Designer Proteins on Demand: An AI Turns Plain-English Requests Into Working Molecules

Westlake University's Pinal writes proteins from typed requests; in lab tests 5 of 20 carbon-shuttle designs and 1 of 29 glowing-protein designs worked. Secondary report; the Nature paper was not opened.

TL;DR

Proteins are the molecules that do much of the work inside living cells, and designing new ones is how scientists build new enzymes and protein medicines. Researchers at Westlake University in China have built an AI system called Pinal that takes a plain sentence describing what a protein should do and writes out candidate designs, according to secondary reports of a paper in Nature. In the lab, 5 of 20 designs for a protein that ferries carbon worked, one in four, and the best raised output to 1.7 times that of the natural version. For a glowing protein, 1 of 29 worked. Going straight from words to working molecules is a step toward ordering a custom protein the way people order from a menu. To see our odds on when describing a protein becomes a reliable way to get one, and the timeline of breakthroughs that leads there, subscribe.

Below the line, for paid subscribers: The full story: where it stands, how we got here and what's dated next; our labeled speculation on what it adds up to; and every source.