Integrating machine learning and WGCNA to identify core candidate genes and decipher immune-related molecular networks in TcdA/TcdB-associated CDI.

Clostridioides difficile infection (CDI) remains a leading cause of healthcare-associated diarrhea and is characterized by high recurrence rates and inadequate predictive tools despite available therapeutic options. Current toxin- or nucleic acid-based assays do not fully capture host inflammatory dynamics, highlighting the need for mechanistically informed host-response studies that may complement existing microbiological approaches. This study aims to clarify the molecular mechanisms through which Clostridioides difficile (C. difficile) toxins toxin A (TcdA) and toxin B (TcdB) induce CDI and establish the theoretical foundation for more targeted interventions. We systematically identified core candidate genes associated with CDI using differential expression analysis of multiple datasets, weighted gene co-expression network analysis (WGCNA), and integration of TcdA/TcdB-related gene sets derived from multiple databases. We further integrated machine learning algorithms, protein-protein interaction (PPI) network analysis, and immune infiltration analysis to investigate the relationships between TcdA/TcdB-associated host responses and immune microenvironment remodeling, defined here as inferred changes in immune-cell composition. We identified 66 candidate genes located at the intersection of CDI-associated differential expression, WGCNA-derived genes, and TcdA/TcdB-related gene sets. Machine learning algorithms further narrowed these to nine core genes: Areg, Ptgs2, Cd40, Tlr7, Aif1, Cxcl10, Il6, Cybb, and Nos2. These genes occupy key positions within inflammatory and immune-related molecular networks associated with CDI. In addition, immune infiltration analysis suggested that dysregulation of the host immune microenvironment contributes substantially to disease progression. TcdA and TcdB may contribute to the onset and progression of CDI through host transcriptional and immune programs associated with specific inflammatory signaling pathways. The nine core genes identified through machine learning, together with immune infiltration and functional enrichment analyses, support the relevance of TcdA/TcdB-associated host-response networks in CDI. These findings provide a useful foundation for future mechanistic and experimental studies aimed at clarifying the molecular basis of TcdA/TcdB-associated CDI.