Nuclear Receptor Ligand Activity Prediction

We collected ligand activity data for 12 nuclear receptor targets and developed prediction models for 10 targets. The workflow performs binder screening followed by functional classification for predicted active endpoints.

Explore Services

Hierarchical Workflow

From SMILES input to interpretable attention maps through an integrated prediction workflow.

12

curated NR datasets

10

modeled NR targets

Visits since launch 212

From Prediction to Interpretation

Each stage is tuned for clarity and interpretability. You get immediate signal, then deeper mechanistic insight when needed.

Binder Screening

Ensemble ML + GNN blending to detect binders with calibrated confidence.

Mechanism Classification

Agonist or antagonist classification for active targets, ready for review.

Attention Visualization

Atom-level highlights to explain why predictions land where they do.

Ensemble stacking Attention-aware outputs Complete workflow