Molecular Modeling

Molecular modeling is the use of computational tools to perform physical calculations, with the goal of determining the kinematic and kinetic characteristics of a molecule. In short, it does physics to a molecule so you don't have to.

Molecular modeling, along with molecular dynamics simulation, allows us to quantitatively pre-test key features of proteins without using molecular biology techniques. This saves time and valuable resources, while also allowing us to focus those resources on fewer mutations later.

Stage 1: Homology Modeling

Homology modeling involves using the known structure of an existing protein to generate digital models of closely related proteins. We typically use homology modeling to generate structure files depicting a protein with a few point mutations. We do this using Modeller.

Stage 2: Input Generation

In order to simulate a protein, it must be dissolved or inserted into a solution of physiologically relevant substances. If your protein is on or within a cell membrane, for instance, the simulation system needs to include a membrane, a cytoplasm, and an extracellular matrix. Input generators (such as CHARMM-GUI) allow you to insert a protein into a realistic simulation system, usually in the shape of a box, that allows you to simulate a protein's interaction with its physiological environment.

Stage 3: Molecular Dynamics Simulation

Molecular dynamics simulation (referred to elsewhere on this Wiki as MD Simulation) is the process of calculating the forces between atoms in a simulation system, calculating the resulting movement of those atoms, and then repeating for a specified amount of iterations. Because so many calculations need to be performed, high-grade hardware is required to run a simulator. We often use Bridges-2 at the Pittsburgh Supercomputing Center, running GROMACS, to simulate systems of interest.

Stage 4: File Conversion and Analysis

Because we use GROMACS for simulation, but NAMD for analysis, we need to convert files. We do this using VMD, which can read and write in several file formats as well as calculate important data for analysis, including root-mean-square deviation data. Once the files are converted, we can use NAMD to calculate interaction energies and more detailed RMSD data for analysis.

All of the following software is being continuously updated and changed, and we may discontinue use of certain software tools over time. It is important to know what you are downloading and why before defaulting to the Wiki's recommendation. The tools listed below are in use in our lab as of July 2021.

• Homology modeling: Modeller 10.1, Python 3.9

• Input generation: CHARMM-GUI

• MD Simulation: GROMACS 2020.4 (on Bridges-2), GROMACS-2021.2 (on personal computers)

• Conversion and Analysis: VMD 1.9.4, NAMD 2.14, Python 3.9

Some protocols have already been written for existing projects. These protocols are specifically designed for their respective projects, and may need to be altered before they will be applicable to your project's needs.

CFTR Project

1. Homology modeling: Alignment File Creation, Modeller, Mock Modeling, Modeling Training

2. Input generation: CHARMM-GUI, DESMOND

3. MD simulation: XSEDE Account, Benchmarking, Equilibration, Gromacs, SA Simulation

4. File conversion and analysis: Residue Pairs, RMSD Analysis

COVID-19 Project

1. Homology modeling: Using Modeller

2. Input generation: Using CHARMM-GUI

3. MD simulation: Installing GROMACS, Running MD Simulations

4. File conversion and analysis: Using VMD, Using Tool.py, Using NAMD Energy