SMOG: Structure-based Models for Biomolecules
A very brief introduction to structure-based modelingWhile originally developed for the study of protein folding, over the last 10 years we have worked to extend the applicability of structure-based models to investigate a broad range of biomolecular dynamics, including domain rearrangements in proteins, folding and ligand binding in RNA, and large-scale rearrangements in ribonucleoprotein assemblies. In its simplest form, a structure-based model defines a particular structure (normally obtained from X-ray, or NMR, methods) as the global energetic minimum. Oftentimes, this single-basin description can provide rich information about complex processes. When this simple representation is insufficient to account for the dynamics of interest, these models can be easily extended to include multiple basins of attraction, as well as describe non-specific effects (e.g. electrostatic or solvation effects). While coarse-grained variants of this model have had considerable success in expanding our understanding of biomolecular folding and function, in recent years, we have extended the Calpha model to an all-atom representation for structure-based simulations of biomolecules (complete descriptions can be found elsewhere). Additionally, with the computational simplicity of the models, they are proving to be very useful for structural modeling purposes, including atomic modeling of cryo-EM (e.g. MDfit) and SAXS data.
To increase the accessibility of these modeling techniques, we provide the smog-server webtool interface that will give you all the necessary information for performing structure-based simulations of proteins, nucleic acids, as well as many supported ligands. In addition, the webtool can handle all preprocessing steps that are necessary to prepare a structure-based forcefield for a given molecular structure. For in-depth descriptions of the underlying models, and for examples of how to use the models, we refer you to the original publications and recent review articles.
How to prepare and simulate a SMOG model for a specific biomoleculeTo run a single-basin structure-based model, the following steps should be adhered to.
Step 0: Install the Gromacs molecular simulation package (Version 4.5.X). Gromacs will be the engine of the simulations. There are several excellent general tutorials available on the Gromacs webpage. We have found Gromacs to work very well for performing these simulations, both in serial and parallel, and it provides an excellent set of post-processing tools. However, to further extend the utility of these models, we are working to implement them in NAMD. Once that option is available, it will be announced here.
Step 1: Prepare the structure-based potential input files for Gromacs. (All-atom or Calpha). Since this model is not included in any major software distribution, we have made it freely available through our web interface. All you need to do is provide a pdb structure file and specify which parameters you would like to use, and we will generate the Gromacs input files for you.
Alternate Step 1: If you are familiar with the tool, try the beta version SMOGv2.0.beta that can be run locally on your machine. While we believe you will find it useful, it is still an experimental version and we appreciate feedback.
Step 2: Run the simulation. Due to some technical differences between structure-based models and many other commonly used models, there are a few additional considerations that will ensure maximum performance of your calculations. Click here for a details on how to run a SMOG model in Gromacs. It is highly recommended that you read this information, even if you are an expert Gromacs user.
Step 3: Analyze the results. Once your simulations have finished, there are many forms of post-analysis that may be of interest. While there are many analysis tools available with the Gromacs distribution, we have also made scripts and programs available here. If you would like to contribute analysis scripts, please send us a copy and we will make them available.