Open data phylometabolomics reveals turnover-dominated chemical divergence and clade-specific physicochemical regimes across angiosperms.

Specialized metabolites are central to plant defense, signaling and ecological interactions, yet we still lack a macroevolutionary framework explaining how this diversity is structured across angiosperms. Here, we integrate the open LOTUS chemical repository with standardized taxonomy to map a curated 'chemical core' comprising 77 404 family-supported occurrences across 12 representative families spanning Magnoliids, Monocots, and Eudicots. Chemical composition shows strong higher-level structure, delineating major lineages while revealing a striking evolutionary topology: inter-lineage divergence is dominated by metabolite replacement rather than accumulation. Across family pairs, β-diversity is overwhelmingly explained by turnover, indicating that chemical disparity rarely arises from one lineage retaining subsets of another's repertoire. Despite this universal turnover regime, lineages occupy distinct physicochemical and elemental neighborhoods, consistent with divergent evolutionary strategies under shared allocation constraints. Magnoliids define a lipophilic boundary characterized by greater investment in nitrogen-bearing and aromatic-rich defenses; Monocots occupy a more hydrophilic, high-molecular-weight and structurally saturated niche; and Eudicots expand oxygen-rich carbon scaffolds with reduced nitrogen dependence. Together, our results indicate that angiosperm chemical evolution is highly dynamic at the compositional level, yet constrained at the architectural level, with persistent turnover generating lineage-specific chemical identities within inherited physicochemical corridors. This work provides a reproducible open data foundation for testing mechanistic links between biosynthetic organization, ecological antagonists, and evolutionary diversification of plant chemistry.