Molecular structure-based methods of property prediction in application to lipids: A review and refinement

The paper is a review of the combined group contribution (GC)–atom connectivity index (CI) approach for prediction of physical and thermodynamic properties of organic chemicals and their mixtures with special emphasis on lipids. The combined approach employs carefully selected datasets of different pure component properties to develop simultaneously two parallel models, one based on group contribution and another based on atom connectivity, for each property. The lipids present in the database are regarded as a separate class, for which special models for pure component properties, primary and temperature dependent, have been developed. For mixtures, properties related to phase equilibria are modeled with GE-based models (UNIQUAC, UNIFAC, NRTL, and combined UNIFAC-CI method). The collected phase equilibrium data for VLE and SLE have been tested for thermodynamic consistency together with a performance evaluation of the GE-models. The paper also reviews the role of the databases and the mathematical and thermodynamic consistency of the measured/estimated data and the predictive nature of the developed models.