Skip to content

samlobe/FastHydroMap

Repository files navigation

FastHydroMap

PyPI version Python versions DOI

FastHydroMap predicts per-residue dewetting free energies (Fdewet) from protein structures and trajectories.

FastHydroMap overview

Quick Start

Use a fresh Python environment. Python 3.11 to 3.14 are supported.

pip install fasthydromap
fasthydromap install-torch
fasthydromap predict your_structure.pdb -o outputs/your_structure_fdewet

fasthydromap install-torch defaults to the CPU build, which is usually the right choice for current FastHydroMap workloads because SASA preprocessing dominates runtime.

Advanced installation options, Docker usage, GPU Torch variants, and release workflows are documented in INSTALL.md and PYPI_RELEASE.md.

Inputs

FastHydroMap supports:

  • Single protein structures in PDB format
  • Protein trajectories in DCD or XTC format together with a matching topology PDB

Typical usage:

# Single structure
fasthydromap predict examples/1A1U.pdb -o outputs/1A1U_fdewet

# Trajectory
fasthydromap predict-trajectory examples/proteinG.pdb examples/proteinG_short.dcd -o outputs/proteinG_fdewet

Outputs

For a single structure, FastHydroMap writes:

  • *.csv: one row per residue with Fdewet; with --parts, intrinsic and context columns are included
  • *.pdb: a copy of the input structure with predicted Fdewet written to B-factors

For a trajectory, FastHydroMap writes wide CSV files containing one row per frame and one column per residue. Use --parts to also write intrinsic, context, and per-frame summary CSVs.

Model Scope

FastHydroMap was trained on structured single-chain proteins and the 20 canonical amino-acid chemistries. Predictions for PTMs and other non-canonical chemistries should be treated cautiously.

Visualization

FastHydroMap writes Fdewet values to the B-factor column of output PDBs, so you can color structures directly in molecular viewers.

ChimeraX:

color bfactor range 4,6.5 palette ^lipophilicity

PyMOL:

spectrum b, red_white_blue, minimum=4, maximum=6.5

For dynamic hydrophobicity visualization in a MD trajectory, see the teaching-oriented example script scripts/chimerax_fdewet_trajectory_example.py with a ChimeraX implementation you can adjust.

Citation

If you use FastHydroMap in your research, please cite the software release:

DOI

Lobo, S. FastHydroMap (Version 0.1.2) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.19744336

When the manuscript becomes available, please cite that as well.

Acknowledgements

Shell Lab and Shea Group

About

Rapid dewetting free energy prediction using a GNN

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors