Applications of neural networks in quantitative structure-activity relationships of dihydrofolate reductase inhibitors.

摘要: Back propagation neural networks is a new technology useful for modeling nonlinear functions of several variables. This paper explores their applications in the field quantitative structure-activity relationships. In particular, ability to fit biological activity surfaces, predict activity, and determine "functional forms" its dependence on physical properties compared well-established methods field. A dataset 256 5-phenyl-3,4-diamino-6,6-dimethyldihydrotriazines that inhibit dihydrofolate reductase enzyme used as basis comparison. It found lead enhanced surface fits predictions relative standard regression methods. Moreover, they circumvent need ad hoc indicator variables, which account significant part variance linear models. Additionally, elucidation "cross-products" effects correspond trade-offs between effect activity. first demonstration latter two findings. On other hand, due complexity resulting models, an understanding local, but not global, relationships possible. The must await further developments. Furthermore, longer computational time required train somewhat inconveniencing, although restrictive.