Skip to content

Repo associated with paper "Computing committors via Mahalanobis diffusion maps with enhanced sampling data" by Luke Evans, Maria K. Cameron, and Pratyush Tiwary, https://arxiv.org/abs/2208.13772.

Notifications You must be signed in to change notification settings

aevans1/targetmeasure-mmap

Repository files navigation

Under Construction!

The file for the diffusion map code is src/diffusion_map.py. This uses scikitlearn extensively and is based largely on the "pydiffmap" library.

The examples available in rough .ipynb form are:

  • a 2d toy system with position-dependent diffusion, see 2D_example.ipynb.
  • a 2d molecular dynamics example, alanine dipeptide with two dihedral angles, see aladip_2var.ipynb
  • a 4d molecular dynamics example, alanine dipeptide with four dihedral angles, see aladip_4var.ipynb

There are python scripts available for the `Ksum test' on the datasets, Ksumtest_datasetname.py

In progress:

  • Step-by-step process for alanine dipeptide in vacuum data (at current state these aren't runnable by themselves, they currently serve as a reference):
  • Note: The plumed used for this was a custom plumed recompiled particularly with an adjusted DUMPPROJECTIONS (plumed documentation) in plumed/src/core/value.cpp so that it computes with the mass matrix as in equation (6) of our tm-mmap paper.
  • Cleaning up the .ipynb for readability

About

Repo associated with paper "Computing committors via Mahalanobis diffusion maps with enhanced sampling data" by Luke Evans, Maria K. Cameron, and Pratyush Tiwary, https://arxiv.org/abs/2208.13772.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published