DIAGNOSTIC TRANSCRIPTOMIC SIGNATURES FOR MYCOPLASMA PNEUMONIAE PNEUMONIA TO GUIDEPEDIATRIC PATIENT TREATMENT

Mycoplasma pneumoniae is a frequent cause of atypical pneumonia in children and young adults. Its lack of a cell wall makes it resistant to β-lactam antibiotics, the first-line treatment for classical bacterial pneumonia. Current microbiological diagnostic tests for M. pneumoniae are often slow, insensitive or non-specific, which leads clinicians to prescribe empirical antibiotics. Therefore, there is an urgent need for accurate, rapid and host-based diagnostic tools to guide antimicrobial therapy more precisely. This study aims to delineate the host blood transcriptomic response associated with M. pneumoniae pneumonia and to derive minimal gene expression signatures that reliably distinguish M. pneumoniae pneumonia from other bacterial and viral pneumonias in pediatric patients. Using blood microarray data from 107 children hospitalized with pneumonia — including 30 with confirmed M. pneumoniae infection — we performed differential expression analyses followed by a resampling-based penalized regression (LASSO) approach to identify optimal small-size transcriptomic signatures. These signatures (3 to 10 transcripts) were then validated in an independent RNA-seq cohort of pediatric pneumonia patients. In addition, we performed pathway-level analysis (GSVA) to characterize functional host-response differences between M. pneumoniae and viral pneumonias. We identified eight robust transcriptomic signatures (each comprising 3–10 genes) that discriminate M. pneumoniae pneumonia from other bacterial or viral pneumonias with high accuracy (AUC range: 0.84 – 0.95). Existing broadly used signatures to distinguish bacterial from viral infections/pneumonias performed poorly in identifying M. pneumoniae . Our new signatures retained strong performance in the independent RNA-seq validation cohort, demonstrating their robustness and reproducibility. Pathway analysis revealed that M. pneumoniae pneumonia triggers distinct host immune and inflammatory pathways compared to viral pneumonia, supporting the biological plausibility of the identified signatures. Mycoplasma pneumoniae pneumonia produces a unique host transcriptomic response that can be captured by small, robust signatures. These diagnostic gene signatures have the potential to significantly improve the timely and accurate identification of M. pneumoniae in pediatric pneumonia patients, inform antimicrobial decision-making (e.g., macrolide vs β-lactam), and help reduce inappropriate antibiotic use. Adoption of such transcriptome-based diagnostics could enhance patient management and antimicrobial stewardship in clinical practice.