Creating coarse-grained models for the MARTINI force field using Martinize2

The Martinize2 SAMSON Extension uses Martinize2 and Vermouth ¹ and makes it possible to create coarse-grained (CG) structures and topologies (e.g., for GROMACS) from an atomistic structures for the MARTINI force field ².

A CG structure is a simplified representation of an all-atom structure achieved by grouping multiple atoms into CG beads (e.g., all amino acid backbone atoms are represented by a single bead). This reduces the complexity of a modeled system allowing to speed up simulations by several times.

Creating CG model from atomic structure

Let's see how to create CG structures and topologies on an example of the Ubiquitin protein (1UBQ).

1. Load in SAMSON the atomic system(s) for which you want to create CG model(s) (structures and topologies).

   Tip

   To download a PDB structure from RCSB PDB, go to Home > Fetch and provide the PDB code. You can choose any format: PDB, mmCIF/PDBX, mmTF.

2. Prepare the protein system(s) using Home menu > Prepare. Choose to remove alternate locations, water, ions, ligands, and other molecules used for crystallization.

   

3. Open Martinize2 (Home > Apps or via Find everything...)

4. Select the system in the document.

   

   Note

   Each structural model (the top-level structural node as shown in the image above) will be treated as a separate input model by Martinize2, i.e. separate coarse-grained models will be generated.

5. Click Set in Martinize2.

   

6. Check the Martinize2 options, e.g.:

   • The force field to use: martini3001.
   • The position restraints (they will be added in the generated include topology files (itp)):
     • backbone - create position restraints for backbone beads in proteins,
     • all - create position restraints for all beads,
     • none - no position restraints will be generated in the output itp file.
   • Apply side chain corrections.
   • Set neutral termini, charged is the default.
   • etc.

   

   Tips

   Hover above options to check the tool tips.

   Scroll to see all the options.

   You can always click Restore default values to reset options to default.

7. Specify the results folder where the project folders should be created.

8. Click Create coarse-grained models. You should see the progress in the log output. The results should appear in a timestamped project subfolder in the chosen results folder. The project subfolder contains:

   • The inputs subfolder with input PDB structures, one per input structural model.
   • A log file.
   • The outputs subfolder with the generated CG models, one subfolder per input model. There you can find PDB and GRO files for the CG model, the topology files (TOP and ITP file(s)) for GROMACS, and a log file with the martinize2 command info.

   If you generated a CG model for a single structure then it will be loaded in SAMSON.

Note

SAMSON tries to detect CG systems when loading files to visualize them as connected beads.

Creating CG model for multiple replicas

Let's say you want to create a CG model for a system that contains multiple replicas of the same protein.

We will be using the same example of the Ubiquitin protein (1UBQ).

Creating replicas

You can create copies of your system using the Molecular Box Builder, some Python script, or manually in SAMSON.

If you already have your replicas created, please go to the Renumber chains and residues section.

Creating replicas manually

Let's see how to create replicas manually.

First, let's ensure that all atoms are visible - this will help us in copying and placing the replicas. Toggle the structural model visibility by unchecking and checking the check box of the structural model as shown on the image below:

This should result in the fully visible atomic structure of the structural model.

Let's now create replicas:

1. Select the chain(s) from the structural model.

   

2. Do a copy-paste: press Ctrl/Cmd + C and then Ctrl/Cmd + V, This should create a new copy of the chain in place, i.e. in the same structure and at the same position.

   

3. While having the newly copied chain selected in the document view, as shown in the image above, switch to one of the move editors via the editors toolbar on the left side of the viewport, e.g. use the global move editor (shortcut: K) and move the newly created chain.

   

   Tip

   You can specify the translational and rotational snapping in the top-left corner of the viewport. See the gif below.

4. Proceed with steps 1-3 until you have the desired number of replicas.

   Tip

   You can copy and move multiple chains at the same time.

   

   Tip

   Once you've done, you can switch back to the rectangle selection editor (shortcut: R).

Renumber chains and residues

Once you have all the replicas created, you need to ensure that residues have unique IDs and chains have unique IDs and names. Otherwise, there might be issues when creating topologies.

1. Renumber residue IDs: right-click on the structural model and, in the context menu, go to Structural model > Renumber residues and structural groups:

   

   In the pop-up dialog, leave the default value to 1 and click OK.

   

2. Renumber chain IDs: right-click on the structural model and, in the context menu, go to Structural model > Renumber chain IDs:

   

   In the pop-up dialog, leave the default value to 0 and click OK.

   

3. Rename chains to ensure unique names. You can rename chains directly in the document via F2 or right-click and Rename:

   

   or via the Inspector

   

Tip

Save the system into a file to store your work.

Next

Once you have your replicas created and ensure unique numbering and naming, you can then proceed with the same steps as in the Creating CG model from atomic structure section.

See also

For the documentation on Martinize2 and Vermouth see https://vermouth-martinize.readthedocs.io.

References

---

1. Peter C. Kroon, Fabian Grunewald, Jonathan Barnoud, Marco van Tilburg, Paulo C. T. Souza, Tsjerk A. Wassenaar, Siewert-Jan Marrink. Martinize2 and Vermouth: Unified Framework for Topology Generation. https://arxiv.org/abs/2212.01191 ↩

2. P.C.T. Souza, R. Alessandri, J. Barnoud, S. Thallmair, I. Faustino, F. Grünewald, et al., Martini 3: a general purpose force field for coarse-grained molecular dynamics, Nat. Methods. 18 (2021) 382–388. DOI: https://doi.org/10.1038/s41592-021-01098-3 ↩