Molecular Modeling

In general, models are very popular in chemistry, especially in modern drug design. Computer graphics are a versatile tool for displaying structures and models along with various properties that can be assigned and geometrically superimposed on these molecules. Structures can be calculated by starting from first principles and trying to get as close to the physics as possible. This is done using quantum mechanical calculations. Since these methods can easily become computationally intensive, empirical approaches are an alternative. They are based on much simpler physics, usually classical mechanics, and treat molecules as a set of point charges in space connected by springs following harmonic potentials. Empirical approaches can only be used if sufficient experimental data are available to parameterize and calibrate the empirical concepts. Therefore, large databases have been developed to collect knowledge about molecular properties. Meanwhile, attempts are also being made to calculate small building blocks of molecules quantum-chemically to build a large database that can then be used to parameterize force-fields using machine learning methods. Molecular mechanics to compute the geometry of molecules are based on empirical force fields. They comprise multiple energy terms that describe mutual interactions either through bonds or through space. Special potentials are used to describe the torsional barrier to rotation around single bonds and non-bonded interactions. The accuracy and required computational power of quantum-chemical approaches depend on the sophistication of the basis sets of atomic or molecular orbitals used for the calculations. Parameterization of some parts of the calculations with empirical data can significantly reduce the computational requirements. Density functional theory is a faster approach that works with electron density distributions instead of orbitals. Combinations of quantum-chemical methods and force-field approaches have been developed to handle large systems such as protein-ligand complexes. Different types of surfaces have been defined, such as the van der Waals surface or the solvent-accessible surface. Properties such as charges can be represented on the surface of molecules. Molecular dynamics simulations are typically based on potentials derived from empirical force fields. They consider the properties of a molecule under dynamic conditions by solving Newtonian equations of motion. As a result, the motion of a molecule over time can be evaluated by analyzing the so-called molecular trajectory. Molecular dynamics simulations can be used to study the flexibility of a protein next to its ligand-binding site. Such simulations can reveal multiple conformations of the protein capable of accommodating different ligands. Computer simulations can be used to enumerate the possible properties of molecules under different experimental conditions. They help to interpret the results of experiments or to predict the properties of molecules to better plan the next experiments. https://sn.pub/bwhyti