1. [Untitled]
- Author
-
Abilev Sk, Zefirov Ns, Gal'berstam Nm, Baskin, Paliulin Va, and Liubimova Ik
- Subjects
Quantum chemical ,Artificial neural network ,Chemistry ,Stereochemistry ,Chemical structure ,fungi ,General Biochemistry, Genetics and Molecular Biology ,Partition coefficient ,Nonlinear system ,Linear regression ,Molecule ,Multiple linear regression analysis ,General Agricultural and Biological Sciences ,Biological system - Abstract
The relationship between mutagenic activity and chemical structure was studied for 54 polycyclic compounds using two approaches: multiple linear regression analysis and artificial neural networks. Structural fragments, quantum chemical indices, and hydrophobicity (octanol–water partition coefficient) were used as descriptors (properties of the molecules introduced in the model). Both linear regression equations and nonlinear relationships obtained with the help of a neural network were shown to accurately predict mutagenic activity for the compounds structurally similar to those in the training sample. The introduction of experimentally selected descriptors is substantiated to verify the proposed mechanism of related compounds mutagenic activity.
- Published
- 2001
- Full Text
- View/download PDF