← Tools
Fold Watch
Paste a protein sequence. Get structure prediction, misfolding risk, aggregation hotspots, and drug strategy. Runs in your browser — your data never leaves your machine.
WHAT IT DOES
Structure
Radius of gyration, fold class, Chou-Fasman secondary structure with nucleation
Risk
Per-residue aggregation propensity, burial-aware scoring, polyQ/prion detection
Strategy
CHARGE (amyloid), HYDROPHOBIC (IDP), STABILIZE (unstable fold)
Speed
Milliseconds per protein. 11.5 million/sec on GPU (Metal batch). 66K/sec single-thread C. Your laptop is enough.
HOW IT WORKS
Chou-Fasman secondary structure with nucleation rules (helices need 4+ consecutive high-propensity residues). TANGO-like aggregation scoring per residue (charged residues suppress aggregation). Flory scaling for Rg with fold-class corrections. No training data — pure physics.
INSTALL
pip install begump
from gump.foldwatch import analyze, fold, water_fold_batch, keystones
# Analyze sequence
r = analyze("MVLSPADKTNVKAAWGKVGAH...")
print(r['misfolding_risk']) # low / medium / high
# GPU batch folding (11.5M/sec on Apple Silicon)
results = water_fold_batch(["FVNQHLCGS..."] * 10000)
# Find structural keystones (destabilization targets)
k = keystones("FVNQHLCGSHLVEALYLVCGERGFFYTPKT")
print(k['keystones'][0]) # highest-impact residue
Your data stays on your machine. No cloud. No API keys. Deterministic.