Science Friday How Alphafold Has Changed Biology Research, 5 Years On
6 snips
Nov 18, 2025 
In this engaging discussion, John Jumper, a scientist at Google DeepMind and co-recipient of the 2024 Nobel Prize in Chemistry, dives into the revolutionary impact of AlphaFold on protein folding and biological research. He explains how this AI tool predicts protein structures, enhancing our understanding of diseases and guiding vaccine design. John also discusses the challenges of AI in drug discovery and shares insights into future advancements like AlphaFold 3, which predicts interactions with DNA and RNA, paving the way for AI-driven drug design.
AI Snips
Chapters
Transcript
Episode notes
Why Protein Structure Matters
- Proteins fold into functional 3D shapes encoded by their amino-acid sequence and physics.
- Determining those structures experimentally is slow and expensive, often taking years and ~$100,000.
AlphaFold Trains On Historical Structural Data
- AlphaFold is a deep learning system trained on decades of experimentally determined structures in the Protein Data Bank.
- It predicts 3D protein structures and its confidence, often matching experimental accuracy.
Use Predictions To Guide Lab Work
- Use AlphaFold as a hypothesis generator to guide experiments and interpret data.
- Always validate AlphaFold predictions with laboratory experiments before drawing firm conclusions.