Tutorial 6 — Protein–RNA complex

Goal: build and simulate a multi-component system — a protein chain plus an RNA chain — the minimal ingredients of a ribonucleoprotein (RNP) condensate. You’ll learn how COSMO handles mixed biomolecule types and why the model choice is constrained when nucleic acids are involved.

Time: a few seconds on a CPU.

Files in this folder

File

Role

protein.pdb

The protein component (all-atom).

rna.pdb

The RNA component (all-atom).

combine.py

Merges the two into one PDB (RNA → chain A, protein → chain B).

combined.pdb

The pre-merged input used by md.ini (ready to run).

md.ini

Configuration — note model = hps_kr.

run_simulation.py

Thin runner wrapper.

Background concepts

  • One bead per residue, two chemistries. COSMO coarse-grains a protein at the Cα and a nucleotide at the phosphate (P). A mixed system is just a topology that contains both kinds of bead, with the appropriate per-type parameters.

  • Model choice is constrained. Only models that carry nucleic-acid parameters can score protein–RNA and RNA–RNA interactions. In COSMO that is hps_kr (Kapcha–Rossy scale, with RNA and phospho-protein parameters), so this tutorial uses model = hps_kr. Using a protein-only model (hps_urry, hps_ss) here would lack RNA parameters. (mpipi also has RNA parameters; see the model table in Tutorial 2.)

  • Everything is in one PDB. COSMO reads a single pdb_file. Multi-component and multi-chain systems are assembled upstream into one PDB whose chains are distinct — exactly what combine.py does here.

Step-by-step

1. (Optional) Rebuild the combined input

combined.pdb is already provided, but you can regenerate it to see how a multi-component input is assembled:

python combine.py        # reads rna.pdb + protein.pdb -> writes combined.pdb

It puts the RNA on chain A and the protein on chain B and renumbers residues so the merged file is clean.

2. Run the complex

python run_simulation.py -f md.ini

Outputs go to traj/complex.* (see Tutorial 1/4 for the full file list). The build log reports two chains and uses the hps_kr force field, which scores the protein–protein, protein–RNA, and RNA–RNA contacts together.

3. Watch them interact

vmd traj/complex.psf traj/complex.dcd

With attractive protein–RNA interactions the two chains associate rather than drifting apart — the elementary event behind RNP condensate formation.

Scaling up to a condensate

This two-chain system is the building block. To study an actual RNP condensate you would:

  1. replicate many protein and RNA chains into one PDB (the upstream assembly step, generalizing combine.py),

  2. put them in a periodic box (Tutorial 3) and use the slab protocol (Tutorial 5) to measure coexisting dense/dilute phases,

  3. scan the protein:RNA stoichiometry and temperature — RNA can both promote and (in excess) dissolve condensates, the well-known reentrant behavior.

Try next

  • Swap model = mpipi (also nucleic-acid capable) and compare the strength of association.

  • Put the complex in a box (pbc = yes, box_dimension = …) and run NVT, as a bridge toward the slab setup in Tutorial 5.