Analyzing docking results by substructure search in euklidean space

CINF 72

Thomas Zuhl, Thomas.Zuhl@biosolveit.de, Marcus Gastreich, marcus.gastreich@biosolveit.de, Christian Lemmen, and Holger Claußen, Holger.Claussen@biosolveit.de. BioSolveIT GmbH, An der Ziegelei 75, 53757 Sankt Augustin, Germany
The scoring problem remains unsolved for molecular docking. Target-specific scoring and other tweaks in the docking process are current work-arounds for this problem. Our Docking Database (DDB) provides a powerful analysis tool for this task.

We have introduced a DDB substructure search in 3D that allows users to analyze spatial arrangements of substructures of docking poses. In addition to simple selection and filtering of particular substructures, the new mechanisms enable users to generate statistics of, for example, the distribution of functional groups within the active site and to test the effects of FlexX-Pharm constraints by eliminating unwanted results.

In order to yield fast analyses, all 3D coordinates of both protein structure and all ligand poses are preprocessed and stored in an underlying ORACLE database. The substructures are specified as SMARTS expressions.

We will present first test results and an outlook on further potential applications.