A novel bioinformatics approach to reveal the role of circadian oscillations in AD development

Abstract

Altered circadian gene expression may contribute to Alzheimers disease (AD) progression. Unfortunately, sampling the central nervous system (CNS) at multiple time points is not feasible. Moreover, there are no AD-related time-series transcriptome datasets available for studying these circadian patterns and their impacts on AD development. In this study, we introduce a novel computational platform, Event-driven Sample Ordering for Circadian Variation Detection (ESOCVD), to reveal rhythmic patterns of gene expression of AD using untimed transcriptome datasets. ESOCVD was applied to 20 untimed gene expression profiles of 16 brain regions from approximately 3000 AD patients in public transcriptome databases. Our analysis revealed five types of circadian alteration patterns in ~2,000 circadian genes in different brain regions of AD patients. Further analyses of additional databases confirmed that our analytical platform can be applied to identify the evolutionary dynamics of circadian variation during the process of AD development. Through the gene expression correlation analysis for our 8 circadian genes identified from AMP-AD MSBB cohorts, we identified stage-specifically enriched biological processes with anticipated context. Gene expression analysis of AD mouse brain tissues further substantiated the predictions of the ESOCVD model. In summary, ESOCVD is highly versatile in bridging circadian research and precision medicine.