RAD-Blood: a database to identify RNAs in blood as potential biomarkers of Alzheimer's disease by integrating bulk- and single-cell transcriptome profile

Abstract

Background The early diagnosis of Alzheimer's disease (AD) in large-scale high-risk population is a major challenge. Blood-based biomarkers could enable widespread testing for AD. RNA-seq technology is becoming an effective method in investigating diagnostic biomarkers for diseases, but platforms exploring RNA-seq data in AD blood are lacking. Methods We collected the raw RNA-seq data in the blood of AD patients or AD mouse models, mild cognitive impairment (MCI) patients, and normal people or wild type mouse from the Gene Expression Omnibus (GEO) and Synapse databases. And the RNA-seq data was analyzed by the standard pipeline. We applied R-Shiny to develop the website of RAD-Blood (RNA-seq analysis of AD blood, http://www.bioinform.cn/RAD-Blood/) to present the plentiful analysis results. Results RAD-Blood was specifically designed to analyze existing blood RNA-seq data sets (mRNA-seq, miRNA-seq, and scRNA-seq) from patients and mouse models with AD pathology. The RAD-Blood provides differential expression, gene set enrichment, immune abundance and its correlation with gene expression, cell type annotation, T cell receptor, and cell communication analyses for RNA-seq data in AD/MCI/normal blood, with rich results forms and colorful figures. We used a case study to show the capacity of RAD-Blood in finding blood biomarkers in AD/MCI blood. By using RAD-Blood, we found 274 protein-coding genes whose mRNA expression was consistently up-regulated or down-regulated from normal to MCI to AD. Among the consistently down-regulated genes, four are the markers of the blood erythroid cell. Compared with normal people, the population of erythroid cells in AD patients decreased. Despite the reduction in cell count, interactions between blood erythroid cells with other cells increased dramatically, which is mainly mediated through the major histocompatibility complex I (MHC-I) signaling pathway. These findings have not been reported by existing studies, which suggests that RAD-Blood is a solution for finding potential novel signatures in the blood of AD and MCI. Conclusions RAD-Blood is a user-friendly web server for multi-level analysis and visualization of gene/miRNA expression and immune profile in AD blood, and allows broad utility in exploring potential AD blood biomarkers, testing hypotheses, guiding experiment design, and investigating the peripheral pathogenic mechanisms and proposing potential early diagnosis standard of AD.