跨祖先影像衍生表型多效性分析提升抑郁风险分层
该研究整合欧洲与东亚人群的抑郁和影像衍生表型(IDP)大规模GWAS汇总统计,通过MTAG、孟德尔随机化、聚类和机器学习框架,识别多效性变异与六类生物学通路,并构建抑郁风险分层模型(N=16,166)。
要点
- 抑郁与IDP存在遗传相关和多效性SNV,涉及代谢、神经递质和神经免疫通路。
- 孟德尔随机化在EUR和EAS中分别建立抑郁与IDP的因果关联。
- 基于独立SNV的无监督聚类识别六个具有组织/细胞类型特异性的生物学通路。
核心结论
- 研究推进了对抑郁多尺度病因的理解,并提供动态风险分层以支持精准预防。
研究设计与样本
- 收集欧洲(EUR)抑郁GWAS N=1,293,933和IDP N=33,224,东亚(EAS)抑郁N=82,874和IDP N=7,058的汇总统计。
- 采用MTAG分析抑郁与IDP的多效性,TwoSampleMR进行孟德尔随机化,MAGMA评估组织/细胞类型特异性。
- 从多效性分析中选择独立SNV进行无监督聚类,并用机器学习框架(N=16,166)细化风险分层。
主要发现
- MTAG clarify了抑郁与IDP的遗传相关,识别了多效性SNV和基因并赋予功能见解。
- 孟德尔随机化在EUR和EAS祖先中分别建立了抑郁与IDP的因果关联。
- 六个聚类 delineate 了代谢调节、神经递质动态和神经免疫交互的 distinct 生物学通路。
- 机器学习框架 dissect 了抑郁与多基因风险评分、IDP及可改变生活方式因素的关系。
临床怎么用
- 可利用多基因风险评分、IDP和生活方式因素进行抑郁风险分层,辅助精准预防。
- 识别出的多效性通路可能为干预提供潜在靶点,但需进一步验证。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
Depression arises from dynamic interactions among genetic predisposition, brain alterations, and environmental stressors. Despite genome-wide association studies (GWAS) identifying risk loci, the mechanisms translating genetic variation into brain changes remain elusive. Imaging-derived phenotypes (IDPs) were the intermediate traits linking genetic architecture to neural circuit dysfunction. Here, we collected large-scale GWAS summary statistics of depression and IDPs across European (EUR; N = 1,293,933 and 33,224, respectively) and East Asian (EAS; N = 82,874 and 7058, respectively). In the multiple-trait analysis between depression and IDPs, we clarified their genetic correlation through MTAG, identified the pleiotropic single nucleotide variants (SNVs) and genes with functional insight, and established the causal relationship through Mendelian randomization via TwoSampleMR in EUR and EAS ancestry, respectively. To discern the heterogeneous genetic drivers, we selected independent SNVs from the multiple-trait analyses to perform unsupervised clustering. Six clusters delineated distinct biological pathways for metabolic regulation, neurotransmitter dynamics, and neuroimmune interactions, with tissue/cell type specificity through MAGMA. Finally, we dissected relationships between depression and polygenic risk score, IDPs, and modifiable lifestyle factors, and introduced a machine learning framework to refine risk stratification (N = 16,166). Our study advanced the understanding of the multiscale etiology of depression while providing dynamic depression risk stratification for precision prevention.
引用
- Feng Y, Guo X, Huang P, et al. Cross-ancestry pleiotropic analysis of imaging-derived phenotypes enhances risk stratification of depression. Mol Psychiatry. 2026. PMID: 42387104. doi:10.1038/s41380-026-03730-0
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。