AI GlossaryㄱTechnical words in the news
structure predictor
An AI tool that calculates the actual 3D folded shape of a molecule just by looking at a one-dimensional string of information, like an amino acid sequence.
In plain words
A structure predictor is a program that looks only at a sequence—a one-line blueprint—and computes ahead of time what three-dimensional shape that molecule will fold into. Think of it like looking at origami instructions and picturing in your head what the finished shape will look like. A protein is a long, thread-like chain of amino acids, and what it does depends entirely on how that thread twists and folds.
This prediction matters because if you want two molecules to bind tightly to each other, you first need to know each one's shape. Get the shape wrong, and no matter how carefully you design the sequence, the result won't actually bind in the lab. That's why, in the work of designing new proteins, a structure predictor is used together with the task of choosing a sequence.
Recently, rather than being used on its own, this kind of predictor has been evolving to also calculate the shape a target and a new protein form together when they meet. This kind of calculation—folding two structures at once—has its own name, which you can find in a related entry.
See also
Stories using this term
- Anthropic Releases Optimization Code for Biology ModelsAI · 2026.09.19
- Claude Beats Industry Average Success Rate in Designing Drug-Binding ProteinsAI · 2026.08.19
- Google Research confirms limits of transfer learning in genetic risk predictionAI · 2026.09.04
- Liquid AI open-sources on-device benchmark PipetteAI · 2026.08.25
- Anonymous model Ox Alpha matches GLM-5.2 on all 60 tokenizer testsAI · 2026.08.23
- Google unveils GlucoFM, a dual-stream glucose prediction modelAI · 2026.08.27
