From Classification to Regression Multitasking QSAR Modeling Using a Novel Modular Neural Network: Simultaneous Prediction of Anticonvulsant Activity and Neurotoxicity of Succinimides

• DOI: 10.1021/acs.molpharmaceut.7b00582
• Publication Date: 2017-11-04T17:11:38Z
• Authors: Davor Antanasijević; Jelena Antanasijević; Nemanja Trišović; Gordana Ušćumlić; Viktor Pocajt
• Keywords: Succinimides

Succinimides, which contain a pharmacophore responsible for anticonvulsant activity, are frequently used antiepileptic drugs and the synthesis of their new derivatives with improved efficacy tolerability presents an important task. Nowadays, multitarget/tasking methodologies focused on quantitative-structure activity relationships (mt-QSAR/mtk-QSAR) have role in rational design since they enable simultaneous prediction several standard measures biological activities at diverse experimental conditions against different targets. Relating to this very topic, mt-QSAR/mtk-QSAR methodology can give only binary classification models, as such, study regression mtk-QSAR (rmtk-QSAR) model based novel modular neural network (MNN) has been proposed. The MNN uses models input modules, while is performed by output module. rmtk-QSAR successfully developed neurotoxicity succinimides, satisfactory accuracy testing (R2 = 0.87). Thus, proposed method be regarded viable alternative QSAR methodology.