De Novo Design of Protein Nanopores: From Minimal Peptides to AI-Driven Design

Protein nanopores offer a powerful platform for analytical applications by directly converting chemical properties into information-rich, single-molecule electrical readouts. However, engineering natural nanopores for sensing remains challenging due to the limited number of available scaffolds and evolutionary constraints on their geometries and chemistries. Recent advances in AI- and data-driven protein design are enabling increasingly complex membrane protein architectures, opening new possibilities for designing synthetic nanopores with tailored folds and functions. This review surveys the evolution of de novo nanopore design, from minimal sequences to AI-supported approaches, highlights emerging strategies, and outlines key challenges─including data scarcity, membrane modeling, and experimental characterization─that must be addressed to realize robust, programmable nanopores for next-generation sensing technologies.