Quantitative Structure-Activity Relationships Based on Functional and Structural Characteristics of Organic Compounds

Abstract

In the present quantitative structure-activity relationship (QSAR) modeling, organic compounds, including priority pollutants, have been considered and classified based on their functional and structural characteristics. Five physico-chemical characteristics have been used to develop a QSAR model for Pimephales promelas, by means of multiple regression analysis. Collinearity diagnostics was carried out using two different approaches based on condition index and K correlation index. The outlier analysis was carried out using the variable subsets obtained through both the approaches. An attempt has been made to justify the deletion of outliers in each group referring to their physicochemical characteristics. The expressions obtained by using both approaches provide almost the same prediction accuracy, however, the latter approach resulted in expressions with reduced number of molecular descriptors. The QSARs obtained through this exercise would certainly assist in designing environment-friendly molecules with lower toxicity.

Key Words:

• Ecotoxicity
• Collinearity diagnostics
• Condition index
• Variance-decomposition proportions
• K correlation index
• K inflation factor