adme-pred-py

This library supports computational drug discovery by implementing several druglikenss filters, medicinal chemistry filters, and provides an easy to use wrapping API for common cheminformatics calculations.

We want to replicate most of the functionality of SwissADME. We want the core code to be abstract enough that it can produce a extensive report on a single compound like SwissADME, or be used as a filter in a screen. (Ex. for compound in compounds if lipinski_druglikeness then ...) We will probably also want to implement some functionality from Open Source Bayesian Models.

Getting started

We use Anaconda as the base Python, so first install it from here: https://www.anaconda.com/products/individual.

Then to get up and running, create the environment:

conda create --name adme-pred-py
conda activate adme-pred-py
conda install -c conda-forge -n adme-pred-py rdkit
conda install -c conda-forge -n adme-pred-py matplotlib

Examples

Asprin is druglike because it does not violate the Lipinski Rule of 5:

chem = "O=C(C)Oc1ccccc1C(=O)O"
mol = ADME(chem)
print(mol.druglikeness_lipinski(verbose=True))

No violations found

Paromomycin is not a small molecule drug, and so druglikeness filters will screen it out:

chem = "O=S(=O)(O)O.O([C@H]3[C@H](O[C@@H]2O[C@H](CO)[C@@H](O[C@H]1O[C@@H](CN)[C@@H](O)[C@H](O)[C@H]1N)[C@H]2O)[C@@H](O)[C@H](N)C[C@@H]3N)[C@H]4O[C@@H]([C@@H](O)[C@H](O)[C@H]4N)CO"
mol = ADME(chem)
print(*mol.druglikeness_lipinski(verbose=True), sep="\n")

H Bond Donors 15>5
H Bond Acceptors 21>10
Molecular Weight 713.263680664>500

Thalidomide triggers the Brenk medicinal chemistry filter:

chem = "O=C(N1C2CCC(NC2=O)=O)C3=CC=CC=C3C1=O"
mol = ADME(chem)
print(mol.brenk())

True