/torsional-strain

Primary LanguagePythonApache License 2.0Apache-2.0

Calculation of Torsional Strain Energy

Setting up the Environment

This step will get you setup with an evironment within which you can run the code in this repository using Orion.

  1. Get access to OpenEye Orion and Magpie. If you are a licensed Orion user and don't have access, please contact OpenEye Support.
  2. Activate a conda distribution (typically something like source $CONDA/activate)
  3. Create an environment conda env create -v -f environment.yaml -n torsion
  4. Activate the environment conda activate torsion
  5. Install OpenEye packages
    a. Insert your magpie token into the first line of requirements_dev.txt:
    --extra-index-url https://token:YOUR_TOKEN@magpie.eyesopen.com/simple/
    b. Install packages with pip install -r requirements_dev.txt
  6. Configure the orion CLI
    ocli --profile=default config profile

You should now have an environment capable of running the orion code.

Uploading the Floe to Orion

  1. Activate the environment
  2. Run invoke clean package upload

This will add a package called torsional-strain 1.0.0 with a floe called Torsional Strain Energy to Orion. Floe Screenshot

Running the Floe

The floe requires only one input: a dataset of molecules for which we want to calculate torsional strain.
The floe generates two datasets as primary outputs:

  1. output: one output molecule per input molecule, with total molecular strain, strains at every torsion, and the torsional energy profiles of all the torsions attached as SD properties
  2. qm_confs: QM-optimized conformers for each torsional fragment for each molecule

The floe also generates datasets for torsional fragments, pre-QM torsional conformers, and conformers that fail various QM stages.