| |
Abstract: The novel C-Jun N-terminal kinase 3(JNK3) inhibitor studied in this article is a promising therapeutic option for Alzheimer's disease, as it is believed to play a significant role in its treatment. The inhibitory effect of certain compounds can be accurately identified by measuring their IC50 values. However, traditional IC50 determination methods require substantial human and material resources. To better understand the Quantitative Structure-Activity Relationship (QSAR), we employed various machine learning techniques, including heuristic method(HM), random forest(RF), support vector machine(SVM), generalized regression neural network(GRNN) to establish a comprehensive QSAR model that better describes this relationship. Among them, the support vector machine model demonstrated the best performance, with R2 and MSE values of 0.914, 0.944, 0.066, and 0.041 on the training and testing sets, respectively. The leave-one-out cross validation(LOO-CV) yielded Q2 values of 0.92 and 0.905 for the training and testing sets, respectively. These results demonstrate that the support vector machine model can be used to predict the inhibitory activity of this novel inhibitor, providing valuable assistance in the design and synthesis of new JNK3 inhibitors based on 2-aryl-1-pyrimydyl-1,4,5,6-tetrahydrocyclopenta[d]imidazole-5-carboxamide derivatives. Furthermore, the five key descriptors revealed in this study will aid in the design and screening of novel drug molecules for Alzheimer's disease.
|
|
|
