Bio Molecular Io

Reading Molecules

Goal: Load molecules from SMILES strings, SDF files, or SMILES files into RDKit molecule objects.

Approach: Use Chem.MolFromSmiles for individual SMILES, SDMolSupplier for multi-molecule SDF files, and file iteration for SMILES files, filtering out parse failures.

from rdkit import Chem
from rdkit.Chem import AllChem

# From SMILES
mol = Chem.MolFromSmiles('CCO')

# From SDF file (single molecule)
mol = Chem.MolFromMolFile('molecule.mol')

# From SDF file (multiple molecules)
supplier = Chem.SDMolSupplier('library.sdf')
molecules = [mol for mol in supplier if mol is not None]
print(f'Loaded {len(molecules)} molecules')

# From SMILES file
with open('compounds.smi') as f:
    molecules = []
    for line in f:
        parts = line.strip().split()
        if parts:
            mol = Chem.MolFromSmiles(parts[0])
            if mol:
                mol.SetProp('_Name', parts[1] if len(parts) > 1 else '')
                molecules.append(mol)

Writing Molecules

from rdkit import Chem

# To SMILES
smiles = Chem.MolToSmiles(mol)  # Canonical SMILES
smiles_iso = Chem.MolToSmiles(mol, isomericSmiles=True)  # With stereochemistry

# To SDF file
writer = Chem.SDWriter('output.sdf')
for mol in molecules:
    writer.write(mol)
writer.close()

# To MOL block (string)
mol_block = Chem.MolToMolBlock(mol)

Structure Standardization

Goal: Normalize molecular representations to a canonical form for consistent comparison and analysis.

Approach: Apply a multi-step pipeline: sanitize valences, normalize functional groups, neutralize charges, canonicalize tautomers, and strip salts using rdMolStandardize.

Use rdMolStandardize module (Python MolStandardize was removed Q1 2024).

from rdkit import Chem
from rdkit.Chem.MolStandardize import rdMolStandardize

def standardize_molecule(mol):
    '''
    Full standardization pipeline.
    Order: Sanitize -> Normalize -> Neutralize -> Canonicalize tautomer -> Strip salts
    '''
    if mol is None:
        return None

    # Sanitize (assign valences, kekulize)
    try:
        Chem.SanitizeMol(mol)
    except Exception:
        return None

    # Normalize (standardize functional groups)
    normalizer = rdMolStandardize.Normalizer()
    mol = normalizer.normalize(mol)

    # Neutralize charges where possible
    uncharger = rdMolStandardize.Uncharger()
    mol = uncharger.uncharge(mol)

    # Canonicalize tautomers
    enumerator = rdMolStandardize.TautomerEnumerator()
    mol = enumerator.Canonicalize(mol)

    # Remove salts/fragments (keep largest)
    remover = rdMolStandardize.FragmentRemover()
    mol = remover.remove(mol)

    return mol

# Standardize a library
standardized = [standardize_molecule(m) for m in molecules]
standardized = [m for m in standardized if m is not None]

Open Babel Conversion

For format conversions not supported by RDKit.

# Open Babel 3.x import (not 'import pybel')
from openbabel import pybel

# Read MOL2 (better supported in Open Babel)
mols = list(pybel.readfile('mol2', 'ligands.mol2'))

# Convert to SDF
output = pybel.Outputfile('sdf', 'output.sdf', overwrite=True)
for mol in mols:
    output.write(mol)
output.close()

# Format conversion
for mol in pybel.readfile('pdb', 'complex.pdb'):
    mol.write('mol2', 'ligand.mol2', overwrite=True)

Molecular Drawing

Use rdMolDraw2D (legacy Draw.MolToImage deprecated).

from rdkit import Chem
from rdkit.Chem.Draw import rdMolDraw2D

def draw_molecule(mol, filename, size=(400, 300)):
    '''Draw molecule to PNG file.'''
    drawer = rdMolDraw2D.MolDraw2DCairo(size[0], size[1])
    drawer.DrawMolecule(mol)
    drawer.FinishDrawing()
    with open(filename, 'wb') as f:
        f.write(drawer.GetDrawingText())

# Draw with highlighting
def draw_with_substructure(mol, pattern, filename):
    '''Highlight substructure match.'''
    match = mol.GetSubstructMatch(Chem.MolFromSmarts(pattern))
    drawer = rdMolDraw2D.MolDraw2DCairo(400, 300)
    drawer.DrawMolecule(mol, highlightAtoms=match)
    drawer.FinishDrawing()
    with open(filename, 'wb') as f:
        f.write(drawer.GetDrawingText())

Related Skills

• molecular-descriptors - Calculate properties after loading
• similarity-searching - Compare loaded molecules
• virtual-screening - Prepare ligands for docking