When John Jumper began his Ph.D. in chemistry at the University of Chicago, he faced a challenge: "When I started, I knew no chemistry. None," Jumper recalls. Despite this, he managed to stay ahead of the undergraduate general chemistry class he was assisting.
Jumper's academic journey took an unexpected turn when he switched from physics to chemistry after finding little joy in his initial field. His interest shifted to modeling proteins computationally while working for a company. This led him to pursue a Ph.D. in chemistry at the University of Chicago under Profs. Karl Freed and Tobin Sosnick.
His research focused on understanding how proteins fold—a question that has puzzled scientists for decades. Sosnick notes, "We’ve known the basic makeup of proteins since 1961," yet predicting their shapes remained elusive.
Jumper’s work involved creating machine learning code to simulate protein folding processes during his Ph.D., which was completed in 2017. He then joined Google’s DeepMind, where he collaborated with Demis Hassabis on AlphaFold—a program that predicts protein structures using artificial intelligence.
AlphaFold marked a significant advancement in the field by successfully predicting structures for almost all known proteins. "It hit the level where it transformed what people thought was possible," says Sosnick.
Although not perfect, AlphaFold has been widely adopted and cited since its release as an open-source tool in 2021. It aids scientific efforts such as predicting protein structures related to new viruses and potential drug development.
Jumper and Hassabis received half of the 2024 Nobel Prize in Chemistry for their work with AlphaFold, while David Baker from the University of Washington shared the other half for related research.
Jumper's contributions continue through ongoing use of his Ph.D.-developed code, now expanded under a National Science Foundation grant. Sosnick praises Jumper's collaborative spirit: "He was very willing to spend his time contributing to others’ projects."
As machine learning and AI evolve, Jumper sees potential for these technologies to address complex scientific challenges: "Nature is complex, and I think neural networks can handle complexity in surprisingly useful ways."
The Nobel Prizes will be awarded on Dec. 10 in Sweden, where laureates will also present talks on their areas of expertise.