Body Composition Architecture and Injury Topology in Physically Active Young Adults: A Tanglegram-Based...

Background/Objectives: Injury occurrence in physically active young adults is considered a multifactorial phenomenon influenced by body composition and training-related characteristics. This study aimed to investigate the multidimensional relationships between body composition, training context variables, and injury phenotypes in physically active university students using exploratory multivariate approaches. Methods: The study included 418 physically active university students. Participants completed questionnaires regarding injury history, physical activity, and sport participation and underwent standardized anthropometric and body composition assessments. Analyses included Kendall's Tau correlations, multiple correspondence analysis (MCA), hierarchical clustering, heatmap phenotyping, tanglegram-based clustering analysis, and distance-based redundancy analysis (dbRDA). The tanglegram analyses were intended as exploratory structure-matching procedures designed to evaluate similarities in hierarchical organization between domains rather than direct biological associations, causal relationships, or predictive effects. Results: Weak but significant associations were observed between selected body composition variables and injury outcomes, particularly for skeletal-muscle-related indicators and lower limb injuries. MCA and clustering analyses identified partially differentiated sport-training profiles and exploratory injury-burden phenotypes. Topology-based cross-domain matching analyses suggested partial structural correspondence between body composition, training context, and injury phenotypes; however, the most anatomically coherent patterns were observed for local body composition variables. Nevertheless, overall cross-domain concordance remained weak-to-moderate. dbRDA demonstrated statistically significant but weak associations for body composition (adjusted R² = 0.027, p = 0.001) and the combined explanatory model (adjusted R² = 0.022, p = 0.023), whereas the training context model was not significant (adjusted R² = 0.002, p = 0.304). Conclusions: Injury occurrence was weakly associated with body composition and training context characteristics within a multidimensional exploratory framework. The findings are consistent with the interpretation of injury occurrence as a heterogeneous and predominantly multifactorial phenomenon and highlight the utility of multidimensional exploratory approaches for investigating complex injury-related patterns.