Received on November 26, 1999; revised on December 11, 1999; accepted on March 21, 2000 | Alexey Lagunin, Alla Stepanchikova, Dmitrii Filimonov and Vladimir Poroikov
The article introduces the concept of the biological activity spectrum, which describes the properties of biologically active substances. The PASS (Prediction of Activity Spectra for Substances) software, developed by Alexey Lagunin, Alla Stepanchikova, Dmitrii Filimonov, and Vladimir Porokov, predicts over 300 pharmacological effects and biochemical mechanisms based on the structural formula of a substance. This software can be used to find new targets for ligands and vice versa. A WWW interface for PASS has been created, allowing users to predict the biological activity spectra of substances online. The software uses multilevel neighborhoods of atoms (MNAs) as descriptors to represent chemical structures and calculates the probability of various activities using a specific algorithm. The predictive accuracy is around 89% (leave-one-out cross-validation), and the program can predict biological activity spectra for 319 types of pharmacological effects. Users can input a standard Molfile or draw a structural formula to obtain the predicted biological activity spectrum, which includes the likelihood of each activity being revealed or not. The article provides an example of the predicted biological activity spectrum for nipradilol, highlighting its ambiguous interactions with beta-1 adrenoceptors.The article introduces the concept of the biological activity spectrum, which describes the properties of biologically active substances. The PASS (Prediction of Activity Spectra for Substances) software, developed by Alexey Lagunin, Alla Stepanchikova, Dmitrii Filimonov, and Vladimir Porokov, predicts over 300 pharmacological effects and biochemical mechanisms based on the structural formula of a substance. This software can be used to find new targets for ligands and vice versa. A WWW interface for PASS has been created, allowing users to predict the biological activity spectra of substances online. The software uses multilevel neighborhoods of atoms (MNAs) as descriptors to represent chemical structures and calculates the probability of various activities using a specific algorithm. The predictive accuracy is around 89% (leave-one-out cross-validation), and the program can predict biological activity spectra for 319 types of pharmacological effects. Users can input a standard Molfile or draw a structural formula to obtain the predicted biological activity spectrum, which includes the likelihood of each activity being revealed or not. The article provides an example of the predicted biological activity spectrum for nipradilol, highlighting its ambiguous interactions with beta-1 adrenoceptors.