Skip to content

Latest commit

 

History

History
63 lines (42 loc) · 2.29 KB

README.rst

File metadata and controls

63 lines (42 loc) · 2.29 KB
Raw

Nearl

Python 3.9 License Documentation Status

Nearl is an open-source machine learning framework designed to mine protein dynamics information from molecular dynamics (MD) trajectories. The current release focuses on a 3D voxel-based representation for use in 3D convolutional neural networks (3D-CNNs).

Key Features

  • Automated pipeline for featurizing MD trajectories
  • Flexible definition of features and trajectory supplier
  • User-friendly API for customizing featurization workflow
  • Support for true 3D dynamic and static features
  • Pre-built 3D-CNN models for machine learning training and development

Documentation

The installation guide and tutorials are available on ReadTheDocs.

Quick Start

Below is a simple example demonstrating how to featurize an example trajectory set with Nearl. The resulting feature, with dimensions of 32×32×32 and a grid resolution of 0.5, represents the mass distribution of substructures near all ARG residues. A small example dataset (approximately 26MB) will be downloaded to the directory /tmp/test. For more details, please refer to the documentation.

import nearl
import nearl.featurizer, nearl.features, nearl.io

loader = nearl.io.TrajectoryLoader(nearl.get_example_data("/tmp/test")["MINI_TRAJSET"])
feat = nearl.featurizer.Featurizer({"dimensions": 32, "lengths":16, "time_window":10})
feat.register_feature(nearl.features.Mass(outkey='mass', outfile="/tmp/test.h5", sigma=1.5, cutoff=5.0))
feat.register_focus([":ARG"], "mask")
feat.register_trajloader(loader)
feat.main_loop()

License

This project is licensed under the MIT LICENSE.