Integrated machine learning to predict the prognosis of lung adenocarcinoma patients based on SARS‐COV‐2 and lung adenocarcinoma crosstalk genes

Abstract Viruses are widely recognized to be intricately associated with both solid and hematological malignancies in humans. The primary goal of this research is to elucidate the interplay of genes between SARS‐CoV‐2 infection and lung adenocarcinoma (LUAD), with a preliminary investigation into their clinical significance and underlying molecular mechanisms. Transcriptome data for SARS‐CoV‐2 infection and LUAD were sourced from public databases. Differentially expressed genes (DEGs) associated with SARS‐CoV‐2 infection were identified and subsequently overlapped with TCGA‐LUAD DEGs to discern the crosstalk genes (CGs). In addition, CGs pertaining to both diseases were further refined using LUAD TCGA and GEO datasets. Univariate Cox regression was conducted to identify genes associated with LUAD prognosis, and these genes were subsequently incorporated into the construction of a prognosis signature using 10 different machine learning algorithms. Additional investigations, including tumor mutation burden assessment, TME landscape, immunotherapy response assessment, as well as analysis of sensitivity to antitumor drugs, were also undertaken. We discovered the risk stratification based on the prognostic signature revealed that the low‐risk group demonstrated superior clinical outcomes ( p < 0.001). Gene set enrichment analysis results predominantly exhibited enrichment in pathways related to cell cycle. Our analyses also indicated that the low‐risk group displayed elevated levels of infiltration by immunocytes ( p < 0.001) and superior immunotherapy response ( p < 0.001). In our study, we reveal a close association between CGs and the immune microenvironment of LUAD. This provides preliminary insight for further exploring the mechanism and interaction between the two diseases.