Protein Structure Prediction

Proteins do most of the work inside living cells, and what they do depends largely on the three-dimensional shapes they fold into. Protein structure prediction is the field devoted to inferring those shapes computationally from a protein's amino acid sequence. For decades this was one of biology's hardest open problems, while experimental methods such as X-ray crystallography and cryo-electron microscopy remained slow and expensive.

Deep learning changed that. Systems like DeepMind's AlphaFold showed that neural networks could predict many structures with accuracy approaching experimental methods. The approach has since spread across academic labs, startups, and large pharmaceutical companies. Researchers now also use these models to design entirely new proteins, from enzymes to binders engineered to latch onto disease targets.

Why it matters now

The technology is moving from research benchmarks into commercial pipelines. AI-assisted drug candidates are advancing into late-stage clinical testing, protein design success rates are improving, and pharma companies and AI-native biotechs are striking more partnerships and investment deals. Large structural datasets, including collections covering viral proteins, are also expanding what models can learn from.

What you'll find here

• New models and methods for predicting and designing protein structures
• Drug discovery milestones, including clinical trial progress for AI-derived candidates
• Deals, funding, and partnerships across pharma, biotech, and AI companies
• Datasets and infrastructure that power structural biology research
• Enzyme and therapeutic design advances and their real-world applications

This hub tracks how predicting protein shapes is changing the way medicines and materials are made, for readers who follow computational biology, pharmaceutical R&D, or applied AI.