Anthropic and a research team used Claude to determine the first three-dimensional structure of a prokaryotic vault, cutting a purification step from days to 90 minutes and structure computation to 34 hours. It is a rare public case of an LLM embedded across wet-lab and structural-biology workflows, though methodological detail remains thin.
Protein Structure Prediction
蛋白质结构预测
Computational inference of the three-dimensional conformation formed by amino acid sequences
Curated Intelligence
Anthropic's biology team used Claude to show for the first time that bacterial and phage relatives of vault proteins can assemble into vault particles, cutting particle isolation from about three days to 90 minutes and producing a first 3D structure model in 34 hours. The project took under four months, with researchers deciding and Claude executing and suggesting, demonstrating LLM productivity in structural biology discovery.
Bloomberg reports that Isomorphic Labs, the Alphabet unit spun out of Google DeepMind, is negotiating a roughly $5B raise at about a $50B valuation, despite having no drug yet tested in humans. Its core asset is a drug-design engine built on AlphaFold 3 that predicts protein folding and how candidate molecules bind their targets, letting chemists rank and discard weak designs on a computer before paying for synthesis and animal studies. Without human data, the $50B price rests on prediction accuracy and pharma partnerships rather than evidence that its molecules work in patients.
Anthropic's life sciences team used Claude to determine the first 3D structure of a vault particle from bacteria, cutting a task that would have taken weeks to months of manual work down to days. The work also shows that vault-forming proteins exist in bacteria and their viruses, offering simpler models for probing vault function in human cells, while the team candidly documents the simple mistakes Claude made and how they were fixed.
The DeepMind AlphaFold team published a five-year retrospective highlighting the model's breakthrough in protein structure prediction since 2020. The article emphasizes that AlphaFold has accelerated global biological discovery and served as the first major proof point for AI advancing science, earning Nobel Prize recognition.