SMILES

SMILES (Simplified Molecular Input Line Entry System) is a line notation that encodes the structure of a chemical species as a short ASCII string. It was devised in the 1980s by David Weininger at the US EPA laboratory in Duluth, Minnesota, and subsequently developed and commercialized by Daylight Chemical Information Systems. Because Daylight's specification left some behaviour implementation-defined, the Blue Obelisk open-source chemistry community published OpenSMILES in 2007 as an open, precisely-worded standard; more recently IUPAC has worked toward a standardized SMILES+ variant.

At its core SMILES is a graph serialization. A molecule is modeled as a chemical graph whose nodes are atoms and whose edges are bonds. To produce the string, hydrogens are (usually) made implicit, the graph is reduced to a spanning tree by a depth-first traversal, and the atoms are written in visit order. Ring cycles that the traversal broke are reconnected with matching numeric labels (e.g. cyclohexane is C1CCCCC1), and branches are enclosed in parentheses such as FC(F)F. Bonds are single by adjacency or written explicitly as = (double), # (triple), $ (quadruple), and : (aromatic); disconnected components are joined by a period. Organic-subset atoms (B, C, N, O, P, S, halogens) can be written bare with implied valence-filling hydrogens, while brackets carry isotope, charge, explicit hydrogen count and stereochemistry, as in [NH4+] or [13C].

The notation captures more than plain connectivity. Aromatic atoms are written in lowercase (benzene is c1ccccc1); tetrahedral chirality uses @ and @@; and double-bond cis/trans geometry uses the directional bonds / and \. Isomeric SMILES bundles these stereo and isotope features. A single molecule admits many equally valid strings (CCO, OCC and C(O)C all denote ethanol), so toolkits apply canonicalization algorithms to derive a unique string for database indexing—though canonical output is toolkit-specific and not interchangeable across vendors.

SMILES is the de facto interchange format for 2D chemical structure and underpins a family of related notations: SMARTS for substructure queries, SMIRKS for reaction transforms, and BigSMILES for polymers, alongside the registry-oriented InChI and InChIKey and the connection-table formats Molfile and SDF. It is read and written by essentially every cheminformatics toolkit, including RDKit, Open Babel, and ChemDoodle. Its strengths are compactness, human readability and graph-theoretic rigor; its main limitations are the non-canonical ambiguity noted above and its restriction to topology and stereochemistry, omitting 3D coordinates and conformation.