Nonlinear mapping techniques for prediction of pharmacological properties of chemical compounds

摘要

The review concerns the state of the art in the development and application of a special group of data analysis methods called mapping algorithms, which allow one to reduce the dimension of multidimensional spaces of the properties of objects under study (series of chemical compounds) down to corresponding two- or three-dimensional projections (images) with retention of the initial space topology. It is shown how can such maps be used to visually analyze complex structure—property relationships in order to predict the pharmacologically significant properties of chemical compounds. The bibliography includes 86 references.