Hello! This is a quick guide for how to use Modeller to generate .pdb files for new molecule designs.
Getting started
You'll need to have Modeller installed. To do that, you'll need to download Modeller and Python. Be sure to get the correct one for your system. If you're on Mac, I recommend following the instructions on that page to download Modeller using Homebrew. The most recent version as of July 2021 is 10.1, which defines the commands we use here. If you download a later version, be sure to change the command (to mod10.2 or whatever yours is). Once you have Modeller installed, open a command line (Terminal on Mac and Linux, Command Prompt on Windows) and change your directory to the folder with your input files.
Input files
I've made template files for you to make it easier to get started. Both the Windows and Mac/Linux templates are the same, just with different compression formats to make it easier for your system. You should only need to edit the alignment file – everything else will usually take care of itself. I recommend keeping the compressed archive somewhere handy so you can extract it as many times as you need for future decoys.
Changing the sequence
To alter the sequence, all you need to do is edit the alignment file. It opens in a basic text editor like TextEdit or Notepad, and it uses single-letter amino acid codes that you can change to create a new model. The upper sequence is the model going in, and the lower sequence is the model coming out, so only change the bottom sequence. On two-chain systems (sace2-rbd, rg-rbd), the forward slash divides the first chain from the second.
Running Modeller
Once you've edited the sequence for the appropriate system, you can run Modeller. From your terminal in the correct working directory, type mod10.1 model.py and let it run. When it's working properly, nothing appears in the terminal, but files will be generated in your working directory. If you'd like to see what it's doing, open the .log file.
Choosing a decoy
Once Modeller has finished, open the .log file and scroll to the bottom. There should be a list of five decoys with test values for each of them. Find the output file with the lowest DOPE score (these numbers are negative, so the absolute value will be higher – e.g. -24852.87500 is better than -24722.09961). Then, find that file in the working directory and open it in PyMOL.
Visual inspection
Once you have the file open in PyMOL, visually check to make sure everything looks right. Check that the intended mutations are present, that disulfide bonds are marked where they should be, and that the protein is connected and in a reasonable shape.
Editing the file
Once you've visually checked your decoy, you'll need to renumber the sequence. Unless you've added amino acids (not just changed one amino acid to another, which is 99% of what we're doing), you can follow these instructions exactly. Ask for help if you're not sure.
You'll need to install a plugin for PyMOL before you can do this. Once you've installed it, you never have to install it again. On PyMOL, go to Plugin > Plugin Manager and click Install New Plugin. In the URL box, copy the URL https://raw.githubusercontent.com/Pymol-Scripts/Pymol-script-repo/master/renumber.py and paste it into the box, then click Fetch. The plugin should be installed.
In the command box on PyMOL, type renumber chain A, 21. Then, if you have an RBD system, type renumber chain B, 336.
Go to File > Export Molecule and change the Selection dropdown to all. Click Save, then name the file according to the mutation set you used. Be sure to change the file type to .pdb, and save it.
Finishing
Post your .pdb to the Google Drive, probably in the Designed Decoys folder underneath COVID-19 Project. Make the mutation set you used part of the folder name, and keep your naming consistent across files. Add a link to it on the COVID-19 Progress page.