🧬 De Novo Protein Design Lab
Interactive 3D simulator where a residue chain begins as scrambled noise and a diffusion-style denoising process gradually pulls it into a chosen folded shape, showing how generative AI sculpts brand-new protein backbones from randomness.
The simulation demonstrates the reverse-diffusion process at the heart of generative protein design: a chain of residues starts as a disordered random walk and is iteratively denoised toward a chosen target fold family.
🔬 What It Demonstrates
The simulation demonstrates the reverse-diffusion process at the heart of generative protein design: a chain of residues starts as a disordered random walk and is iteratively denoised toward a chosen target fold family.
🎮 How to Use
Pick a target fold family, adjust the diffusion speed, and press Play to watch the chain organize itself from noise into structure; press Rebuild to scramble a fresh random design and run it again.
💡 Did You Know?
Real tools like RFdiffusion run this same coordinate-denoising process hundreds of times per second across GPU clusters, generating thousands of candidate protein backbones before a handful are ever synthesized in a lab.
Interactive 3D simulator where a residue chain begins as scrambled noise and a diffusion-style denoising process gradually pulls it into a chosen folded shape, showing how generative AI sculpts brand-new protein backbones from randomness.
3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install