Screening of key genes in infectious ARDS and development of a disease risk prediction model

Abstract

Infection-associated acute respiratory distress syndrome (ARDS) is a common and severe clinical disease in critical care medicine, with high mortality and a significant burden on intensive care resources. Current treatments mainly rely on supportive care, and there is still a lack of effective targeted drugs and diagnostic tools. Therefore, in-depth exploration of its molecular mechanisms and identification of key genes have important clinical significance. Objective To screen infectious ARDS key genes, construct a risk prediction model, and explore targeted drugs. Methods We extracted GSE32707 data from the GEO database, analyzed infection-related ARDS samples and control samples, screened for differentially expressed genes, and identified signaling pathways using GO and KEGG methods. Core genes were subsequently selected for molecular docking using HadDOCK. Notably, this study integrates multiple analytical approaches, including differentially expressed gene analysis, pathway enrichment, ceRNA network construction, immune infiltration characterization, and molecular docking into a systematic framework, which to our knowledge has not been previously applied to infectious ARDS in a coordinated manner. By integrating ceRNA networks to reveal immune characteristics, we validated key gene expressions and predicted targeted therapies. Results In patients with infectious ARDS, 3,034 differentially expressed genes were identified, enriched in T cell receptors, autophagy, and the PI3K/Akt/mTOR signaling pathway. Network analysis pinpointed key genes including ATG5, AKT1, ATG7, and BMP9, with ATG5 strongly associated with CD4+ T cells and BMP9 linked to NK cells. Validation revealed elevated expression of BMP9, AKT1, and ATG5 in ARDS patients, while ATG7 showed low expression, with these genes showing certain diagnostic potential that warrants further validation. Based on the modeling of key genes to predict disease risk, the calibration curve showed acceptable reliability. Molecular docking suggested that sorafenib could bind to these target genes at the molecular level, indicating its theoretical potential as a candidate worthy of further investigation, though this computational finding requires experimental validation. Conclusion BMP9, AKT1, ATG5, and ATG7 may contribute to ARDS pathogenesis through immune cell infiltration and signaling pathway regulation. The risk model and sorafenib-targeted strategy provide preliminary insights and new directions for clinical diagnosis and treatment though further validation is needed.