menger-curvature 1.1.0

This project aims to provide a simple MDAkit for JIT accelerated Menger curvature calculation

Menger_Curvature

Latest release
Status
Community

This project aims to provide a simple MDAkit for JIT accelerated Menger curvature calculation. The idea is to associate a value of curvature to as many residues as possible in a polymer. If one has access to several conformations , the average value of the curvature (LC) and its standard deviation (LF) are valuable information to characterize the local dynamics of the backbone.

Range of proteic menger curvature (PMC) values and their associated structural elements. Backbone representations are extracted from the single chain tubulin simulation. Backbone is represented in licorice, Cαs involved in the PMC calculations are in black Van de Waals.

Cite

Please consider citing the BiorXiv preprint at BiorXiv.

Installation

User

With pypi project

The easiest way to install Menger_Curvature is through pip:

pip install menger-curvature

This will install the latest stable release from PyPI.

Developper

Clone the repository and enter it:

git clone https://github.com/EtienneReboul/menger_curvature.git
cd menger_curvature

To build Menger_Curvature from source, we highly recommend using virtual environments. If possible, we strongly recommend that you use Anaconda as your package manager. Below we provide instructions both for conda and for pip.

With conda

Ensure that you have conda installed.

Create a virtual environment and activate it:

conda create --name menger_curvature
conda activate menger_curvature

Install the development and documentation dependencies:

conda env update --name menger_curvature --file devtools/conda-envs/test_env.yaml
conda env update --name menger_curvature --file docs/requirements.yaml

Build this package from source:

pip install -e .

If you want to update your dependencies (which can be risky!), run:

conda update --all

And when you are finished, you can exit the virtual environment with:

conda deactivate

With pip

To install from pipy, run :

pip install menger-curvature 

To build the package from source, run:

pip install .

If you want to create a development environment, install the dependencies required for tests and docs with:

pip install ".[test,doc]"

Quick Start

We expect the calculation to take less than a minute for a trajectory of 441 alpha carbon with 20,000 frames Calculate Menger curvature for the chain A of a tubulin protein trajectory in serial mode:

import MDAnalysis as mda
from menger.analysis.mengercurvature import MengerCurvature
from menger.data import files

# replace by your own filepaths 
topology = files.TUBULIN_CHAIN_A_PDB 
trajectory = files.TUBULIN_CHAIN_A_DCD
u = mda.Universe(topology, trajectory)

# run analysis in serial mode 
menger_analyser = MengerCurvature(
    u,
    select="name CA and chainID A",
    spacing=2
    )
menger_analyser.run()

# retrieve results data
average_curvature = menger_analyser.results.local_curvatures
flexibility = menger_analyser.results.local_flexibilities
menger_curvature = menger_analyser.results.curvature_array

Calculate Menger curvature for the chain A of a tubulin protein trajectory in parallel mode:

import MDAnalysis as mda
from menger.analysis.mengercurvature import MengerCurvature
from menger.data import files

# replace by your own filepaths 
topology = files.TUBULIN_CHAIN_A_PDB 
trajectory = files.TUBULIN_CHAIN_A_DCD
u = mda.Universe(topology, trajectory)

# run analysis in parallel 
menger_analyser = MengerCurvature(
    u,
    select="name CA and chainID A",
    spacing=2,n_workers=4
    )
menger_analyser.run(backend="multiprocessing", n_workers=4)
average_curvature = menger_analyser.results.local_curvatures
flexibility = menger_analyser.results.local_flexibilities
menger_curvature = menger_analyser.results.curvature_array

Tutorial

We provide a more comprehensive tutorial in a jupyter notebook

Code of conduct

Menger_Curvature is bound by a Code of Conduct.

Copyright

The Menger_Curvature source code is hosted at https://github.com/EtienneReboul/menger_curvature and is available under the GNU General Public License, version 2 (see the file LICENSE).

Copyright (c) 2025, LBT

Acknowledgements

Project based on the MDAnalysis Cookiecutter version 0.1. Please cite MDAnalysis when using Menger_Curvature in published work.