Skip to content

Code for "A small molecule stabilises the disordered native state of the Alzheimer’s Aβ peptide"

License

Notifications You must be signed in to change notification settings

vendruscolo-lab/ab42-g5-ensemble

Repository files navigation

A small molecule stabilises the disordered native state of the Alzheimer’s Aβ peptide

This repository contains the full code to reproduce our results of kinetic ensembles of amyloid-β 42 with urea and the small molecule 10074-G5. See our previous work for details on the method and the original unbound ensemble.

Reproducibility information

We used the same Google compute engine instance as for the previous work. Conda environments for training (env-tf.txt) and analysis (env-analysis.txt) are provided, although we strongly recommend using a custom tensorflow install.

Dataset

The full dataset is available on zenodo.

Notebooks

See our previous work on how to handle the notebooks. They contain the following:

  • msm-vampe-hyperpar.ipynb: Hyperparameter search code, can be run with papermill and the env-tf.txt environment.
  • msm-vampe-training.ipynb: Training code, can be run with papermill and the env-tf.txt environment.
  • msm-vampe-convergence.ipynb: Convergence analysis code, can be run with papermill and the env-tf.txt environment.
  • msm-vampe-analysis.ipynb: Analysis and plotting code, can be run with papermill and the env-analysis.txt environment.
  • model.py: The neural network model code.
  • data.py: The tensorflow-independent part of model.py, including wrappers for datasets.

About

Code for "A small molecule stabilises the disordered native state of the Alzheimer’s Aβ peptide"

Resources

License

Stars

Watchers

Forks

Languages