NVIDIA OpenFold3 · NIM API

Metabolite
Structure Explorer

Upload your cancer metabolite panel as a CSV. We resolve each compound via PubChem, then let you predict its co-folded 3D structure against oncology-relevant protein targets using NVIDIA OpenFold3.

01
Upload CSV
Metabolite names auto-resolved to SMILES via PubChem
→
02
Select Target
Choose from curated cancer-relevant protein targets
→
03
Predict & Visualize
3D co-folded structure rendered with pLDDT confidence coloring

Upload Your Panel

CSV must have a metabolite_name column. Optional: cancer_type column.

⬆
Drop your CSV here or browse
Accepts .csv · max 2 MB
Expected format
metabolite_name,cancer_type
3-Hydroxybutyric acid,breast
L-Glutamine,breast
Citric acid,colorectal
Palmitic acid,prostate
🔬
3 Predictions / Session
Hard limit to manage NVIDIA NIM API usage. Choose your combinations wisely.
🧬
PubChem Auto-Resolution
Metabolite names are resolved to canonical SMILES automatically. Failed lookups are flagged for manual entry.
⚡
OpenFold3 NIM
NVIDIA's GPU-optimised microservice predicts co-folded protein-ligand structures on par with AlphaFold3.