Home / Science / AI Solves Protein Folding: Biology's iPhone Moment
AI Solves Protein Folding: Biology's iPhone Moment
24 Dec
Summary
- AlphaFold predicts protein shapes with atomic accuracy.
- A database of over 200 million protein structures is available.
- AlphaFold 3 now extends AI capabilities to DNA and RNA.
Until November 2020, DeepMind was primarily recognized for its AI mastering the game of Go. However, the organization then shifted its focus to protein folding, a complex scientific challenge. The resulting system, AlphaFold2, achieved remarkable accuracy in predicting the three-dimensional shapes of proteins.
This breakthrough culminated in a comprehensive database containing over 200 million predicted protein structures, representing the majority of known proteins. This resource is now indispensable for approximately 3.5 million researchers across 190 countries. The foundational 2021 Nature article detailing AlphaFold2 has garnered tens of thousands of citations.
The latest iteration, AlphaFold 3, arrived last year, expanding AI's predictive power to include DNA, RNA, and drug interactions. While challenges like 'structural hallucinations' in disordered protein regions persist, this advancement signifies a crucial progression toward future scientific understanding and application.
