Daylight Theory — Cheminformatics Fundamentals
The Daylight Theory Manual is the canonical reference for the molecular languages underlying modern cheminformatics: SMILES, SMARTS, SMIRKS, and fingerprints. These are not Daylight-proprietary — they are industry-standard formats implemented in RDKit, OpenBabel, CDK, and every major cheminformatics toolkit.
Router — What to Read
| Topic |
Reference |
| Molecular graph model, aromaticity, chirality, SSSR, reaction representation |
references/molecules.md |
| SMILES syntax: atoms, bonds, branches, rings, stereochemistry, reactions |
references/smiles.md |
| SMARTS query language: primitives, operators, recursive SMARTS, reaction queries |
references/smarts.md |
| SMIRKS reaction transforms: atom maps, grammar, stereochemistry |
references/smirks.md |
| Fingerprints: structural keys, path-based FP, folding, Tanimoto, Tversky, all similarity measures |
references/fingerprints.md |
| Chemical database concepts: hash tables, identifiers, in-memory search, pools, hitlists |
references/cheminformatics-databases.md |
Core Languages at a Glance
| Language |
Purpose |
Example |
| SMILES |
Encode a specific molecule |
CC(=O)Oc1ccccc1C(=O)O |
| SMARTS |
Describe a molecular pattern |
[OH]c1ccccc1 (phenol) |
| SMIRKS |
Encode a reaction transform |
[C:1][Br:2]>>[C:1][I:2] |
Key Principles
- Molecular graph: atoms = nodes, bonds = edges; properties are explicit or derived
- Aromaticity: Hückel 4N+2 rule, all ring atoms sp² — a derived, not stored, property
- SMILES organic subset: B, C, N, O, P, S, F, Cl, Br, I — no brackets needed at normal valence
- SMARTS unspecified = unrestricted:
O in SMARTS matches any aliphatic oxygen; in SMILES it means water
- Tanimoto: c/(a+b+c) — the standard fingerprint similarity; double-zero independent
Related ALKYL Skills
rdkit — implementation of these concepts in Pythonscientific-skills:matchms — spectrum similarity (same mathematical ideas)
1---2name: daylight-theory3description: Daylight Theory — Cheminformatics Fundamentals4---56# Daylight Theory — Cheminformatics Fundamentals78The Daylight Theory Manual is the canonical reference for the molecular languages underlying modern cheminformatics: SMILES, SMARTS, SMIRKS, and fingerprints. These are not Daylight-proprietary — they are industry-standard formats implemented in RDKit, OpenBabel, CDK, and every major cheminformatics toolkit.910## Router — What to Read1112| Topic | Reference |13|-------|-----------|14| Molecular graph model, aromaticity, chirality, SSSR, reaction representation | `references/molecules.md` |15| SMILES syntax: atoms, bonds, branches, rings, stereochemistry, reactions | `references/smiles.md` |16| SMARTS query language: primitives, operators, recursive SMARTS, reaction queries | `references/smarts.md` |17| SMIRKS reaction transforms: atom maps, grammar, stereochemistry | `references/smirks.md` |18| Fingerprints: structural keys, path-based FP, folding, Tanimoto, Tversky, all similarity measures | `references/fingerprints.md` |19| Chemical database concepts: hash tables, identifiers, in-memory search, pools, hitlists | `references/cheminformatics-databases.md` |2021## Core Languages at a Glance2223| Language | Purpose | Example |24|----------|---------|---------|25| SMILES | Encode a specific molecule | `CC(=O)Oc1ccccc1C(=O)O` |26| SMARTS | Describe a molecular pattern | `[OH]c1ccccc1` (phenol) |27| SMIRKS | Encode a reaction transform | `[C:1][Br:2]>>[C:1][I:2]` |2829## Key Principles3031- **Molecular graph**: atoms = nodes, bonds = edges; properties are explicit or derived32- **Aromaticity**: Hückel 4N+2 rule, all ring atoms sp² — a derived, not stored, property33- **SMILES organic subset**: B, C, N, O, P, S, F, Cl, Br, I — no brackets needed at normal valence34- **SMARTS unspecified = unrestricted**: `O` in SMARTS matches any aliphatic oxygen; in SMILES it means water35- **Tanimoto**: c/(a+b+c) — the standard fingerprint similarity; double-zero independent3637## Related ALKYL Skills3839- `rdkit` — implementation of these concepts in Python40- `scientific-skills:matchms` — spectrum similarity (same mathematical ideas)
Kdevos12