Predicting a Variety of Constant Pure Compound Properties by the Targeted QSPR Method

The application of a modified version of the Targeted QSPR (Quantitative Structure-Property Relationship, Brauner et al., Ind. Eng. Chem. Res., 45, 8430, 2006) method for predicting a wide variety of constant properties is considered. Prediction of a (target) property for a particular (target) compound is carried out in two stages. The first stage involves the identification of a similarity group and a training set whose members are structurally related to the target compound. In the second stage, a Dominant Descriptor (DD), which is collinear with the target property values for the training set members, is identified. The linear QSPR derived in terms the DD is used to predict the property value of the target compound. By proper adjustment of the training set, the great majority of the constant properties can be predicted within the experimental error level, provided that a sufficient number of predictive compounds, which are similar to the target compound, are included in the data base.