Biophysics and artificial intelligence: convergent approaches to characterize pathological protein aggregates in neurodegeneration

To understand what has gone wrong in a neurodegenerative disease, one must first establish what is actually different. In many disorders, that difference is carried by the protein assemblies themselves: the same polypeptide arranged into a distinct conformer, polymorph, oligomeric state, or kinetic species. Detecting and quantifying these differences is therefore not a preliminary step, but the foundation on which mechanistic understanding, neuropathological marker development, and therapeutic strategies all rest. Before anything else, it is an analytical problem.

This Research Topic addresses that problem by bringing together quantitative biophysics and artificial intelligence. Biophysical approaches provide measurable characteristics of protein assemblies, including their size, stoichiometry, conformation, stability, interaction partners, and the kinetics of their assembly and propagation. Artificial intelligence can help resolve differences hidden within heterogeneous, rare, or high-dimensional observations; predict aggregation propensity and partner binding; and enhance the design, analysis, and interpretation of biophysical measurements. Together, these approaches can help define the features that distinguish pathological assemblies from benign or physiological species and translate those features into transferable platforms for detecting and classifying pathological protein assemblies across neurodegenerative diseases.

We welcome experimental, computational, and hybrid contributions in which a well-defined parameter is measured, validated, and connected to mechanistic insight, neuropathological marker development, or therapeutic strategy. Contributions may address pathological protein assemblies associated with a range of neurodegenerative disorders and should emphasise quantitative, interpretable, and reproducible approaches across laboratories. Particular interest will be given to studies that demonstrate how artificial intelligence can complement biophysical measurements, improve the detection or classification of disease-associated species, or generate experimentally testable hypotheses. Both methodological advances and applications to established or emerging proteinopathies are encouraged.