Dr. Sherman works closely with researchers using Schrödinger software
for molecular modeling and drug design projects. In this newsletter,
Dr. Sherman talks about R-group analysis, structural interaction
fingerprints, and more. All of the scripts discussed here may be
downloaded free of charge from the Schrödinger Script Center, or by
using 'Update...' from the Scripts menu in Maestro.
Q: I have a set of congeneric molecules and would like to analyze the R groups. How can I do this with the Schrödinger tools?
A: We have a new tool to do R-group analysis (r_group_analysis.py). Using your input set of molecules, the script will determine the core structure either from CombiGlide results, a user-submitted SMARTS pattern, or an automated Canvas Maximum Common Substructure (MCS) search. You can then view the core, view a list of fragments at each position in a 2D viewer, and export files containing these fragments.
Q: I have been hearing a lot of good things about structural interaction fingerprints (SIFt or IFPs). How can I analyze my docking results using this technique?
A: You can use the Interaction Fingerprints script (interaction_fingerprints.py), which implements a variant of the method described in the paper "Structural Interaction Fingerprint (SIFt): A Novel Method for Analyzing Three-Dimensional Protein-Ligand Binding Interactions" Deng et al, J. Med Chem (2004), 47, 337. The script's graphical interface allows you to compute interaction fingerprints between a receptor and ligands. The resulting fingerprints can be visualized in an interactive interaction matrix, sorted by similarity, and clustered. Clicking cells in the clustering matrix shows 2D structures for each pair of ligands, and results can be incorporated into the Project Table.
Q: I just read the paper "Energetic analysis of fragment docking and application to structure-based pharmacophore hypothesis generation" by Loving et al. in J. Comp. Aid. Mol. Des., which describes the use of an automated script used to generate structure-based pharmacophores from fragment docking. Is this script available for download?
A: Yes, we recently posted the script to do this (xpdes_score_phase_hypothesis.py). It creates a Phase hypothesis from one or more ligand poses by mapping the Glide XP descriptor energies onto each pharmacophore feature and then selecting the top scoring features. The input can be a single pose (e.g., from a crystal structure) or multiple poses (i.e., from fragment docking). For fragment docking input, the poses can be clustered by volume overlap to obtain spatially diverse poses. Receptor-based excluded volumes can optionally be included. A Glide poseviewer file and an XP descriptor file are required as input.
Q: I'm interested in running some semiempirical quantum mechanics calculations using Jaguar. How can I get started with this?
A: We have a new graphical interface to Jaguar's MOPAC implementation (mopac_gui.py) that can be run from either inside or outside of Maestro. The interface allows you to run RM1, AM1, PM3, MNDO, or MNDO/d calculations either with or without geometry optimization. You can also use the interface for non-default calculations by specifying additional keywords.
Q: I am looking for a more complete plotting tool. For example, I would like to plot four different properties simultaneously, with the 3rd and 4th dimension represented by color and size. Is there a way to do this?
A: We recently developed a new plotting tool (projectplot.py) that is available on the Script Center. In addition to scatter plots with four variables, it can also create 1D or 2D histograms, and can manage different plots using an intuitive plot manager.