Data alignment to model Alzheimer's disease (AD). (December 2021)

• Record Type: Journal Article
• Authors: Olayan, Rawan; Uyar, Asli; Onos, Kristen; Preuss, Christoph; Howell, Gareth R; Carter, Gregory W
• Is Part Of: Alzheimer's & dementia. Volume 17(2021)Supplement 3
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-12
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.051753
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Alzheimer's disease (AD) is a heterogenous disease that is associated with complex brain changes and major alterations in gene expression. The analysis of correlated transcriptomic variation can help in understanding the biological mechanisms underlying the disease. Having a comprehensive atlas of changes in human brain transcriptome, enables the application of data alignment to other models like mouse to validate, assess, and compare such changes. Identifying the optimal mouse model can further advance the discovery of the early disease characteristics and development of potential drugs. Method: Our methodology starts by considering distinct human late‐onset AD (LOAD) subtypes based on case heterogeneity, each is examining a specific group of molecular profiles. Moreover, we performed comprehensive transcriptome profiling of brains from a panel of early‐onset AD (EOAD) mouse models at 8 months of age, including six distinct genetic backgrounds carrying the same APP/PS1 transgenic construct. Next, we generate informative features that determine the specific characteristic in each population of samples in mouse and human. Then, we identify and measure the overall relationships between human and mouse via the dimension reduction technique canonical correlation analysis. By cross‐aligning samples, we are able to identify which mouse model shows correlated changes with which human subtype. Finally, we identify and rank the major translatable altered pathways in which the mouse model effects mimic disease effects in human subtypes. Result: The selected set of pathways represents AD‐related pathway modifications that varies across mouse models. Moreover, based on the gene overlap, this ranked set of pathways reveals cell‐type‐specific pathways associated with LOAD pathogenesis in the brain. Our results suggested that cross‐species data alignment reveals similar samples between human and mouse populations. Conclusion: We identify the potential mouse models, with the potential AD‐related pathways that have similar genetic characteristics to human subtypes. We believe that using diverse mouse models to model AD in human can better recapitulate the molecular changes during the origin and progression of LOAD.