Absorption and distribution
Compare permeability, intestinal absorption, blood-brain barrier, plasma protein binding, solubility, and distribution signals in the context of molecular properties.
ADME-Tox and molecular property prediction
ADMET prediction that keeps uncertainty visible.
Prioritize small molecules using absorption, distribution, metabolism, excretion, toxicity, physicochemical properties, drug-likeness, and model applicability—without confusing an in silico prediction with experimental evidence.
ADME-ToxDMPKDrug-likenessPhysicochemical propertiesToxicity predictionApplicability domain
A connected ADMET profiler
Review developability alongside potency, structure, and provenance.
ADMET is not a single score. MolexIO organizes endpoint-level predictions around the decisions medicinal chemistry and DMPK teams make during hit triage and lead optimization. Each result remains attached to the molecule, prediction engine, run settings, and available measured data.
Metabolism and clearance
Inspect CYP interaction, metabolic stability, clearance, and related DMPK liabilities while retaining the endpoint definition and prediction source.
Toxicity and safety flags
Surface hERG, hepatotoxicity, mutagenicity, and other supported toxicity endpoints as prioritization signals that require appropriate experimental follow-up.
Drug-likeness and descriptors
Use molecular weight, logP, TPSA, hydrogen-bond counts, rotatable bonds, QED, and rule-based filters to understand why a candidate passes or fails.
Confidence and applicability
See whether a model is supported for the molecule and avoid turning unsupported or out-of-domain outputs into false precision.
Multi-parameter decisions
Carry ADMET evidence into MPO, selectivity, synthetic feasibility, docking, and assay review instead of optimizing one endpoint in isolation.
From structure to a reviewable decision
A practical ADMET workflow for hit triage and lead optimization.
- 01
Standardize
Resolve chemical identity and calculate reproducible physicochemical descriptors from the submitted structure.
- 02
Predict
Run supported ADME and toxicity models while preserving engine, version, input, and endpoint metadata.
- 03
Interpret
Compare values, confidence, structural alerts, applicability, and agreement instead of relying on a traffic-light label alone.
- 04
Prioritize
Combine developability evidence with potency, selectivity, docking, synthesis, and assay plans for the next DMTA cycle.
Scientific boundary
Prediction supports prioritization. It does not establish safety or pharmacokinetics.
An ADMET model estimates properties from available training evidence and molecular representation. Results can guide which compounds to test, which liabilities to investigate, and where chemistry may need to change. They are not clinical, regulatory, or experimental determinations. MolexIO makes measured, calculated, predicted, and missing evidence distinguishable so teams can act without overstating certainty.
Evaluate ADMET in the context of the whole discovery program.
Bring a molecule, series, or virtual-screening result and review developability signals alongside the evidence that produced the candidate.