Predict condensate phase diagrams from protein sequence
In plain words
A disordered protein's amino-acid sequence decides at what concentration it forms droplets. A method that predicts this concentration from sequence, across salt and temperature, would connect sequence to physics directly.
Precise statement
For intrinsically disordered regions of 50 to 500 residues, predict the saturation concentration $c_{\mathrm{sat}}(T,\mathrm{salt})$ and the dense-phase concentration from sequence alone, within a factor of 2 for $c_{\mathrm{sat}}$, across sequence variants with altered aromatic and charge patterning (stickers-and-spacers framework). Answer: a model validated on blind variants and on two-component mixtures.
What would settle it
Blind predictions of $c_{\mathrm{sat}}$ for newly designed sequence variants compared with measured phase diagrams.
Status in the literature
Unverified note
Residue-level coarse-grained models reproduce $c_{\mathrm{sat}}$ trends within families of disordered regions; accuracy for arbitrary sequences and mixtures is not established as of 2026, to this survey's knowledge.