If you are looking for instructions for the COVID-19 Project, please go to the Using Modeller page.
If you are looking for specific instructions for another project, such as CFTR, try looking under Projects.
In order to run Modeller, you first need a template, which consists of an initial structure file, an alignment file, and a Python script. You will need Modeller, PyMOL, and Python to complete this protocol. See the Molecular Modeling page for download links.
The initial structure file can be generated using PyMOL. To do this, download the relevant system from RCSB (in .pdb format). Open the system in PyMOL and delete the atoms not relevant to the simulation. You will need to delete all water molecules and ions present in the system, and you may also want to remove glycosides (usually displayed as NAG). You can see the sequence, including all components of the structure, by selecting the S button in the bottom-right corner of the screen. Once finished, go to File > Export molecule, change the dropdown to all, and save the protein as In.pdb.
The alignment file will need to be generated manually. Using an example alignment file and the (relevant portion of the) FASTA sequence for the protein of interest from RCSB, write an alignment file and title it model.ali. Note the following:
You need two copies of the full sequence of the entire system: one under In, and one under Out.
If you are simulating a system with more than one chain, you need to insert a forward slash (/) between the last amino acid of one chain and the first amino acid of the next, and you need to append an asterisk (*) after the end of the last chain. The FASTA sequence will list the chains in different paragraphs; this will not work for the alignment file.
You will need to edit the residue ID and chain ID of the very first and very last amino acids. In the example file, the protein started with chain A, residue ID 21; then ended with chain B, residue ID 518. Make sure to edit these numbers to accurately reflect your own system.
You should not change any other part of the alignment file unless you have thoroughly researched its function. The Modeller website has some useful information.
You will need to download the example Python script. This Python script can stay mostly the same between systems, with one exception: you will need to edit the disulfide bond locations. This can be found on RCSB. Keep the following in mind:
You need one patch for each disulfide bond in your system. If you are adding patches to increase the number of disulfides, be sure to match the indents. If you need fewer (or no) disulfide bonds, simply remove all of the lines beginning with self.patch (and the short line beginning with self.residues just beneath each).
You need to use the RCSB website, as well as a close reading of the sections of your protein of interest you are retaining, to determine which disulfide patches to add. Modeller will automatically delete disulfide bonds, so you must list them specifically.
When adding disulfide patches, Modeller uses the residue numbers for the output file, which are generally different from the residue numbers of the input file. Modeller writes chain A as starting with residue ID 1 and shifts all other residues down to match. It then starts chain B with the residue number following the last new number in chain A. Chain C starts with the residue number following the last new number in chain B, and so on. Thus, you need to determine what the output residue numbers will be, and use those when writing disulfide patches.
It is generally helpful to test Modeller templates with a wild type (unmutated) run. Then, you can compress a work directory and add it to the Drive folder for your project.
You will need Modeller to complete this protocol. See the Molecular Modeling page for a download link.
Make a copy of the template files you previously prepared. You need only one template for each input base. You'll be working with the copy for the rest of this protocol.
Open the alignment file. Edit the out sequence (the bottom paragraph) using one-letter amino acid codes according to the point mutation of interest. Save and close the alignment file.
Open a Terminal window in the directory (or cd to the directory). Run mod10.1 model.py. If you are using a different version of Modeller, change 10.1 to the correct version on your machine.
Open model.log and scroll to the bottom. Unless you changed the number of output systems to produce, there should be five systems listed. For simulations, select the one with the lowest DOPE score. DOPE scores are negative, so the lowest score is the score with the highest absolute value (e.g. -25000 is preferred over -24800).
Copy the .pdb file that matches the selection you made from the log. This is the structure file you will use for your mutated system.
Note: When working with only one computer, if you need more than one structure file for a single mutation set, it is appropriate to select the best two or three Modeller outputs from a single run. Running Modeller more than once on the same computer will yield identical results, so running Modeller with the same settings three times and selecting the best from each run will yield three identical systems, which is not useful for simulations.