Abstract: Objective To screen for potential biomarkers of abdominal aortic aneurysm(AAA)using single-cell RNA(scRNA)data analysis, weighted gene co-expression network analysis(WGCNA), machine learning, and immune infiltration analysis. Methods The scRNA sequencing data containing AAA and normal aorta control(NAC)in the gene expression database were downloaded and processed by data quality control, dimensionality reduction, differential analysis, and cell type annotation. Chronological analysis was proposed to screen for the earliest differentiated cell types during AAA genesis, and to screen for differentially expressed genes(DEGs). High dimensional WGCNA(hdWGCNA)was performed to identify AAA-related gene modules, and enrichment analysis was conducted. Conventional transcriptome sequencing data containing AAA and NAC was downloaded for differential analysis and WGCNA. DEGs of scRNA samples, DEGs of conventional transcriptomes and WGCNA results were integrated to screen genes associated with AAA lesions. Gene ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)signaling pathway enrichment analysis were carried out. The potential biomarkers of AAA were screened using the least absolute shrinkage and selection operator(LASSO), support vector machine recursive feature elimination(SVM-RFE), and random forest(RF)machine learning methods. The immune infiltration analysis was performed. Results The results of scRNA data analysis showed that endothelial cell was the earliest cell type to differentiate during AAA development, and a total of 853 scRNA DEGs were obtained. hdWGCNA identified 2 gene modules associated with AAA, which were significantly enriched in the signaling pathways of T helper 17 cell differentiation, and T helper 1 and 2 cell differentiation. Conventional transcriptome analysis yielded a total of 162 DEGs. Integration yielded 17 AAA-associated genes, which significantly enriched in signaling pathways such as chemokines, T helper 17 cell differentiation, and T helper 1 and 2 cell differentiation. The machine learning algorithm identified a potential biomarker for AAA, ecotropic viral integration site 2B(EVI2B). The expression of EVI2B was higher in AAA samples than in NAC samples. The immune infiltration results showed that the proportions of naive B cells, plasma cells, activated dendritic cells and neutrophils were higher in AAA samples than in NAC samples. EVI2B was positively correlated with M2 macrophages, M1 macrophages, CD8 T cells, plasma cells, helper follicular T cells, M0 macrophages, and neutrophils; and it was negatively correlated with resting dendritic cells. Conclusion AAA pathogenesis involves a variety of immune cells and signaling pathways, and EVI2B expression is significantly increased in AAA samples, correlating with a variety of immune cells, which may be a new target for AAA treatment.

Key words: Abdominal aortic aneurysm, Weighted gene co-expression network analysis, Machine learning, Biomarkers, Immune infiltration