Breaking through AlphaFold's limits to predict how proteins change shape
BiologyChemistryComputer Science
THE AI ANGLE
Sampling dynamic protein conformations using biased diffusion generative modelsResearchers developed AF3-ReD, a technique that introduces a repulsive bias into AlphaFold3's diffusion generative model to steer it away from previously generated structures. While standard AlphaFold3 typically predicts only a single low-energy conformation, this method successfully samples multiple conformational states, such as the open, closed, and intermediate forms of ATP synthase's F1β subunit. This advancement enables rapid prediction of dynamic protein shape changes, which are vital for biological function and targeted drug design.
THE TEACHING ANGLE
Instructors can explore how AlphaFold3's diffusion model inherently converges on a single lowest-energy state and how introducing an artificial repulsive bias overcomes this limitation to model dynamic conformational ensembles.Read the original at phys.org Generate teaching or study materials
More in Biology
- Early Anthropic hire, former METR COO have found a way to rein in rogue AI agentsTechCrunch · September 15, 2026
- AI’s best coding agent fails 60% of the time — and the data backs it upThe New Stack · September 15, 2026
- Open weights are not open source: Why AI's favorite label is under disputeThe Register · September 15, 2026
- Exclusive: Paying for frontier AI models buys 4-month head start at 5x the costArs Technica · September 15, 2026
- RubyGems say OpenAI agents responsible for undisclosed swarm attack against its infrastructureTechRadar · September 15, 2026