SenescenceBioKG reveals full-text-supported biomaterial design, mechanisms, and regeneration patterns in senescence-oriented tissue engineering

Background Senescence-oriented biomaterial studies provide evidence regarding material design, responsive behavior, cellular targets, mechanisms, regeneration, and validation across different full-text sections. However, publication-level labels cannot reliably answer questions that require several of these relations to occur in the same study. Methods We constructed SenescenceBioKG based on 283 original experimental or translational full-text studies. Candidate relations were generated under a typed ontology, adjudicated against same-record evidence, linked to unique evidence identifiers, and canonicalized into a queryable graph. Fifteen competency questions were evaluated using graph traversal and an independently implemented dataframe method. A local read-only R Shiny Explorer was developed for evidence inspection. Results The frozen graph contained 654 canonical nodes, 1,868 supported edges, and 1,868 linked evidence records. Within the curated corpus, strict material responsiveness comprised 64 RESPONDS TO edges from 55 studies and 55 material platforms, led by acidic pH and reactive oxygen species. Same-study analyses recovered responsive-design modules and 35 complete target-cell–mechanism–regeneration modules. Mesenchymal stem/stromal cells, senescent cells, and macrophage or innate immune populations were recurrent targets, whereas oxidative-stress control, mitochondrial homeostasis, and NF-κB/Senescence-associated secretory phenotype (SASP) signaling frequently accompanied bone, wound/skin, cartilage, and disc endpoints. Validation was concentrated in Cell and Small-Animal studies; Organoid/chip and Large-Animal evidence was sparse, and no Clinical edge was supported. All 15 queries achieved exact-set agreement between the two implementations, with macro precision, recall, and F1 values of 1.00. Conclusion SenescenceBioKG provides a full-text-supported structure for examining smart and responsive biomaterial design, fabrication, cellular evidence, regenerative outcomes, and validation gaps. Its provenance model separates co-supported evidence from causal inference and supports reproducible, material-centered retrieval.