Machine Learning-Based Identification of Diagnostic Markers for Allergic Constitution and Bioinformatics Exploration of the Mechanism of"Allergy Kang"in Regulating Constitution

YANG Fan; MA Zhaotian; WU Qinglu; LI Wanyang; KONG Jingwei; ZONG Yuhan; WANG Manting; Wendusubilige; DONG Guangda; WANG Ji

doi:10.3969/j.issn.1006-3250.2026.01.012

Abstract：Objective To identify diagnostic markers and key immune cells for allergic constitution using machine learning methods,and to explore the mechanism of"Allergy Kang"in regulating allergic constitution.Methods Differentially expressed genes(DEGs)in individuals with allergic constitution—comprising patients with various allergic diseases from the Gene Expression Omnibus(GEO)database—were analyzed compared to normal individuals.Weighted Gene Co-expression Network Analysis(WGCNA)was used to identify genes associated with allergic constitution.Protein-protein interaction(PPI)analysis,Gene Set Enrichment Analysis(GSEA),and Gene Ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)pathway enrichment analyses were performed.Machine learning algorithms(LASSO regression and SVM-RFE)were employed to identify diagnostic markers for allergic constitution.The CIBERSORT algorithm was used to assess the infiltration levels of 22 types of immune cells,and the correlation between diagnostic markers and immune cells was analyzed.Active components and related targets of"Allergy Kang"were obtained from the batman-TCM and TCMSP platforms,and an"Allergy Kang-Active Components-Allergic Constitution Targets"network was constructed to explore the mechanism of constitution regulation.Results A total of 360 DEGs were identified in individuals with allergic constitution compared to normal individuals.WGCNA revealed 816 genes significantly associated with allergic constitution(correlation coefficient=-0.47,P<0.001),which were enriched in epithelial cell dysfunction and complement and coagulation cascades.Machine learning identified three genes as diagnostic markers for allergic constitution,with SERPINB2 demonstrating the most stable diagnostic performance in external validation.Five immune cells showed differential infiltration in allergic constitution,with activated dendritic cells and mast cells being significantly upregulated(P<0.001).SERPINB2 was closely correlated with key immune cells involved in allergic processes.Core targets of"Allergy Kang"in regulating allergic constitution included ALDH7A1,EDN1,TGFβ1,TP53,and AR.The molecular mechanisms primarily involved the regulation of histidine metabolism,cell surface receptor signaling,MAPK signaling pathways,and PPAR signaling pathways.Key active components(such as ganoderic acid DM and 5-O-methylvisammioside)showed significant binding to core targets of these signaling pathways(e.g.,MAPK13,PPARγ).Conclusion SERPINB2 is a key diagnostic marker for allergic constitution.Upregulated infiltration of activated dendritic cells and mast cells represents an immune characteristic of allergic constitution.Targeting MAPK and PPAR signaling pathways to modulate immune cell communication is a potential mechanism by which"Allergy Kang"regulates allergic constitution.