This AI Solved a 50-Year Mystery and Won a Nobel Prize

In December 2018, at eleven at night in a hotel room in Cancún, John Jumper got a phone call: DeepMind wanted to go all in on one of biology's oldest problems, and he was in charge. Proteins fold into shapes that decide what they do, and for fifty years working out a single shape could take a lab months or years. Two years later, in lockdown, his team's AlphaFold predicted shapes almost as accurately as experiments, and in 2024 it won the Nobel Prize in Chemistry. This video follows the people behind the paper: the restart that made things worse before they got better, the "distogram barrier", the teammates' questions that became the Evoformer and the triangle updates, the 50-paper test that showed they were only halfway, the confidence score that came out of the pandemic, and the long nights on a coronavirus protein called ORF8. Along the way it explains how AlphaFold actually works: reading evolution from related sequences, building a protein in 3-D from a cloud of pieces, and scoring how sure it is. Chapters: 0:00 Intro 1:18 The shape of life 2:55 Fifty years of measuring 5:24 The competition 7:00 A games lab 8:23 The first AlphaFold 9:45 The strike team 10:19 AlphaFold two 11:47 The barrier 12:20 Reading evolution 13:01 The Evoformer 14:16 How it learned 15:12 The second half 17:12 The result 18:54 After 20:06 What it can't do 20:53 What it means Paper: Jumper et al. Highly accurate protein structure prediction with AlphaFold. Nature 596, 583–589 (2021). https://www.nature.com/articles/s4158... Sources and further reading: Jumper et al. (2021). Highly accurate protein structure prediction with AlphaFold. Nature 596, 583-589. https://www.nature.com/articles/s4158... Senior et al. (2020). Improved protein structure prediction using potentials from deep learning (AlphaFold at CASP13). Nature 577, 706-710. https://www.nature.com/articles/s4158... Kryshtafovych, Schwede, Topf, Fidelis, Moult (2021). Critical assessment of methods of protein structure prediction (CASP), Round XIV. Proteins 89, 1607-1617. https://pmc.ncbi.nlm.nih.gov/articles... Ewen Callaway (2020). 'It will change everything': DeepMind's AI makes gigantic leap in solving protein structures. Nature 588, 203-204. https://www.nature.com/articles/d4158... DeepMind (2020). AlphaFold: a solution to a 50-year-old grand challenge in biology. https://deepmind.google/discover/blog... DeepMind (2022). AlphaFold reveals the structure of the protein universe. https://deepmind.google/discover/blog... EMBL-EBI. AlphaFold Protein Structure Database. https://alphafold.ebi.ac.uk/ The Royal Swedish Academy of Sciences (2024). The Nobel Prize in Chemistry 2024: press release and popular information. https://www.nobelprize.org/prizes/che... Christian B. Anfinsen (1972). Studies on the principles that govern the folding of protein chains. Nobel Lecture. https://www.nobelprize.org/uploads/20... Cyrus Levinthal (1969). How to fold graciously. In Mössbauer Spectroscopy in Biological Systems, University of Illinois Press, 22-24. Cooper et al. (2010). Predicting protein structures with a multiplayer online game. Nature 466, 756-760. https://www.nature.com/articles/natur... Mosalaganti et al. (2022). AI-based structure prediction empowers integrative structural analysis of human nuclear pores. Science 376, eabm9506. https://www.science.org/doi/10.1126/s... John Jumper (2024). Nobel Lecture: AlphaFold. https://www.nobelprize.org/prizes/che... Yasemin Saplakoglu (2024). How AI Revolutionized Protein Science, but Didn't End It. Quanta Magazine. https://www.quantamagazine.org/how-ai... Jumper et al. (2021). Applying and improving AlphaFold at CASP14. Proteins 89, 1711-1721. https://pmc.ncbi.nlm.nih.gov/articles... Tunyasuvunakool et al. (2021). Highly accurate protein structure prediction for the human proteome. Nature 596, 590-596. https://www.nature.com/articles/s4158... Documentary footage: The Thinking Game (2024), dir. Greg Kohs. #AlphaFold #AI #Biology