Project ID: plumID:23.009
Name: Deep Learning Collective Variables from Transition Path Ensemble
Archive: https://github.com/dhimanray/TPI_deepTDA/archive/main.zip (browse)
Category: methods
Keywords: TPI-Deep-TDA, Deep-TDA, Transition Path, OPES, OPES Flooding, Machine Learning, Protein folding, Ligand binding
PLUMED version: 2.9
Contributor: Dhiman Ray
Submitted on: 06 Mar 2023
Publication: D. Ray, E. Trizio, M. Parrinello, Deep learning collective variables from transition path ensemble. The Journal of Chemical Physics. 158 (2023)
PLUMED input files
Last tested: 16 Jan 2025, 20:26:03
Project description and instructions
Introducing the Transition Path Informed Deep Targeted Discriminant Analysis (TPI-Deep-TDA) method for improving deep neural network-based collective variables by utilizing information about the transition pathways. It is tested on Muller potential, folding/unfolding of chignolin, and ligand binding in the OAMe-G2 complex studied in SAMPL5 challenge. The MD engine used is GROMACS 2021. The Deep-TDA and TPI-Deep-TDA CV were trained using Pytorch 1.8.2 and mlcvs package.
Submission history
[v1] 06 Mar 2023: original submission
Badge
Click on the image below and get the code to add the badge to your website!
