Generative Protein Design for Physiological Intervention: Progress and Horizons

Export The rapid convergence of de novo protein design, deep learning, and high-throughput screening has redefined what proteins can be engineered to do. Once limited to stabilizing structures or optimizing existing binders, computational design now enables the creation of entirely new protein architectures with programmable dynamics, specificity, and therapeutic function. This mini-review highlights key advances that collectively establish a framework for de novo-designed protein therapeutics, from miniprotein binders and molecular sensors to switches and regulatory systems. We summarize representative breakthroughs showing how rational design and generative modeling are beginning to capture the complexity of physiological regulation, enabling a priori avoidance of off-target interactions, precise antigen recognition on the cell surface, engineering of intracellular protein–protein interaction (PPI) blockers, and selective targeting of PPIs mediated by intrinsically disordered regions. Together, these developments illustrate a growing capacity to move beyond empirical optimization toward the predictive design of proteins as molecular tools for precise physiological intervention.