/ddGlearn

Primary LanguagePython

ddGlearn

Repository of scripts to generate (eventually) a machine-learned correction term for alchemical free energy calculations. Scripts pull ligand data from datasets/input/[target]/poses/* and datasets/input/[target]protein/[target]/*. Note that each builder script requires strict adherence to directory formatting in datasets/input/.

Prerequisites:
  • System:
    • Python 3.x, OBabel
  • Python:
    • RDKit, DeepChem, Mordred, ODDT, MDTraj
    • Tensorflow, Scikit-Optimize
To do:

  • pFP:

    • adjust to take input parameters

  • dDESCR:

    • adjust to take input parameters

  • dPLEC:

    • adjust to take input parameters
    • create PLEC-time matrix

  • General:

    • work towards first learned model
    • integrate scripts into a notebook