Many Copies, One Topology: Prepare Protein Replicas Before Coarse-Graining

Building a system with several copies of the same protein can look straightforward: duplicate a chain, move the copy, and repeat. But before generating coarse-grained models and GROMACS topologies, there is an important cleanup step. Replicas should have unique residue IDs, chain IDs, and chain names. If they do not, topology generation may run into issues.

This is especially easy to overlook because the copies can look perfectly separate in the viewport while still carrying duplicated identifiers in the structural model. A little organization before running Martinize2 can make the conversion workflow easier to manage and its outputs clearer.

First decide what counts as a separate input

Martinize2 treats each top-level structural model in the SAMSON document as a separate input model. In practice, that means the way you organize the document matters: if you want several replicas handled together as one input, place them in the same structural model. Separate top-level structural models will produce separate coarse-grained models.

Start with a prepared atomistic structure. The tutorial uses ubiquitin (PDB 1UBQ) as an example. If you are downloading a structure in SAMSON, use Home > Fetch and provide its PDB code. Then use Home > Prepare to clean the system as appropriate—for example, by removing alternate locations, water, ions, ligands, and other crystallization molecules.

Make and place the replicas

You can create replicas with the Molecular Box Builder, a Python script, or manually. For the manual route, make sure all atoms are visible before copying; otherwise, parts of the structure may be harder to see and select.

1. Select the chain or chains you want to duplicate.
2. Copy and paste with Ctrl+C and Ctrl+V on Windows or Linux, or Cmd+C and Cmd+V on macOS. The new chain is initially placed at the same position as the original.
3. With the copied chain selected, choose a move editor and move it to the desired position. The global move editor can be opened with the K shortcut. Translational and rotational snapping can help make placement more controlled.
4. Repeat until you have the desired number of copies. You can also copy and move multiple chains at once.

Give every replica unique identifiers

Once the copies are in place, update their identifiers. This is the step that turns a visually duplicated structure into a clearly organized multi-replica input.

1. Renumber residue IDs. Right-click the structural model and choose Structural model > Renumber residues and structural groups. In the dialog, leave the default starting value of 1 and click OK.
2. Renumber chain IDs. Right-click the structural model and choose Structural model > Renumber chain IDs. Leave the default starting value of 0 and click OK.
3. Rename the chains. Ensure that each chain has a unique name. You can rename a chain in the document with F2 or by right-clicking and choosing Rename; the Inspector is another option.

These steps address different identifiers, so do not treat renumbering residues as a substitute for renumbering chains or giving chains unique names. Save the system once the replicas are organized, so you can return to the prepared input if needed.

Then generate the coarse-grained model

Open the Martinize2 extension from Home > Apps or through Find everything…. Select the structural model in the document and click Set in Martinize2. Review the options for your intended workflow; for example, choose the force field and decide whether position restraints should be generated for backbone beads, all beads, or none.

Choose a results folder and click Create coarse-grained models. The timestamped project folder contains the inputs, a log, and outputs organized by input model. Output files include coarse-grained structures in PDB and GRO formats and GROMACS topology files (.top and .itp). If the input was a single structural model, its generated coarse-grained model is loaded in SAMSON.

The practical takeaway: arrange replicas in the structural model you intend to process, then make residue numbering, chain IDs, and chain names unique before generating topology files. For the complete walkthrough and additional options, see the Martinize2 documentation.

Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.