How proteins fold in milliseconds despite astronomical impossibility
Connected to the 2024 Nobel Prize in Chemistry
In 1969, molecular biologist Cyrus Levinthal posed a devastating question: A protein with just 100 amino acids can assume 1047 different shapes. Testing each at nanosecond speed would take 1027 years — longer than the age of the universe.
Yet real proteins fold correctly in milliseconds. How?
If proteins tried each conformation randomly:
That's 1017 times longer than the universe has existed!
Real proteins fold via the energy funnel:
Result: Correct fold in milliseconds to seconds!
Protein folding isn't a random search through a flat landscape — it's a guided descent through a funnel. The funnel's walls are made of energy: wrong conformations are energetically unfavorable, naturally pushing the protein toward the correct native state at the bottom.
Consider a 20-amino acid chain with hydrophobic (H) and polar (P) residues:
The chain naturally folds to hide hydrophobic residues in the core — this hydrophobic collapse is the main driving force, creating a natural energy funnel!
Demis Hassabis and John Jumper created AlphaFold, an AI that predicts protein structures with near-experimental accuracy.
David Baker used similar principles to design entirely new proteins that don't exist in nature.
By October 2024, AlphaFold had been used by 2+ million researchers in 190 countries. What once took years now takes minutes.
Levinthal's paradox was cited on the first page of the Nobel Prize scientific background — it perfectly illustrates the scale of the problem that computational methods have now cracked.
1969: Levinthal poses the paradox — random folding is impossible
1990s: Energy landscape theory explains how nature solves it
2020: AlphaFold achieves human-level accuracy in structure prediction
2024: Nobel Prize awarded for computational protein science
Understanding protein folding enables: