A molecule is an electrically neutral group of atoms held together by chemical bonds. Atoms share electrons in covalent bonds, while weaker noncovalent interactions influence molecular shape and associations with other molecules. In drug discovery, molecules are practical starting points for representing chemical structures as inputs to machine-learning models. section 1.2 section 1.2.1
Computational workflows commonly represent molecules in several forms:
- SMILES encodes atoms, bonds, and their connections as a compact, space-free text string governed by a grammar. section 1.2.1
- Molecular graphs represent atoms and bonds as connected computational structures for machine-learning models. section 1.2
- Compounds are molecules containing at least two different element types in a fixed proportion; the terms are used interchangeably here. section 1.2.1
Reading simple SMILES
Examples include CC for ethane, O=C=O for carbon dioxide, C#N for hydrogen cyanide, C1CCCCC1 for cyclohexane, and c1ccccc1 for benzene.

