered as an extended mechanical model consisting of a set of point masses located in a force field represented by a set of potentials. The terms of the MM force field include elements responsible for the equilibrium values ​​of bond lengths, bond and torsion angles, as well as Coulomb charges on atoms and van der Waals dispersion interactions. The MM is a completely empirical method, since experimentally determined sets of parameters are used in the force field components. Now there are many different force fields. Some experimentally inaccessible parameters for force fields are obtained from quantum chemical calculations.

MD is a combination of the MM force field and classical Newtonian laws of dynamics. In contrast to the static models of all the methods listed above, the advantage of MD is that it becomes possible to analyze the evolution of a molecular system over a certain period of time. It is important to emphasize that the results of calculations in MM and MD of the energies of intermolecular interactions, which are the most important task of modern molecular modeling, depend very strongly on the choice of the force field.

d) Combined methods. When analyzing large molecular ensembles, it is often necessary to find a compromise between the resource intensity and the quality of the calculation method. In the process of such a search, researchers came up with a number of non-standard solutions that combine two or more calculation methods in one model. For example, in the 1990s, Morokuma first proposed using a “multilayer” quantum-chemical-molecular-mechanical model to calculate a molecular system. In his model, different parts (“layers”) of the system are analyzed by different levels of theory, called basis sets – large for the most interesting part of the system and small for peripheral, accepted as less important parts of the molecular ensemble. Now one can find many publications in which the authors combine quantum chemical theory with molecular dynamics in the most diverse way. For example, the charges on the atoms of the molecule are calculated using a reliable quantum chemical basis set, and then the resulting charges are used in MD. There are already a large number of examples demonstrating that the competent use of such combinations of computational methods makes it possible to reproduce experimental data with a high degree of reliability.

Prospects for the development of molecular modeling

Given the fact that the quantum chemical model is the most reliable description of any molecular system, the development of molecular modeling is associated primarily with the development of computer technology. As mentioned above, there is now an active trend in the development and use of a wide variety of combined methods for calculating large molecular systems. There is every reason to believe that it will continue in the next decade. In addition to developments and combinations of well-known computational tools, new algorithms for using the quantum-chemical apparatus are being developed and will be improved to analyze, first of all, intermolecular interactions, which are the essence of the design of bioactive molecules and materials. An example of such algorithms is the recently developed pixel method.