Standalone methodology for building QSAR models: an antioxidant QSAR model of di(hetero)aryl amines and amides as a case study

Abstract

Abstract QSAR modeling is a methodology used in various scientific fields to correlate molecular descriptors to the properties or biological activities of compounds of interest. Several steps are needed to construct a QSAR model, including chemical structure preparation, molecular descriptor calculation and selection, and model building and validation. We present a complete methodology for preparing QSAR models using free and open-source software tools. A detailed step-by-step protocol is provided with the complete process of QSAR modeling, from compound library preparation to statistical validation. A QSAR model was developed as a case study to model the antioxidant activity, particularly the radical scavenging activity of 70 di(hetero)aryl amine and amide compounds. The OCHEM platform was used to calculate the 12,072 molecular descriptors. These molecular descriptors and the experimental pIC50 for each compound were introduced in PyQSAR software, and a genetic algorithm was used to select four molecular descriptors to build the QSAR model: B06[C-O], Eig04_AEA(dm), JGI2 and J_Dz(p). The QSAR model was then implemented by applying multiple linear regression, and a final equation was obtained. The QSAR model presents excellent statistical parameters that verify its robustness and predictability, namely, the correlation coefficient (R2 = 0.8905), the mean score value of the MLR method (Q2CV = 0.8676) and the relative standard deviation of the residuals (RSR = 0.3320 and RSRCV=0.3518). This QSAR model will guide the synthesis of new di(hetero)aryl amines or amides with improved antioxidant activities. All files and the complete protocol are provided to replicate the building of the presented antioxidant QSAR model, and researchers will be able to prepare other QSAR models using different compound libraries and different biological activities.