简介：

The heart of any drug discovery effort lies in elucidating structure-activity relationships (SARs), whether the endeavor is based on testing a handful of compounds or assaying thousands of compounds by high-throughput screening. Distill™ is a hierarchical clustering tool that classifies compounds according to their common substructures and organizes the results in a display that enables visualization of SARs.

Distill's clustering algorithm creates a hierarchy in which each node contains compounds that share a common structural core. The results are presented graphically in an interactive dendrogram, where each node represents a substructure and the set of compounds containing that substructure. By averaging a property such as activity for the compounds at a node, and then coloring each node according to this average, Distill can visually relate substructures to activity. Navigation tools facilitate movement within the hierarchy.

说明文档下载

Download Detailed Distill Information (560k)

Distill's interface showing a dendrogram color-coded by activity for a validation test set, and the linked display of the common substructures.

Key Benefits

• Because Distill orders and sorts compounds the way a medicinal or synthetic chemist would, researchers can objectively evaluate their own conclusions about structure-activity relationships for a given receptor.
• Distill exploits the power of the human eye for pattern recognition of chemical structures, making it possible to easily formulate SAR hypotheses from overwhelming volumes of data.
• Unlike other clustering tools that relate descriptors to properties, Distill relates the components of structure (atoms, bonds, and connectivity) to a specific property, facilitating the next generation design by chemists.
• Users can select nodes in the dendrogram from which Distill constructs UNITY^® 2D search queries in order to identify similar interesting compounds from another database.