QUANTITATIVE STRUCTURE–ACTIVITY RELATIONSHIP (QSAR) STUDY OF LIVER TOXIC DRUGS

Abstract

Drug-induced liver injury (DILI) is one of the most severe adverse effects (AEs) causing life-threatening conditions, such as acute liver failure. t has also been recognized as the single most common cause of safety- related post-market withdrawals or warnings Due to the nature and idiosyncrasy of clinical forms of DILI, attempts to develop new predictive approaches to evaluate the risk of a medication being a hepatotoxicant have been difficult. The FDA Adverse Event Reporting System (AERS) provides post-market data illustrating AE morbidity. A quantitative structure –activity relationship (QSAR) model for DILI prediction with satisfactory output is urgently needed. In this study, we documented a high-quality QSAR model for predicting the hepatotoxicity risk of DILI by integrating the use of eight effective classifiers and molecular descriptors given by the VlifeMds program. For the present QSAR study, data set of 99 compounds (withdrawn and approved drugs) collected from different databases were taken. Multiple linear regression and partial least square analysis methods had developed two dimensional QSAR models, and then validated for internal and external predictions. The 2D QSAR model developed was statistically important, and was highly predictive. The validation methods presented essential statistical parameters that proved the model's predictive ability. The developed 2D QSAR model revealed the significance of SsssCE-index, SsOHcount, SsssNcount and SdssPcount descriptors. These findings will prove to be an important guide for furtherdesigning and developing new hepatotoxicity activity.