pubchem-guide

Scope

PubChem aggregates what depositors submit: compound records with computed properties, GHS classifications from regulators and companies, bioassay results from screening centers and from papers extracted by curated databases, and cross-references. Keyless. It reports what was deposited and computes nothing for a structure it has no record of.

Hand off to a sibling source when the question needs:

• article text or abstracts behind the PubMed IDs these tools return → a literature server
• curated potency comparisons, standardized per measurement type and target → a medicinal-chemistry bioactivity database
• approved-drug labels, adverse events, or drug–drug interactions → a drug regulatory or drug-label source
• protein function, sequence, or 3D protein structure behind a target accession → a protein knowledgebase or structure database

Identifiers

• CID — from pubchem_search_compounds in identifier mode. Besides names, SMILES, InChI, and InChIKey, a CAS Registry Number resolves as a name (67-64-1 → acetone, CID 180).
• Salt and parent are separate CIDs. "metformin" resolves to CID 4091 and "metformin hydrochloride" to CID 14219. Each carries its own assay data; nothing merges them.
• AID — from pubchem_search_assays (by target), or the AID on any bioactivity row or target entry.
• NCBI Gene ID — from a compound's target entries (GeneID:5743 in the target kind of pubchem_get_compound_interactions) or the target Gene ID on a pubchem_get_bioactivity row from a targeted assay (rows from untargeted screens carry none). pubchem_search_assays returns AIDs only, never Gene IDs.
• Protein accession (UniProt) — from the targets of an assay summary (pubchem_get_summary), or the target accession on a bioactivity row.
• Gene symbol — pubchem_search_assays wants the official symbol (PTGS2); an alias such as COX-2 matches nothing. Gene and protein summaries list aliases and the species once you hold the Gene ID or accession.
• Taxonomy ID — from a gene or protein summary.