Machine learning and bioinformatics to identify 8 autophagy-related biomarkers and construct gene regulatory networks in dilated cardiomyopathy

Abstract

Dilated cardiomyopathy (DCM) is a condition of impaired ventricular remodeling and systolic diastole that is often complicated by arrhythmias and heart failure with a poor prognosis. This study attempted to identify autophagy-related genes (ARGs) with diagnostic biomarkers of DCM using machine learning and bioinformatics approaches. Differential analysis of whole gene microarray data of DCM from the Gene Expression Omnibus (GEO) database was performed using the NetworkAnalyst 3.0 platform. Differentially expressed genes (DEGs) matching (|log2FoldChange ≥ 0.8, p value < 0.05|) were obtained in the GSE4172 dataset by merging ARGs from the autophagy gene libraries, HADb and HAMdb, to obtain autophagy-related differentially expressed genes (AR-DEGs) in DCM. The correlation analysis of AR-DEGs and their visualization were performed using R language. Gene Ontology (GO) enrichment analysis and combined multi-database pathway analysis were served by the Enrichr online enrichment analysis platform. We used machine learning to screen the diagnostic biomarkers of DCM. The transcription factors gene regulatory network was constructed by the JASPAR database of the NetworkAnalyst 3.0 platform. We also used the drug Signatures database (DSigDB) drug database of the Enrichr platform to screen the gene target drugs for DCM. Finally, we used the DisGeNET database to analyze the comorbidities associated with DCM. In the present study, we identified 23 AR-DEGs of DCM. Eight (PLEKHF1, HSPG2, HSF1, TRIM65, DICER1, VDAC1, BAD, TFEB) molecular markers of DCM were obtained by two machine learning algorithms. Transcription factors gene regulatory network was established. Finally, 10 gene-targeted drugs and complications for DCM were identified.

Diagnostic and prognostic value of autophagy-related key genes in sepsis and potential correlation with immune cell signatures

“…Through differential expression analysis and WGCNA, we obtained 14 autophagy-related key genes. Functional enrichment analysis further confirmed that these Frontiers in Cell and Developmental Biology frontiersin.org nephropathy, pulmonary hypertension, dermatomyositis, and rheumatoid arthritis (Du et al, 2020;Tong et al, 2021;Bai et al, 2022;Fan et al, 2022;Wang et al, 2022;Yang et al, 2022;Zhang et al, 2022). Therefore, ROC curve and survival analysis were performed on key genes to explore their diagnostic efficacy and prognostic value in sepsis.…”

“…ARGs have been shown to be of great value in the diagnosis and/or prognostic evaluation of diseases such as cancer, dilated cardiomyopathy, acute myocardial infarction, diabetic nephropathy, pulmonary hypertension, dermatomyositis, and rheumatoid arthritis ( Du et al, 2020 ; Tong et al, 2021 ; Bai et al, 2022 ; Fan et al, 2022 ; Wang et al, 2022 ; Yang et al, 2022 ; Zhang et al, 2022 ). Therefore, ROC curve and survival analysis were performed on key genes to explore their diagnostic efficacy and prognostic value in sepsis.…”

New Insight Into Mechanisms of Hepatic Encephalopathy: An Integrative Analysis Approach to Identify Molecular Markers and Therapeutic Targets

“… 18 , 19 Enrichr is a Web-based platform for analyzing candidate gene sets to predict common annotated biological features (ie, pathways, cellular type, phenotypes and transcription factors (TFs)). 20 For this purpose, 2 significant gene sets from DEGs (upregulated and downregulated) were separately submitted in Enrichr and proceeded through WikiPathway Web tool. Furthermore, to identify and compare the phenotype properties of each gene set, phenotype ontology was performed using Enrichr (Human Phenotype Ontology tool).…”

The role of autophagy in cardiovascular disease: Cross-interference of signaling pathways and underlying therapeutic targets

“…Based on Lasso Cox regression algorithm, eight ATGs (BAD, DICER1 HSF1, HSPG2, PLEKHF1, TRIM65, TFEB, VDAC1) were fond to be closely related to the overall survival (OS) rate of DCM. They may be underlying markers for predicting the danger signals of organic changes in heart disease ( 69 ).…”