Sources
Where the data comes from
Skinceptor is a navigation layer over public scientific resources. The current dataset is curated and static: 48 ingredients, 6 biology maps and 56 studies. It is a demonstration corpus, not an exhaustive index of cosmetic science.
PubChem
Visit source →Chemical identity: molecular formula, molecular weight, CID, SMILES and 2D structure images.
UniProt
Visit source →Protein and enzyme records for biological targets, including accession numbers and functional annotation.
NCBI Gene
Visit source →Gene symbols, official names and identifiers for the genes referenced in networks and topic maps.
Reactome
Visit source →Curated human pathway definitions used for the pathway tier of every network.
Gene Ontology
Visit source →Standardised biological process terms where a Reactome pathway is not the right level of description.
ChEBI
Visit source →Chemical entity ontology identifiers for cross-referencing molecules.
PubMed / Europe PMC
Visit source →Primary literature. Each curated study links to its PMID or DOI; unverified identifiers link to a title search instead.
Known limitations
- Coverage is uneven — well-funded ingredients have far more literature than newer actives.
- Many cosmetic studies are small, short, industry-funded, or in-vitro only; we label these explicitly.
- Absence of a relationship here means absence from our curated dataset, not absence from science.
- Nothing here accounts for formulation, concentration, pH, stability or individual biology.