Bioinformatics combined with machine learning analysis of diagnostic biomarkers related to immune infiltration in the pathogenesis of keloids

Object Exploring key markers in keloid pathogenesis and correlation with immune cells based on bioinformatics and machine learning algorithms.Methods Keloid fibroblast microarray dataset was obtained from GEO database, weighted gene co-expression network(WGCNA) and differential gene analysis were constructed to screen differentially expressed genes (DEGs), Functional enrichment analysis was used to identify the main pathways, protein-interaction network combined three machine learning algorithms to further detect key genes, draw ROC feature curves to diagnose the value of key genes, and analyze the distribution of immune cells and the correlation between key genes and immune infiltrating cells.Results A total of 80 DEGs were identified by WGCNA combined with differential gene analysis, which were mainly enriched in the pathways of regionalization, skeletal system morphogenesis and embryonic skeletal system development. The ROC curve showed that HOXC4 has good diagnostic ability and is highly expressed in keloid patients. Immunocorrelation analysis showed that HOXC4 was significantly positively correlated with resting NK cells and negatively correlated with activated dendritic cells and activated NK cells.Conclusions HOXC4 has been identified as a key biomarker in keloid pathogenesis and is associated with immune cells, providing a theoretical basis for keloid diagnosis.