Identification of plasma cell subsets and molecular markers of Alzheimer's disease based on single-cell weighted gene co-expression network analysis and Mendelian randomization analysis

<title>Abstract</title> Alzheimer’s disease (AD) is a neurodegenerative disorder with a multifactorial pathogenesis, comprising gene expression alterations and abnormal immune cell infiltration. In this study, we aimed at further exploring AD pathogenesis and identifying potential therapeutic targets. We downloaded GSE181279 dataset-derived single-cell data from the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database, then processed and analyzed it using various bioinformatic methods. We screened, standardized, homogenized, and processed the data using principal component analysis and harmony. We identified subgroups using t-distributed stochastic neighbor embedding analysis. For the single-cell data, we performed cell-type annotation and marker analysis, and compared cell proportions between the groups. After confirming the plasma cell subtype, we screened the disease-associated gene modules via high-dimensional weighted gene co-expression network analysis and identified key genes using pathway, Mendelian randomization (MR), sensitivity, and immune cell infiltration analyses. Finally, we analyzed the transcription factor regulatory network and the correlation between key genes, identifying 21 subgroups. The plasma cell subtype proportion significantly increased in patients with AD. We identified brown- and blue-module disease-associated genes. Further pathway and MR analyses identified four key genes (<italic>COX4I1</italic>, <italic>MAL</italic>, <italic>RGS1</italic>, and <italic>RPS5</italic>) and confirmed their association with AD. Immune cell infiltration analysis revealed that the key genes are closely related to immune cells, underlining their potentially important roles in AD-related immune regulation. This study explored various AD pathogenesis-related aspects and identified disease-associated key genes and biological pathways. Our results provide important clues for upcoming AD pathophysiological mechanism-related studies and AD therapeutic target assessments.