A claim-specific evidence-boundary model for plant UDP-glycosyltransferases: from substrate candidates to native function

UDP-glycosyltransferase (UGT) research now generates far more candidate genes and metabolites than can be assigned experimentally. This Hypothesis and Theory article introduces a claim-specific evidence-boundary model developed from potential inferential transitions encountered when interpreting published UGT evidence; no new experimental data are presented. The model separates two entry filters from two complementary but non-interchangeable claim axes. Experimental accessibility determines which UGT–acceptor–donor combinations can be tested, whereas predictive prioritization orders testing and can redirect efforts to obtain compounds or donors. Once a defined combination enters an experiment, enzyme-dependent product detection and structural characterization delimit the chemical claim, while heterologous reconstruction and native perturbation delimit biological claims. We hypothesize that new evidence will preferentially increase confidence in the claim directly addressed by that evidence, whereas cross-axis evidence alone will not resolve an outstanding chemical or biological claim. The model extends identification and reporting guidance to UGT reaction and native-function claims. It predicts uneven confidence gains across the chemical and biological axes and identifies transitions where claims may outpace evidence. The model generates three testable predictions. Aligned evidence should produce larger confidence gains for the corresponding claim than for a non-aligned claim. Claim inflation should be enriched at prediction-to-reaction, feature-to-structure, and non-native-to-native transitions relative to other evaluable inferential transitions. Boundary-specific reporting should improve inter-evaluator agreement and reduce misclassification when factual information is held constant. The framework may support study design, peer review, and reusable UGT records.