Identification of sialoylation-related genes as a new diagnostic biomarker of atherosclerosis by machine learning

Abstract

Abstract Background: Atherosclerosis (AS) is a pathological change based on the disorder of lipid metabolism, which is related to the inflammatory process of vascular wall and the high level of low-density lipoprotein. Sialoylation is a post-translational modification controlled by sialyltransferase, transporter and neuraminidase family. This process plays a key role in a variety of biological functions. Abnormal sialylation is related to a variety of diseases, including cancer, pathogen infection and cardiovascular disease (CVD). Therefore, this study aims to explore the role of sialylation related genes in AS. Methods: Two AS data sets were obtained from the gene expression comprehensive database (GEO). Based on the differentially expressed genes (DEGs) and the sialylation gene set, the differentially expressed sialylation-related genes (De-SRGs) were found. Then, machine learning method is used to find the core gene. The immune cell infiltration method was established to study the immune cell imbalance in AS. Subsequently, we explored two different subtypes based on core genes using 158 AS samples. Gene ontology (GO), Kyoto Encyclopedia of Genes and Genome (KEGG) pathway enrichment, gene set variation analysis (GSVA) and immunoinfiltration analysis are also used to evaluate the different roles of subtypes. Results: A total of 36 De-SRGs were identified. Through machine learning algorithm, 5 core genes were identified and 2 subtypes related to core genes were defined. The results of GSVA showed that type A inflammatory response related pathways were significantly enriched, while type B inflammatory response related pathways were significantly enriched. Conclusion: Through this study, we have revealed the relationship between Sialylation-related genes and AS, as well as the heterogeneity of AS patients with different Sialylation subtypes. Selecting a Sialylation-signature based on five genes as the best machine learning model can accurately evaluate the diagnosis of AS and control patients. Our research results reveal the progress of Sialylation in AS for the first time, and provide new insights for its potential pathogenesis and potential treatment strategies.