Over the last decades, advances in omics technologies have substantially expanded our understanding of the immunological processes underlying allergic diseases, redefining how disease endotypes are established. These high-throughput approaches have important clinical implications, enabling the discovery and screening of novel diagnostic biomarkers and the identification of potential therapeutic targets. Omics disciplines encompass multiple molecular layers, including genomics and epigenomics (DNA and small RNA levels), transcriptomics (RNA level), proteomics (protein level), as well as metabolomics and lipidomics (metabolite level).Nevertheless, the vast amount and complexity of the data generated require sophisticated bioinformatic tools and integrative analytical strategies. Importantly, the integration of multiomics data with non-omics information, such as clinical and laboratory parameters, remains an unmet need in the field.Artificial intelligence (AI), particularly machine learning (ML), offers powerful approaches for data integration and for uncovering how distinct biomarkers are interconnected through immune mechanisms. These strategies allow us to move beyond single-disease analyses toward identifying shared molecular signatures across related diseases, such as allergic rhinitis and atopic dermatitis. In this context, Zhang and colleagues identified 36 differentially expressed genes common to both diseases, mainly associated with epithelial barrier dysfunction and immune activation pathways. Using ML algorithms, five potential biomarkers (CD274, SERPINB4, CYP2E1, SPRR1B, and FOLH1) were highlighted, underscoring the interplay between innate and adaptive immunity and epithelial-immune cell interactions (Zhang et al., 2025).Epigenomics, mainly focused on post-transcriptional gene regulation mechanisms, emerged as a useful tool for distinguishing disease-specific patterns. In this context, 47 genome-wide in children with cow's milk allergy (CMA) revealed 48 differential methylation patterns (DMPs) in specific CpG islands, particularly between IgE-49 mediated, non-IgE-mediated CMA and non-allergic children. These were associated with 50 genes such as LDHC, TRAF3IP3, TMCO3 or BCL11 (Lopez-Gomez et al., 2025). Interestingly, 51 differential CpG methylation was also detected in genes potentially linked to 52 acquisition, distinguishing tolerant from non-tolerant children after six months of a exclusion 53 diet (e.g., RNF39, LTB4R, or LTB4R2). Overall, these findings are relevant for the identification 54 of diagnostic and prognostic biomarkers in food allergy and tolerance development.
Fernández-Santamaría et al. (Thu,) studied this question.