Molecule Similarity Based on Fingerprints with RDKit

Cheminformatics Tutorials - Herong's Tutorial Examples

∟rdkit.Chem.rdmolops - Molecule Operations

∟Molecule Similarity Based on Fingerprints with RDKit

This section provides a tutorial example on calculating molecule simularity based on fingerprints using RDKit library.

RDKit library provides its own fingerprint generation function that can be used to calculate similarity score between two molecules.

f = rdkit.Chem.rdmolops.RDKFingerprint(m) - Returns the fingerprint using RDKit topological algorithm from a molecule.

s = rdkit.DataStructs.FingerprintSimilarity(f1 f2) - Returns the calculated similarity score between two fingerprints.

Here is a Jupyter Notebook example of calculating molecule similarity with RDKit library.

from rdkit import Chem
from rdkit import DataStructs
ms = [Chem.MolFromSmiles('CCCO'), Chem.MolFromSmiles('CCOC')]
fs = [Chem.RDKFingerprint(x) for x in ms]
s = DataStructs.FingerprintSimilarity(fs[0], fs[1])
i = Chem.Draw.MolsToGridImage(ms, subImgSize=(120,120))
display(i, 'Similarity: '+str(s))