Skip to content

README.md

Latest commit

 

History

History
executable file
·
86 lines (64 loc) · 3.45 KB

README.md

File metadata and controls

executable file
·
86 lines (64 loc) · 3.45 KB

Proteus: Automated Polymer Nanoprecipitation Pipeline

Proteus is a computational research pipeline designed to simulate the self-assembly and nanoprecipitation of polymer chains. By automating the transition from chemical text (SMILES) to physical simulation (Molecular Dynamics), it allows researchers to rapidly screen and analyze polymer folding behaviors.

Features

  • Text-to-Structure: Instantly converts SMILES strings into 3D molecular geometries with explicit hydrogens.
  • Automated Topology: Generates LAMMPS-compliant data files with generic force field parameters (Lennard-Jones).
  • Physics Engine: Runs implicit solvent simulations using Langevin dynamics with optimized viscosity for realistic molecular drifting.
  • Built-in Analytics: Automatically parses simulation logs to calculate the final Radius of Gyration ($R_g$) and generates stability plots.
  • Automated Visualization: Uses Ovito to render high-quality, color-coded GIF animations of the simulation.

Installation

Proteus uses uv for high-performance dependency management and virtual environment creation.

1. Install uv

If you don't have uv installed, you can install it using the following command (Linux/macOS):

curl -LsSf https://astral.sh/uv/install.sh | sh

For other platforms or methods, refer to the uv documentation.

2. Clone the repository

git clone <repository-url>
cd proteus

3. Set up the environment with uv

make setup

4. Activate the environment

source venv/bin/activate

Usage

The easiest way to run the pipeline is using the run.sh helper script.

Using the Helper Script

chmod +x run.sh
./run.sh "<SMILES>" "<NAME>" [STEPS]
  • Example (3-unit PEO): ./run.sh "CCOCCOCCO" "PEO_3" 10000
  • Example (Multi-chain): ./run.sh "CCO.CCO.CCO" "Triple_Chain" 10000

Arguments

  • SMILES: The chemical structure. Use a dot . to separate independent molecules.
  • NAME: The folder name for your results in output/.
  • STEPS: (Optional) Simulation duration. Default is 10,000.

Output

Results are saved in output/<NAME>/:

  • animation.gif: High-quality video of the simulation (color-coded by molecule).
  • polymer.data: LAMMPS topology file (3D structure).
  • simulation.in: The generated LAMMPS input script.
  • simulation.log: Raw simulation data (energies, temperatures).
  • stability.png: Equilibrium graph showing Temperature and Potential Energy over time.
  • trajectory.dump: Raw atom positions (viewable in external tools like VMD/Ovito).

Architecture

  1. Simulation Core (src/):
    • Topology Architect (topology.py): SMILES $\to$ 3D Topology via RDKit & UFF.
    • Simulation Engine (simulation.py): Langevin dynamics (300K) with tuned viscosity.
    • Analytics (analysis.py): Log parsing and $R_g$ calculation.
    • Visualization (visualization.py): Headless rendering of trajectories via Ovito.
    • Reporting (report.py): PDF report generation.
    • HTS (hts.py): High-Throughput Screening automation.

Simulation Physics

The simulation uses a Generic Hydrophobic Interaction model:

  • Force Field: Lennard-Jones (epsilon=0.105, sigma=2.5).
  • Solvent: Implicit solvent via Langevin thermostat.
  • Viscosity: Tuned damping for realistic drifting rather than high-frequency vibrations.

CONCLUDING THE PROJECT FOR NOW