Cheminformatics · C01

What is cheminformatics?

A chemical drawing is easy for a person to recognize. A computer needs a representation whose parts can be counted, compared, and transformed.

Reviewed 13 September 2026

From drawing to data

A molecular graph represents atoms as nodes and bonds as edges. This simple idea supports many operations: finding a substructure, comparing neighborhoods, or calculating a molecular property from a defined representation.

The data also need context. A record may include a name, identifier, charge, stereochemistry, measured property, conditions, source, and date. A structure without context can answer only limited questions.

Useful questions

A database can find molecules containing a functional group, retrieve a known compound by identifier, or group similar structures. Algorithms can calculate descriptors and build models that relate a chosen representation to observed data.

These tasks are useful in research, education, safety, and information management. They work best when the input, assumptions, and limits are stated plainly.

Representation is a decision

One molecule can be stored as a drawing, a connection table, a SMILES string, or a fingerprint. Each is convenient for different tasks. A fingerprint can make similarity search fast but is not a complete chemical explanation.

Before trusting a result, ask: what structure was supplied, how were salts and charges handled, which representation and settings were used, and what data support the conclusion?

Continue learning →

Sources and further reading

structuremolecular graphsearchable record
structure → molecular graph → searchable record