Visualization of multi-dimensional data of bioactive chemicals using a hierarchical data visualization technique 'Heiankyo view'

摘要

In the age of combinatorial chemistry and high throughput screening, large-scale data of bioactive chemicals oriented to drug development are being accumulated. Due to the difficulties inherent in understanding such large quantities of data, information visualization techniques are increasingly attractive. Authors apply 'Heiankyo View', which is the technique for the representation of large-scale hierarchical data, for the visualization of multi-dimensional data of bioactive chemicals. In the present study, we investigated applicability of the visualization technique to the structure-activity relationship (SAR) analyses. The study first classifies chemicals according to similarity in their biological actions through self-organizing map analysis. It then applies a recursive partitioning method to find the relationship between biologically based categories and chemical structure, and finally it stores the drugs as hierarchical data. HeiankyoView is suitable for the visualization of such hierarchicaldata. This paper first describes the algorithmic overview of Heiankyo View, and then provides some example of visualization of multi-dimensional data of bioactive chemicals.