SMILES to structure

Each line is read as one SMILES, laid out in 2D with ring templates and real bond angles, and drawn in the style of journal figures. Stereocentres get wedges that reproduce the @ and @@ of the SMILES, and double bonds keep the geometry given by / and \.

The canonical SMILES is written by our own canonicaliser: atoms are ranked by iterated invariant refinement with tie breaking, in the manner of Weininger's CANON algorithm, with stereocentres and double bond geometry taken into the ranking. Two writings of the same molecule give the same canonical SMILES, so it can be compared as text.

SVG and PNG show exactly the preview, with the background you chose; PNG is rendered at two or four times the drawing size for slides and documents. The MOL file is written in BIOVIA's current V3000 connection table, the newest version of the CTfile format, with the 2D coordinates, charges, isotopes and wedge bonds, ready for any structure editor; older editors that only read V2000 need the 2D drawing as SVG or PNG instead.

The 3D view builds a model with every hydrogen from standard bond lengths and angles, keeps the stereocentres and double bond geometry of the SMILES, and relaxes it with a small force field. It is an idealised geometry, not a measured or computed minimum; the 3D SDF carries it for docking or modelling software.

Lower-case atoms are aromatic and are given alternating double bonds; a SMILES whose aromatic ring cannot be written that way, such as a pyrrole written without its [nH], is flagged. Atoms that carry more bonds than their valence allows are flagged too, with the atom number counted from the left of the SMILES.