journal articles
MULTI-OMICS ANALYSIS OF DRUGGABLE GENES TO FACILITATE ALZHEIMER\'S DISEASE THERAPY: A MULTI-COHORT MACHINE LEARNING STUDY
Jichang Hu, Yong Luo, Xiaochuan Wang
BACKGROUND: The swift rise in the prevalence of Alzheimer's disease (AD) alongside its significant societal and economic impact has created a pressing demand for effective interventions and treatments. However, there are no available treatments that can modify the progression of the disease.
METHODS: Eight AD brain tissues datasets and three blood datasets were obtained. Consensus clustering was utilized as a method to discern the various subtypes of AD. Then, module genes were screened using weighted correlation network analysis (WGCNA). Furthermore, screening hub genes was conducted through machine-learning analyses. Finally, A comprehensive analysis using a systematic approach to druggable genome-wide Mendelian randomization (MR) was conducted.
RESULTS: Two AD subclasses were identified, namely cluster.A and cluster.B. The levels of gamma secretase activity, beta secretase activity, and amyloid-beta 42 were found to be significantly elevated in patients classified within cluster A when compared to those in cluster B. Furthermore, by utilizing the differentially expressed genes shared among these clusters, along with identifying druggable genes and applying WGCNA to these subtypes, we were able to develop a scoring system referred to as DG.score. This scoring system has demonstrated remarkable predictive capability for AD when evaluated against multiple datasets. Besides, A total of 30 distinct genes that may serve as potential drug targets for AD were identified across at least one of the datasets investigated, whether derived from brain samples or blood analyses. Among the identified genes, three specific candidates that are considered druggable (LIMK2, MAPK8, and NDUFV2) demonstrated significant expression levels in both blood and brain tissues. Furthermore, our research also revealed a potential association between the levels of LIMK2 and concentrations of CSF Aβ (OR 1.526 (1.155–2.018)), CSF p-tau (OR 1.106 (1.024–01.196)), and hippocampal size (OR 0.831 (0.702–0.948)).
CONCLUSIONS: This study provides a notable advancement to the existing literature by offering genetic evidence that underscores the potential therapeutic advantages of focusing on the druggable gene LIMK2 in the treatment of AD. This insight not only contributes to our understanding of AD but also guides future drug discovery efforts.
CITATION:
Jichang Hu ; Yong Luo ; Xiaochuan Wang (2025): Multi-omics analysis of druggable genes to facilitate Alzheimer's disease therapy: A multi-cohort machine learning study. The Journal of Prevention of Alzheimer’s Disease (JPAD). https://doi.org/10.1016/j.tjpad.2025.100128