[Computer classification models on the relationship between chemical structures of compounds and drugs with their blood brain barrier penetration].

Ability of drugs to cross blood-brain barrier (BBB) (BBB+ for BBB-penetrating and BBB- for non-penetrating compounds) is one of the most important properties of chemicals acting on the central nervous system (CNS). This work presents the results of modelling of the relationship between chemicals structure and BBB-crossing ability. The data set included 1513 compounds BBB+/- (1276 BBB+ and 237 BBB-). Computer modelling of structure-activity relationship was realized by two directions: using the "read-across" method and linear discriminant analysis (LDA) based on physico-chemical descriptors. It was found that a sum of donor-acceptor factors is the principal parameter, which define BBB penetration.