Chapter 8: Small Molecule Cheminformatics & Chemical Databases

Chapter 8: Small Molecule Cheminformatics & Chemical Databases

Structure-based drug design requires representing small molecule drug candidates in computational formats. Cheminformatics provides the algorithms and mathematical graph representations needed to search chemical space, evaluate drug-likeness, and calculate molecular fingerprints.

1. Cheminformatics with RDKit in Python

from rdkit import Chem
from rdkit.Chem import Descriptors, Lipinski, AllChem

# Parse SMILES string into an RDKit Mol object
aspirin = Chem.MolFromSmiles("CC(=O)Oc1ccccc1C(=O)O")

# Calculate physicochemical properties
mw = Descriptors.MolWt(aspirin)          # Molecular Weight: 180.16 g/mol
logp = Descriptors.MolLogP(aspirin)       # Octanol-water partition coefficient: 1.31
hbd = Lipinski.NumHDonors(aspirin)        # Hydrogen bond donors: 1
hba = Lipinski.NumHAcceptors(aspirin)     # Hydrogen bond acceptors: 3
tpsa = Descriptors.TPSA(aspirin)          # Topological Polar Surface Area: 63.6 A^2

# Lipinski's Rule of 5 Check (Oral Bioavailability)
is_drug_like = (mw <= 500) and (logp <= 5) and (hbd <= 5) and (hba <= 10)
print(f"Drug-like: {is_drug_like}")

# Generate Morgan Circular Fingerprint (ECFP4 equivalent)
fp = AllChem.GetMorganFingerprintAsBitVect(aspirin, radius=2, nBits=2048)

2. Major Chemical Databases

• ChEMBL: Curated database of bioactive molecules with quantitative binding affinities (IC50, Ki, Kd) against biological targets.
• PubChem: Massive repository of >110 million chemical substances and biological assay results.
• ZINC20: Curated collection of >1 billion commercially available compounds ready for virtual screening.