Abstract
Summary
Neurodegenerative proteinopathies are defined by distinct proteins whose folding, trafficking and clearance depend on shared proteostasis machinery. Whether these disorders engage common components or only common processes has been unresolved. We mapped 16 disease proteins from 12 disorders onto a fixed 990-protein proteostasis network. Across the 12 disorders, the neighborhoods of these proteins shared specific proteins beyond expectations from process composition and connectivity. The same held for the two disorders defined by more than one protein. Nine proteins contributed disproportionately, and only autophagy’s removal significantly reduced similarity. Removing PRKN, HSPA8 or UBQLN2 in silico reduced path efficiency, marking them candidate shared network vulnerabilities. In human brain proteomes, Alzheimer’s disease-zone proteins showed greater abundance remodeling in the insoluble fraction; recurrence across disorders showed no such association. The result is a component-level architecture prioritizing network-derived candidates for cross-disease functional testing.