Skip to content
Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time


GPathFinder is an extension built over GaudiMM core to allow the identification of ligand binding pathways at atomistic level.

GPathFinder logo


Different options for generate pathways

  • Unbinding routes from a known binding site
  • Binding routes to a known binding site
  • Channel analysis (given starting and final points)

Flexibility for the ligand

  • Dihedral angles
  • Pool of conformations provided by the user (e.g. through conformer generation)

Different levels of flexibility for the receptor

  • Side-chain flexibility using rotamer libraries
  • Global movements by Normal Mode Analysis sampling
  • Global movements extracted from a PCA analysis over a MD trajectory (.dcd file)
  • Pool of conformations provided by the user (e.g. snapshots from a MD trajectory)

Different options for evalute and optimize the solutions

  • Steric clashes
  • Vina scoring function
  • Smina, with the possibility to customize the scoring function and use custom atom parameters
  • Smoothness of the ligand movements

Documentation and support

Documentation source is available in docs/ subdirectory, and also compiled as HTML at this webpage.

If you need help with GPathFinder, please use the issues page of our GitHub repo. You can drop me a message at too.

Developer friendly

If the provided genes and objectives are not enough, you can always code your own ones. Check out the developer docs!


GPathFinder and GaudiMM are licensed under the Apache License, Version 2.0. Check the details in the LICENSE file.

History of versions

  • v1.3.0: 10th November 2020

The user can provide a set of conformations for the ligand or the protein or both. This allows, for example, to use a set of snapshots from a MD trajectory to define the conformational space that GPathFinder will explore.

You have a tutorial on how to use the new feature.

  • v1.2.1: 18th March 2020

Bug fix: problems with the format of some .mol2 files, that produced an error like "ValueError: invalid literal for int() with base 10:" when saving the results of GPathFinder. Thanks to Manish K. from the Nagoya University for reporting it.

  • v1.2.0: 30th October 2019

New PCA analysis functionality. If you have a trajectory file of a molecular dynamics, you can now include low energy motions calculated throughout PCA analysis. GPathFinder would include those movements in its calculations.

  • v1.1.0: 6th September 2019

New smina scoring. Possibility of using its built-in scoring functions or making a custom file with your own one. Also allows to introduce your custom atom parameters file.

  • v1.0.1: 25th July 2019

New summary.csv file in the output with score data of all the solutions.

New format for the .pdb files of the output that can be imported as a MD movie in UCSF Chimera.

New trajectory.pdb file in the output. For each solution, shows the trajectory of the ligand, taking its center as a reference.

Command to lauch the program is now gpath run instead of gaudi run to ensure compatibility with GaudiMM.

Bug fix: similarity between solutions is calculated now at each generation, to avoid some cases where repeated or very similar solutions appeared, especially when using only clashes evaluation.

  • v1.0.0: Release version. Used in the benchmark and cases study of the article.

OS Compatibility

GPathFinder is compatible with Linux and macOS.

If you find some dificulties when installing it in a concrete distribution, please use the issues page to report them.

How to cite this software

To cite this software, please refer to our article in IJMS:

Sánchez-Aparicio, J.-E.; Sciortino, G.; Herrmannsdoerfer, D.V.; Chueca, P.O.; Pedregal, J. .-G.; Maréchal, J.-D. GPathFinder: Identification of Ligand-Binding Pathways by a Multi-Objective Genetic Algorithm. Int. J. Mol. Sci. 2019, 20, 3155.


Identification of ligand binding pathways by a multi-objective genetic algorithm







No packages published