Preamble

This repository contains Python code and Jupyter Notebooks for reproducing the results presented in the manuscript Topological analysis reveals multiple pathways in molecular dynamics.

The same code and processed data are also available on Zenodo: https://doi.org/10.5281/zenodo.14229803

MoKiTo: Molecular Kinetics via Topology

MoKiTo (Molecular Kinetics via Topology) is a Python-based toolkit designed for analyzing and extracting topological insights from Molecular Dynamics (MD) simulations.
It enables researchers to study conformational transitions and kinetic pathways in molecular systems by generating Molecular Kinetics Maps (MKMs).

Installation

git clone https://github.com/donatiluca/MoKiTo.git
pip install -r requirements.txt

How to Use MoKiTo

The workflow depends on whether you're working with toy model systems (typically one- or two-dimensional systems driven by overdamped Langevin dynamics) or all-atom molecular systems.
The directory examples/ contains sample workflows for both use cases. Modify the scripts as needed for your specific data and analysis.

Low-Dimensional Systems

For low-dimensional systems, follow these steps:

1. Use generate_trajectories.ipynb to create the initial trajectory and the short trajectories.
2. Use isokann.ipynb to learn the $\chi$-function.
3. Use mokito.ipynb to load the trajectories and the $\chi$-function to generate the MKM and the energy landscape.

Molecular systems

For MD simulations, ensure the relevant .pdb file is placed into the input/ directory. Then follow these steps:

1. Run generate_initial_trajectory.ipynb to create the initial trajectory.
2. Use generate_short_trajectories.ipynb to sample the system's dynamics.
3. Run calculate_PWDs.ipynb to convert the MD trajectories saved as .dcd files into .pt arrays that contain the pairwise distance matrices.
4. Use isokann.ipynb to learn the $\chi$-function.
5. Use mokito.ipynb to load the pairwise distance matrices and the $\chi$-function to generate the MKM and the energy landscape.