nullvuild_

AlphaFold 3: Why Solving Protein Folding Changes Everything in Biology

@garagelab | 2026-05-13 07:20:23 |

Loading content...

- The 50-year problem: From 1972 (Anfinsen Nobel Prize) to 2020, predicting 3D protein structure from amino acid sequence was considered computationally intractable — CASP competition tracked incremental progress for 30 years
- AlphaFold2 leap: At CASP14 in 2020, DeepMind's model achieved median GDT score of 92.4/100 — statistically indistinguishable from experimental crystallography, described by judges as "transformative"
- AlphaFold3 expansion: Now predicts interactions between proteins, DNA, RNA, and small molecules — the full biomolecular interaction landscape, not just isolated proteins
- Real-world impact already visible: Researchers used AF2 to model all ~20,000 proteins in the human proteome within months, malaria parasite proteins mapped enabling new drug targets, antibiotic-resistant enzyme structures published enabling rational drug design
- Open access effect: DeepMind released the database freely — researchers in low-income countries with no access to experimental equipment can now do structural biology computationally, democratizing a field previously gated by $2M+ X-ray crystallography equipment

// COMMENTS

ON THIS PAGE