Cell‐type specific variations of within pathway interaction in AD using covariance regression. (31st December 2021)

• Record Type: Journal Article
• Authors: Lu, Xiaoyu; Zhang, Chi; Zhao, Yi; Cao, Sha
• Is Part Of: Alzheimer's & dementia. Volume 17(2021)Supplement 5
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-12-31
• 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.057873
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Alzheimer's disease (AD) and related dementias are among the greatest public health challenges in the United States. A deeper understanding of the disease pathology on a molecular and cellular level is critical to devise treatments to slow down the disease progression and better understand the mechanisms behind AD. Recent advancement in single‐cell technology provides new avenues for molecular profiling of AD at the single‐cell resolution, which improves the omics studies by making greater precision and granularity possible. Method: We here present a statistical framework for extracting the variability of gene‐gene interaction in a disease‐, sex‐ and cell type‐ specific manner, to identify AD‐associated markers and pathways and to articulate disease mechanisms on cellular and molecular levels. We applied our method to the Religious Orders Study and Memory and Aging Project (ROS/MAP) single nucleus dataset, with single cells of five cell types including astrocyte, microglial, neuron, oligodendrocyte, and oligodendrocyte progenitor cell (OPC). Result: Existing methods for AD gene and biomarker detection mainly focus on comparative and univariate analysis, with the lack of adequate statistical power in detecting abnormal gene‐gene interaction structure in the same pathway. Our covariance regression based analytical pipeline represents a statistically solid tool to investigate the abnormalities in genetic interaction on pathway‐level using single‐cell RNA sequencing data, and to make inference on key covariates. Importantly, such a pathway‐level analysis is robust to the severe drop out events in single cell RNA‐Seq data. Compared with enrichment‐based pathway analysis, our method has much higher statistical power in detecting significant differentially co‐expressed pathways in comparing AD vs healthy control cells in a sex and cell type specific manner. Conclusion: Majority of the pathways tend to have lower co‐expression strength among its genes in AD cells than healthy cells for female subjects in astrocyte, neuron and microglial. While not much pathways are differentially co‐expressed in male AD vs healthy subjects, neuron and astrocyte are the two major cell types with the largest number of alternatively co‐expressed pathways in female subjects, with 3972 and 4218 pathways. These altered pathways belong to known pathway categories, such as synapse, neuronal development, lipid metabolism, calcium transport, etc.