User manual
ADMETatlas user manual
How to find, read, and reuse the curated peptide ADMET evidence in ADMETatlas — the entry points, record pages, evidence tables, and API, with the interpretation rules that keep each value scientifically faithful.
Molecular recordPeptide records
Peptide pages lead with identity and physicochemical descriptors, then summarize how much ADMET evidence is attached and link every source row.
What you see
- •Canonical identity — name, stable ATL ID, sequence, and representation class.
- •Physicochemical descriptors computed from the sequence (length, mass, net charge, pI).
- •Failure-mode evidence depth across the five formal ADMET modules — Exposure / Absorption, Distribution / Barrier, Metabolic & Proteolytic Stability, Excretion / Persistence, and Toxicity / Safety — plus a Physicochemical category (experimental lipophilicity / logD).
- •A complete evidence index for source-linked review and provenance tracing.
How to use it
- 1Start with the identity block to confirm the peptide and representation class.
- 2Read evidence depth as coverage — how much evidence is attached — not as a safety or developability score.
- 3Open a profile area or ADMET module to read its endpoint, unit, and assay-context detail.
- 4Follow the evidence index references to the original sources for reproducible analysis.
Notes: Coverage meters and profile areas show how much evidence is attached, not an overall safety, BBB, or developability score.
Clinical contextProduct records
Product pages place route, formulation, label, trial, and safety context next to the linked peptide identity.
What you see
- •Product identity, route/form context, sponsor/label context, and registry identity support when available.
- •A peptide link module stating whether product context can be connected to a canonical peptide identity.
- •Exposure / PK evidence and clinical, safety, label, immune-recognition, and nonclinical context lanes.
- •Reference lineage and data-access modules for product-level reproducibility.
How to use it
- 1Confirm the product identity and route/form fields before reading PK or clinical context.
- 2Read the peptide link decision before moving from product context to peptide-level interpretation.
- 3Treat label, trial, adverse-event, immune-recognition, and identity lanes as distinct evidence contexts.
- 4Use product references to audit source-level support, not as endpoint-value replacements.
Notes: Product label, trial, adverse-event, and identity-support entries stay context unless a reviewed link and explicit endpoint rules support peptide-level interpretation.
EvidenceEndpoints & evidence
Endpoint and evidence pages expose definitions, units or metric scales, assay context, values, and references.
What you see
- •Formal endpoint definitions with default units and minimum context requirements.
- •Measurements with endpoint, value, unit or metric scale, assay/model, condition, evidence layer, and reference link.
- •Evidence layers that separate comparable measurements, reported observations, and contextual support.
- •Coverage summaries that describe evidence volume, not safety or efficacy confidence.
How to use it
- 1Start from the endpoint page when comparing a formal ADMET property.
- 2Read unit or metric scale, assay model, matrix, species, route, dose, method, and population before comparing values.
- 3Use Evidence for measurement-level filtering and detail pages for full provenance.
- 4When exporting, keep evidence layer, review status, reuse policy, and provenance attached to each row.
Notes: Do not pool or rank measurements across incompatible assay contexts. Evidence counts and coverage bars indicate evidence availability only.