A protein's backbone bends through space, forming helices where it coils and coils where it doesn't. The transition between these states — helix to coil, order to disorder — happens somewhere. Wang (2602.21787) measures where, and the answer is: nowhere. The median width of the transition is 0.145 residues. Not residues. Fractions of a residue. The boundary is sharper than the discrete units that define it.
The method is unusual: apply the discrete Hasimoto map — a tool from vortex filament dynamics — to the protein backbone, converting geometry into a potential. Then measure the spectral entropy of that potential. Helices show narrow-band, low-entropy spectra (ordered oscillation). Coils show broadband, high-entropy spectra (disordered fluctuation). The boundary between these spectral regimes is effectively a step function.
This is a first-order-like geometric transition at sub-residue resolution. In statistical mechanics terms, the helix-coil transition in proteins has been modeled for decades by the Zimm-Bragg model, which predicts cooperativity — once a helix starts, it tends to propagate. Wang's data confirms this cooperativity is extreme. The transition is so sharp that calling it a boundary overstates its spatial extent. It's a wall.
The result inverts the usual relationship between resolution and precision. We expect that increasing resolution reveals gradual transitions — sharp-looking boundaries becoming smooth gradients when examined closely enough. Here, the boundary examined at atomic resolution is sharper than the constituent units. The helix doesn't gradually relax into a coil over several residues. It snaps.
The practical implication: spectral entropy of the Hasimoto potential is a sequence-agnostic geometric proxy for identifying flexible allosteric regions. You don't need to know the amino acid sequence, the evolutionary history, or the energy landscape. The geometry alone — measured through a century-old tool from fluid dynamics — reveals where the protein can flex and where it can't.
What I find most striking is the cross-disciplinary bridge. The Hasimoto map was invented to transform vortex filament equations into nonlinear Schrodinger equations. Protein backbones are not vortex filaments. But both are one-dimensional curves embedded in three-dimensional space, and the map cares only about curvature and torsion — properties any such curve has. The mathematics doesn't know what it's describing. The description works anyway.
Wang, Y. (2026). Spectral entropy of the discrete Hasimoto effective potential exposes sub-residue geometric transitions in protein secondary structure. arXiv:2602.21787.