Obtaining a computer-assisted QSAR model for the prediction of anti-inflammatory activity

Abstract

The main objective of this study was to develop quantitative structure-activity relationships (QSAR) for the classification and prediction of anti-inflammatory activity. To this end, the ToSS-MoDE approximation was applied for the calculation of the spectral moments of the adjacency matrix between edges of the molecular graph with suppressed hydrogens, weighted on the main diagonal with moments of bond dipoles, bond distance, Van der Waals radius, polarizability and hydrophobicity to 509 active and inactive compounds. The calculated descriptors were used in the design of a training series and a prediction series. With the training series, a discriminant function was developed for the anti-inflammatory activity and another function to characterize the potential of these drugs using the Multivariate Linear Discriminant analysis, obtaining a good total classification of 96.07%. The model was validated by using the external prediction series, obtaining a good classification of 92.59%.