Dr. Sherman works closely with researchers using Schrödinger software for molecular modeling and drug design projects. Here, Woody describes some common hurdles that can be easily overcome with the aid of Schrödinger scripts.
Schrödinger recently introduced Python scripting as a method for our customers to automate complex tasks, build workflows, and implement new functionality using our software. Since its introduction, a significant number of Python scripts have been developed, largely inspired by users of Schrödinger software.
Schrödinger makes many useful scripts available at our Script Center. These scripts can be used out-of-the-box without any modification, or can be customized to suit individual needs. However, because of the large number of scripts available for download, many of our customers may not be aware of existing scripts that are applicable toward their research.
Below, I've collected questions that you may have encountered while working on a particular project. These common research problems are easily solved using tools that can be downloaded from our Script Center.
Q: After docking some ligands to my target, I realized that I had not rotated the protein into a common frame of reference before docking. The ligands are docked correctly, but they cannot be compared with results from calculations on similar targets with different frames of reference. How can I get everything — the protein and all docked ligands — into a new frame of reference?
A: We have a script called rotate_all.py to do exactly this. It first aligns the mobile protein onto a reference protein by running the Protein Structure Alignment tool (structalign) and then applies the resulting rotation matrix to your docked ligands. The resulting Pose Viewer file will now have everything in the correct frame of reference.
Q: After preparing my protein using a different program and then importing it into Maestro, I found that certain PDB atom names were incorrect. For example, the CD1 and CD2 atoms on leucine residues are swapped. How do I correct these types of PDB atom naming problems?
A: The script pdbname.py uses MacroModel substructure definitions to set PDB atom and residue names for a structure. Because it uses substructure definitions, connectivity is the only property used to assign atom names, and thus the script is able to correct PDB atom names independently of the original names.
Q: I've docked my ligands and would now like to filter them based on various types of interactions with the protein. For example, in some cases I want to ensure that an aromatic ring is in a certain pocket of the protein. In other cases, I want to ensure that there are H-bonds with certain residues because I did not run Glide with H-bond constraints.
A: We have a script called pose_filter.py that can filter based on different interaction types between the ligand and the protein. The protein atoms of interest can be specified with an ASL expression and different ligand functionalities (H-bond donor/acceptor, aromatic rings, or hydrophobic groups) can be required to interact with the specified protein atoms.
Q: My group is starting a new project with a target for which there are many crystal structures. As a first step, I would like to dock all PDB ligands into each of the crystal structures. Is there a way to automate this process?
A: Yes, XGlide.pl, which is designed to help researchers run cross-docking and docking accuracy experiments, automates the process of running Glide on multiple receptors. The script docks all ligands into all receptors, and then produces a report summarizing the GlideScores and ligand RMSDs.
Q: I'd like to perform mathematical operations on existing properties in the project table to create new properties. Is there an easy way to do this?
A: We have a script called projectcalc.py that provides a graphical calculator for Maestro's Project Table. A variety of standard functions are available, and complex formulae can be created. The script allows new properties to be created based on the operations that you perform.
Q: I often find that crystal structures have the terminal chi angle of asparagine, glutamine, and histidine incorrectly assigned (i.e., flipped by 180 degrees). I usually manipulate this by hand, but is there a faster way to make these corrections? It would also be helpful to automatically predict the correct tautomer for histidine residues.
A: Our Protein Assignment program, called protassign.py, assigns optimal positions for hydroxyl and thiol protons and assigns histidine protonation and tautomeric states. It also attempts to fix incorrectly positioned side chains of asparagine, glutamine, and histidine (via 180 degrees terminal-chi flips). For additional information, please see our whitepaper.
Q: I am working on a target for which there are several related targets that we would like to avoid interactions. I want to get a quick comparison of the sequences of the active sites, both to see the differences at each position and to get an overall metric of the active site identity.
A: The script active_site_identity.py can do this. It takes a set of aligned proteins and computes the residue identity and homology between the active sites of each structure. The output is a .csv file that reports on active site residues for each protein, and a .sim file with a matrix of the fraction of identical residues between each protein.