Contents of this repository:

For each of GB1 and AvGFP...

* md_inputs - files for running MD simulations
  - .in Amber input files for minimization, heating, and production simulations
  - Initial (pre-minimization) model .inpcrd coordinate and .prmtop topology files
* md_analysis - files relating to extracting features from MD simulations
  - .py scripts for extracting data from MD simulations ("wn" = Wernet-Nilsson H-bond energy; "pca" = pca components weights; "ks" = Kabasch-Sander backbone H-bond energy; "globecomm" = by-residue global communication scores; "sr" = Shrake-Rupley solvent-accessible surface area)
  - "match" .py scripts for creating "matched" datasets (all containing the same sequences in the same order) from raw files produced by previous set of scripts
  - "matched" .txt dataset files for each feature (same sequences in the same order) from previous scripts.
* dynamics_backbones - files relating to training the models that predict features from MD simulations
  - make_dataset.py script for converting "matched" datasets to .pkl files
  - resulting dataset .pkl files used as inputs to .py scripts to train models
  - .py scripts for training the models using Keras
  - .keras model checkpoints for trained models
* train_experimental.py script for training models on experimental data