Leveraging large datasets to determine genomic impact of transposable elements in Alzheimer’s Disease and related dementias

ABSTRACT Alzheimer's disease (AD) is a complex neurodegenerative disorder with genetic and environmental factors contributing to its development. Transposable elements (TEs) have been identified as prevalent in individual neurons within the cerebral cortex, a region closely associated with AD progression. Recent research has indicated a connection between AD-associated Tau protein and TE activity. A closer investigation is highly demanded to explore the prevalence and relevance of TEs in AD and related dementias (ADRD), by overcoming the repetitive nature of TEs and the constraint of small sample size in human studies. This proposal aims to investigate the landscape of TEs in large cohorts and their association with AD and ADRD, further revealing the functional impact of TE on AD and AD-related phenotypes. To achieve this, the project will leverage the extensive datasets from large AD projects, which encompass genomic sequences and detailed clinical information. A novel cloud-based infrastructure will be developed to overcome computational challenges and complexities involved in the analysis of such large-scale datasets on the cloud server. The study will identify patterns of diagnostic nucleotides within TEs using an adapted pipeline from large genomic consortia and reveal the potential functional impact of AD-relevant TEs in the insertion loci. Furthermore, TE-derived expression profiles or TE transcript abundance in different brain regions will be assessed using a tailored pipeline from RNA-seq data. Combining the expression profiles from TEs and gene regions affected by TE insertions with clinical and neuropathological metrics, we will develop a predictive model for dementia diagnoses. The project's deliveries and innovative aspects include the landscape of TEs and their diagnostic nucleotide patterns in large AD and AD-related cohorts, a novel and generalized cloud-based pipeline designed to efficiently package genetic variant calling process in on-cloud large datasets, and a predictive model integrating TE-derived expression profiles and TE diagnostic nucleotides with detailed clinical information for AD diagnosis. Through this project, we set the stage for a deeper understanding of the genetic etiology of AD and ADRD and may open avenues for future studies on novel therapeutic strategies and improved patient care.