Screening of therapeutic targets for mycoplasma pneumoniae pneumonia based on transcriptomics and machine learning

Mycoplasma pneumoniae pneumonia (MPP) is a common respiratory infection in children lacking targeted therapies. This study aims to identify potential therapeutic targets through transcriptomics analysis and machine learning. Transcriptome data from MPP patients and healthy controls (dataset E-MTAB-14588) were analyzed to identify differentially expressed genes (DEGs). Functional enrichment (KEGG, GO) was performed on DEGs. Disease-related genes were retrieved from GeneCards and CTD databases using MPP-related keywords. The intersection of DEGs and disease genes was visualized through protein-protein interaction (PPI) networks. Key genes were selected using LASSO, SVM-RFE, and random forest (RF) algorithms, and their diagnostic performance was assessed using ROC curve analysis. The expression of key genes was further validated by RT-qPCR in relevant cells. Immune cell infiltration and correlations with key genes were assessed using MCPCounter. Potential therapeutics were predicted using the CMAP database. A total of 369 DEGs between MPP patients and controls were identified, including 287 upregulated and 82 downregulated genes enriched in immune response and inflammatory pathways. Intersection with 452 MPP-associated disease genes revealed 18 overlapping targets. Machine learning methods consistently highlighted three key genes, namely BAK1, S100A9, and CR2, as critical in the pathogenesis of MPP. ROC analysis demonstrated excellent diagnostic performance with AUCs of 0.9415 (BAK1), 0.9357 (S100A9), and 0.9298 (CR2). RT-qPCR confirmed significant upregulation of BAK1 and S100A9 in Mp-infected A549 cells and downregulation of CR2 in Mp-stimulated Raji cells. These genes showed significant correlations with macrophage and neutrophil infiltration. Ten candidate small molecules with potential therapeutic effects were predicted. This investigation identified BAK1, S100A9, and CR2 as potential diagnostic and therapeutic targets for MPP, which might offer new insights into the underlying mechanisms and contribute to future research on the diagnosis and treatment of MPP.