MACHINE LEARNING-BASED IDENTIFICATION OF BIOMARKERS AND IMMUNE INFILTRATION IN PEDIATRIC TELOMERE-ASSOCIATED ALLERGIC RHINITIS AND ASTHMA: IMPLICATIONS FOR RESPIRATORY FUNCTION AND PHYSICAL PERFORMANCE

Abstract

Background: Childhood Allergic Rhinitis and Asthma Syndrome (CARAS) is a chronic inflammatory condition affecting both the upper and lower airways, significantly impairing respiratory function, exercise capacity, and physical performance in children. The syndrome is closely linked to immune dysregulation, yet the biomarkers associated with its pathogenesis remain inadequately explored. Identifying key genetic and immune-related markers is crucial for improving early diagnosis, targeted interventions, and the management of physical activity limitations in affected children. Objective: This study aimed to identify telomere-associated biomarkers of CARAS in children using machine learning approaches and to evaluate the role of immune cell infiltration in disease progression, with potential implications for optimizing respiratory function, physical fitness, and sports participation. Methods: Gene expression data from the GSE19187 dataset was retrieved from the Gene Expression Omnibus (GEO) database for pooled analysis. Differentially expressed genes (DEGs) were identified using the ‘limma’ package in R software. Least Absolute Shrinkage and Selection Operator (LASSO) regression and Support Vector Machine Recursive Feature Elimination (SVM-RFE) algorithms were applied to detect key candidate biomarkers. Diagnostic accuracy was assessed using Receiver Operating Characteristic (ROC) curve analysis, while CIBERSORT was utilized to quantify the immune cell infiltration in CARAS samples. Correlations between immune cells and identified biomarkers were analyzed using Pearson's correlation test. Results: A total of five telomere-associated genes (ABCC3, NT5DC2, RRAGD, ST6GAL1, and YBX3) were identified as potential biomarkers of CARAS using machine learning algorithms. Significant correlations were observed between these genes, with NT5DC2 emerging as the most prominent key gene. ROC analysis confirmed strong diagnostic efficacy for these biomarkers. Immune cell infiltration analysis revealed distinct immunological profiles associated with CARAS, highlighting their role in disease pathogenesis and potential implications for exercise-induced respiratory adaptation in children with asthma. Conclusions: The study identifies ABCC3, NT5DC2, RRAGD, ST6GAL1, and YBX3 as potential biomarkers of CARAS, with NT5DC2 being the most critical. The ferroptosis pathway emerged as the most significantly enriched signaling pathway, suggesting a novel target for therapeutic intervention. These findings have direct implications for the development of exercise-based rehabilitation strategies, as optimizing immune response and respiratory function in children with CARAS may enhance aerobic capacity, endurance, and participation in physical activities. Future research should explore the role of these biomarkers in exercise-induced airway modulation and personalized rehabilitation programs for children with respiratory conditions.