Codes

• Expectation Maximized Molecular Dynamics (EMMD)

EMMD is a computationally efficient algorithm based on statistical inference for fast estimation of key features in the two-dimensional FEL. This method avoids direct sampling of high free energy states. Rather, the transition states connecting metastable regions of comparable free energies are estimated using Bayesian likelihood maximization. [code, publication]

• Protein Collective Variable & Membrane Surface Prediction

Here we present an autoencoder and graph attention neural network architecture to get the machine learned protein collective variable, and membrane surface prediction given a protein structure. [code, publication]

• Geometry Optimized Local Direction (GOLD) search

We have developed this algorithm to find the minimum free energy path (MFEP) existing on a free energy landscape (FEL). This is coupled with the EMMD-generated FEL. [code, publication]

• Protein Structural Persistence

The algorithm computes the Structural Persistence of a protein with respect to a reference structure of itself. The repository contains a TCL script for computing the structural persistence of the protein and of the selected region of the protein. [code, publication]