Construction of a New Protein–Protein Interaction and Molecular Biomarkers Networks in Alzheimer’s Disease Patients by Bioinformatics Screening

Abstract

This study aimed to construct the molecular biomarkers of autophagy and endoplasmic reticulum stress (ERS), as well as their corresponding protein interaction network in Alzheimer’s disease (AD) patients with different levels of physical activity (PA) by bioinformatics methods. The expression profiles of the genes were downloaded from Gene Expression Omnibus (GEO). Differentially expressed genes (DEGs) between AD samples with low, moderate and high levels of PA were studied. The autophagy and ERS-related genes (AERSRGs) were extracted from GeneCards and MsigDB databases. The functional enrichment analysis was conducted to determine the function of DEGs. To explore the proteins, miRNAs and transcription factors (TF) interacting with DEGs, the protein–protein network, mRNA-miRNA network and mRNA-TF network were built using Cytoscape software. Then the receiver operating characteristic (ROC) analysis were conducted to verify the diagnostic performance of hub genes. A total of 533 AERSRGs were identified in Group H and 150 AERSRGs were screened in Group M. Functional enrichment analysis suggested genes of AD play vital roles in some biological process (e.g., cell cycle phase transition, mitochondrion organization, proteasomal protein catabolic process). KEGG enrichment analysis suggested that sarcopenia involves the pathways (e.g., GABA, P2Y receptors, serotonin release cycle). A total of 5 hub genes were screened in Group H and 9 were identified in Group M. ROC analysis suggested that several hub genes exhibited a relatively high sensitivity and specificity in both groups of AD. The hub genes screened in this study are closely correlated with autophagy and ERS in AD and can differentiate AD with different levels of PA. SRC, MAPK3 and MAP2K1 exhibit relatively high sensitivity and specificity in diagnosis in Group H; MCM2, CDC42, HNRNPM, ASF1A, NCBP2, SNRNP70 and MCM6 exhibit relatively high sensitivity and specificity in diagnosis in Group M.