Ji, J.; Son, A.; Kang, M.-J.; Yeom, J.; Yoo, H. J.; Kim, K.; Kim, J.-H.; Oh, H. Y.; Kim, S. A.; Lee, S.-Y.; Lee, S.-H.; Hong, S.-J.; Kim, H.

Exposure to humidifier disinfectants has been linked to an array of pulmonary disorders and diminished lung functionality, particularly reduced Forced Vital Capacity (FVC). This investigation sought to identify diagnostic biomarkers for early detection of children at elevated risk of developing chronic respiratory conditions following such exposure. Our research employed a comprehensive multi-omics strategy analyzing 70 pediatric patients alongside 10 controls, seamlessly integrating clinical assessments with transcriptomics, methylomics, proteomics, and metabolomics data. The analytical framework utilized a sophisticated combination of Non-negative Matrix Factorization (NMF), Multi-Omics Factor Analysis (MOFA), and advanced machine learning algorithms. NMF clustering uncovered distinctive protein expression patterns associated with integrin-mediated signaling pathways and immune response mechanisms. Complementarily, MOFA identified latent factors correlating with lung function metrics, highlighting critical molecular pathways involved in integrin cell surface interactions and lipid metabolism regulation. Machine learning-based analysis facilitated the development of a multi-marker panel, comprising IGHV2-70, LysoPC (16:0), and hexadecyl ferulate, which achieved 81.46% accuracy in identifying pulmonary dysfunction cohort. These findings suggest that alterations in integrin-related signaling networks and dysregulation of lipid metabolism play pivotal roles in mediating the long-term pulmonary consequences of humidifier disinfectant exposure. The proposed multi-marker panel offers significant potential for enhanced risk stratification and timely therapeutic intervention.