Proteins aren't static statues; they are shape-shifting ensembles, and we can finally predict all their 'moods' at once.
April 17, 2026
Original Paper
Polyformer: a generative framework for thermodynamic modeling of polymeric molecules
arXiv · 2604.14241
The Takeaway
For years, the gold standard in biology has been finding the single, perfect 3D structure of a protein—the approach made famous by AlphaFold. But in reality, proteins are constantly wiggling and changing shape depending on the temperature, and that movement is often how they actually do their jobs. This new model, Polyformer, doesn't just guess one 'best' shape; it predicts the entire 'cloud' of possible shapes a molecule takes as it heats up or cools down. This is a massive shift because it moves us past a 'frozen' view of biology into a dynamic one. For regular people, this means better drug design, because many diseases happen when a protein 'wiggles' the wrong way, not just when its basic shape is off.
From the abstract
The classic paradigm of structural biology is that the sequence of a biomolecule (protein, nucleic acid, lipid, etc) determines its conformation (shape) which determines its biological function. Protein folding programs like AlphaFold address this paradigm by predicting the single best conformation given a sequence that defines the molecule. However, biomolecules are not static structures, and their conformational ensemble determines their function. We present the Polyformer -- a generative fram