Select a target
Explore mechanisms, observations and sources.
MELANIN TARGET SCREENING
Explore links tomelanin targets.
Virtual screening for melanin-related targets
36Melanin-related targets
—Evidence records
3Project tiers
Start your research here
Assess ADMET. Then explore targets.Open assessment →
Step 1ADMET assessment→Step 2Candidate review→Step 3Target screening
ADMET prediction via admetSAR
Predict associations. Preserve evidence. Associations do not establish skin-lightening efficacy or safety.
Interpret results ↗01 / ADMET PRE-ASSESSMENT
Assess risks before exploring associations
ADMET assessment → Candidate review → Target screening
Single-molecule research starts with ADMET assessment.
02 / CANDIDATE REVIEW
Choose molecules for target screening
Run prediction sends molecular structures to admetSAR at lmmd.ecust.edu.cn. Review each endpoint before choosing candidates.
Import existing normalized results
Use MelanoTarget normalized JSON with source metadata and matching molecule IDs / SMILES. This is an internal interchange format; raw admetSAR exports need an adapter. Imported results are labeled as user supplied.
| Select | Molecule / SMILES | Result status | Endpoints / applicability |
|---|
Nothing is selected automatically. Missing or failed molecule results cannot be selected. Partial results remain marked for review.
01 / MOLECULE ANALYSIS
Explore associations, one molecule at a time
Enter a structure to explore target associations and evidence.
02 / CANDIDATE SCREENING
Organize candidates. Prioritize research.
Upload molecules or import existing predictions.
03 / TARGET EVIDENCE LIBRARY
Every target has a research context
Read original studies, experimental conditions and observed effects.
TASKS & RESEARCH REPORTS
Tasks and research reports
View this session's tasks and download saved reports.
Analyzing
READY WHEN YOU ARE
Start with an analysis
Your results and reports will appear here.
ASSOCIATION-BASED RESULTS
Candidate research priorities
Associations do not establish direction, skin-lightening efficacy or safety. See the evidence library for study observations.
—Input molecules
—Scored molecules
—Excluded / all-zero
—Top-50 boundary ties
| Rank | Molecule / details | Target profile score | L1 / L2 / L3 hits | Notes |
|---|
Ranks apply to this batch only. Tie order does not indicate differences in efficacy.
MOLECULE → TARGET → EVIDENCE
Structure and applicability details
| Target / evidence | Project tier | Full prediction rank | Association score | Score contribution |
|---|
Select a gene for evidence. Association scores establish neither direct binding nor effect direction.
Excluded from prediction
Exclusion reasons are retained. These molecules have no model rank.
| Molecule | Status | Reason |
|---|
USER GUIDE & INTERPRETATION
Understand associations and their limits
A guide to inputs, scores and evidence.
What do predictions mean?
Predictions indicate molecule–target associations for research prioritization, not binding probability, effect direction or skin-lightening efficacy.
Scoring and special cases
Among the top 50 predictions, each L1 hit adds 1 point and each L2 hit adds 0.5. Break ties by L3 hit count, then input order.
Top-50 boundary ties trigger a score sensitivity range, not a confidence interval. All-zero results are not ranked. Scores indicate neither effective dose nor inhibitory potency.
Are observed pigment changes retained?
Yes. Increases and decreases in melanin are retained with their experimental conditions. They do not transfer directly to new molecules.
Interpreting different evidence types
Consider species, model, intervention and endpoint. Opposite findings may coexist across experiments. Pathway support is not target-specific validation.
L1 / L2 / L3 are project tiers, not clinical evidence levels.
What can I use now?
Enter a SMILES string, upload a molecule list or import complete predictions. Run admetSAR prediction directly, or import existing results for candidate review.
Inputs, environment and saved tasks
Molecule lists require compound_id and smiles. Imported predictions must cover all 1,848 targets per molecule.
The eight saved examples can be explored without a new prediction. New molecules are processed by the configured prediction service.
The server runs one target analysis at a time. Shared-server deployments isolate tasks by browser session; keep your browser cookies to return to saved records. If another target analysis is running, wait for it to finish before submitting yours. ADMET supports remote prediction, normalized result import and manual candidate selection.
What is in a report?
Molecule rankings, target details, structure QC, exclusions and an evidence snapshot.
Data and provenance
Reports retain run records, prediction provenance and literature sources separately. Runs created from an ADMET selection include its ADMET results and review record.
The target profile score corresponds to DirectScore. Historical CSV filenames are retained for compatibility.