Design and combinatorial library generation of 1H 1,4 benzodiazepine 2,5 diones as photosystem-II inhibitors: A public QSAR approach

Exponential rise in the population around the word increased the demand of food grains/crops with limited expansion of the agricultural land. To meet the demand, generation of new herbicidal agents is of primary need for the manufacturing firm. In silico tool like QSAR is one of the regularly used in designing newer compounds along with wet experiment. Photosystem-II (PS-II) regarded as one of the major target for the herbicidal agents. With this aim in the present study a series of 1H, 1,4 benzodiazepine 2,5-dione analogues as herbicidal (PS-II inhibitors) agents were subjected to QSAR analysis using 2D PaDEL descriptors (open source). Two different splitting techniques namely, kennard stone based and k-means clustering splitting were used to divide the whole data set and GFA based on MAE criteria was used a statistical method to develop a model to investigate the physicochemical and structural requirement of potential PS-II inhibitors. All the models are statistically robust both internally and externally (Q2: 0.540-0.693, R2pred: 0.722-0.810). The activity is mostly affected by polarizabilities, electro negativities as well as substituents at the phenyl ring. Based on the results, a library of compounds was generated using SmiLib v2.0 tool (open source) and better predicted inside applicability domain compounds were identified by applying three different applicability domain (AD) approaches. Therefore the developed public QSAR models may be helpful for the scientific community for the further research.