What is the Pose Explorer script and how is it useful in identifying promising lead compounds?

The Pose Explorer script is a great tool for interactively exploring poses and properties of hundreds to thousands of poses. The Pose Explorer automatically computes 3D structural interaction fingerprints, 2D ligand fingerprints, and other molecular descriptors using a Glide pose viewer file as input. Poses are then clustered into separate self-organizing maps (SOMs) based on similarity of these different sets of descriptors. The SOM cells can be colored by any property of interest. The SOMs are interactive between each other and with the Maestro workspace and Project Table, so you can view the contents of the cluster map cells in the Workspace and mark poses of interest in the Project Table.

To use the Pose Explorer script, you will need a Canvas license for the SOM generation. To get started, load a pose viewer file into the Workspace. (You can use the example file titled factorXa_sp_pv.maegz located in your $SCHRODINGER/impact-v*/tutorial/glide/ directory.) Then, go to Scripts → Docking Post-Processing and open the Pose Explorer panel. In the Map Clusters tab, select the "Create Maps" button.

Once the maps have been generated, go to the View Property Projections tab to view the SOMs. The cluster map titled "Docking score Mean on Glide" is an n x n grid where each cell in the grid represents a cluster of poses that are similar to one another in Glide scoring terms. The Docking score has been graphically projected onto the grid, and a legend illustrates the color ramp. New maps can be generated with different property projections. Each cell in the cluster map titled "Docking score Mean on SIFt" represents a cluster of poses that are similar to one another in structural interaction fingerprint pattern, and each cell in the cluster map titled "Docking score Mean on 2D fingerprint" represents a cluster of poses that are similar to one another in 2D fingerprints. Each cell in the cluster map titled "Docking score Mean on Molecular Descriptors" represents a cluster of poses similar to one another in the molecular descriptors calculated by the Molecular Descriptors script available from the Schrödinger Script Center.

You will notice that the maps are interactive: for example, if you select a cell in the "Docking score Mean on Glide" map, the location of those poses in the "Docking score Mean on SIFt" map is highlighted by red circles. The size of the circle represents the number of compounds in the cell.

This script should prove useful for almost any researcher who uses Glide to select lead compounds. For example, this script allows users to identify docked compounds that make the same interaction patterns with the receptor, are diverse in 2D fingerprint space, and have specific properties of interest. For example, when identifying promising lead compounds you may be interested in a cell in the "Docking score Mean on SIFt map" that has the best Docking score, visually examine the poses in the Workspace, and see how these poses map into the "Docking score Mean on 2D fingerprint" cluster map.

Keywords: self organizing

Back to Search Results