Predictive QSPR Modelling for the Second Virial Coefficient of the Pure Organic Compounds

Abstract

In this article we developed a system of the predictive models for the second virial coefficients of the pure compounds. Second virial coefficient is the property derived from the virial equation of state, and is of particular interest as it describes pair intermolecular interactions. The two-layer QSPR models were developed, which exploited the well-known physical equations and allowed us to include this information into traditional QSPR methodology. This shows some new perspectives for work with temperature-dependent properties. It was shown that 2D descriptors can be successfully used for modeling of complex thermodynamic properties like virial coefficients.

Keywords: Cheminformatics; Linear free energy relationships; Machine learning; Medicinal chemistry; Physicochemical parameters; QSPR; Structure-property relationships; Thermodynamics.