name: cheminformatics description: Chemical informatics and modeling license: MIT compatibility: opencode metadata: audience: cheminformaticians, computational chemists, researchers category: chemistry
What I do
- Apply computational methods to chemical data analysis
- Design and search chemical databases
- Predict molecular properties and reactivity
- Perform molecular similarity analysis
- Develop quantitative structure-activity relationships (QSAR)
- Visualize and analyze chemical structures
When to use me
- When searching chemical databases
- When predicting molecular properties
- When building QSAR/QSPR models
- When analyzing molecular similarity
- When virtual screening compounds
- When managing chemical data
Key Concepts
Molecular Descriptors
Constitutional Descriptors
- Molecular weight
- Atom counts (C, H, O, N, etc.)
- Number of rings
- LogP (lipophilicity)
Topological Descriptors
- Wiener index
- Balaban index
- Connectivity indices
- Hydrogen bond donors/acceptors
# Example: Simple molecular fingerprint
def morgan_fingerprint(molecule, radius=2):
"""Generate Morgan/ECFP fingerprint."""
# Simplified representation
return {
'features': extract_substructures(molecule, radius),
'bit_vector': encode_as_bits(features),
'similarity': lambda other: tanimoto_coefficient(features, other)
}
Chemical File Formats
- SMILES: String representation
- SDF: Structure-data file
- MOL: MOL file format
- PDB: 3D structure
- InChI: IUPAC identifier
Similarity Metrics
- Tanimoto coefficient (Jaccard)
- Dice similarity
- Cosine similarity
- Euclidean distance