Measuring Depression: Using biomarkers to investigate the biology of depression

摘要 抑郁症是世界范围内导致残疾的主要原因，每6个人中就有1人受到影响。尽管背负着全球负担， 抑郁症的生物学仍然知之甚少。实验室检测为医生提供了有针对性的 生化测量，生物标记物，帮助诊断和治疗各种疾病的患者。 存储在电子健康记录(EHR)中的生物标记物结果在很大程度上是一种尚未开发的研究资源。 以前的流行病学研究确定了抑郁状态与各种生物标记物之间的关联， 最值得注意的是免疫标记。然而，关联的方向和潜在的生物学之间 抑郁症和免疫系统还没有被描述。我们假设，将遗传学与 基于EHR的生物标记物和中介数据将有助于了解抑郁症发生的生物过程 抑郁症。在之前的分析中，我们创建了一个实验室范围的关联研究(Labwas)框架作为 扫描多基因评分(PGS)和生物标记物之间关联的假设生成方法 存储在电子病历中。我们的方法允许以前所未有的规模对生物标记物进行调查， 分析中包括的个体数量和生物标志物的数量，使我们有机会 复制以前的生物标记物关联，并识别新的关联。我们发现了一个关联 在抑郁PGS和免疫标志物白细胞计数(WBC)增加之间 跨越多个生物库。我们计划在整个提案过程中进一步调查这种关系。我们计划 使用两个目标来验证我们的假设：目标1将评估一个表型或遗传因素 解释了抑郁症遗传学和白细胞之间的联系。抑郁症的诊断通常是并存的 以及其他已知对白细胞结果有影响的医疗条件。我们将首先进行敏感性分析 在抑郁症PGS和WBC之间的分析中，通过协变可能会混淆诊断。我们会 然后执行条件分析以解析关联的方向和潜在的遗传区域驱动 协会。目标2将描述抑郁症遗传学在调节关系中的作用 WBC在抑郁症诊断和抗抑郁药物使用之间的关系。抗抑郁药物治疗 此前，抗抑郁药与循环生物标记物的变化有关。通过利用 在EHR中的药物信息，我们计划检查抗抑郁药对免疫生物标记物水平的影响， 确定抑郁PGS的调节作用，并确定免疫生物标记物水平是否与 治疗反应。这个项目的成功完成将是第一个分析抑郁症影响的项目 在生物标志物的尺度上进行遗传学研究，解析关联方向并识别基因 在抑郁症和免疫系统之间的中介，并调查遗传学对 来自抑郁症治疗的免疫系统。这些结果的未来影响可能包括发展 生物知情的治疗选择、诊断面板和表型亚型。

Abstract Depression is a leading cause of disability worldwide, affecting 1 in 6 individuals. Despite the global burden, biology of depression remains poorly understood. Laboratory testing provides physicians with targeted biochemical measurements, biomarkers, to aid in diagnosing and treating patients for a variety of diseases. Biomarker results stored in electronic health records (EHRs) are a largely untapped resource for research. Previous epidemiologic studies identified associations between depression status and various biomarkers, most notably immune markers. However, the direction of association and underlying biology between depression and the immune system has not been described. We hypothesize that integrating genetics of depression with EHR-based biomarker and mediation data will help inform biological processes occurring in depression. In previous analyses, we created a lab-wide association study (LabWAS) framework as a hypothesis-generating approach to scan for associations between polygenic scores (PGS) and biomarkers stored in EHRs. Our method allows for an investigation of biomarkers at an unprecedented scale with both the number of individuals and the number of biomarkers included in the analyses, giving us the opportunity to replicate previous biomarker associations as well as identify novel ones. We discovered an association between depression PGS and an increased immune marker, white blood cell count (WBC) which replicated across multiple biobanks. We plan to further investigate this relationship throughout the proposal. We plan to investigate our hypothesis using two aims: Aim 1 will evaluate whether a phenotype or genetic factors explains the association between depression genetics and WBC. Depression diagnosis is often comorbid with other medical conditions that have known effects on WBC results. We will first conduct sensitivity analyses by covarying for potentially confounding diagnoses in the analysis between depression PGS and WBC. We will then perform conditional analyses to parse the direction of association and underlying genetic regions driving the association. Aim 2 will characterize the role of depression genetics in moderating the relationships between depression diagnosis and antidepressant usage with WBC. Antidepressant treatment with antidepressants has previously been associated with changes in circulating biomarkers. By leveraging medication information in EHRs, we plan to examine the effect of antidepressants on immune biomarker levels, determine the moderating role of depression PGS, and determine if immune biomarker levels associate with treatment response. Successful completion of this project would be the first to analyze the effect of depression genetics on the landscape of biomarkers at scale, parse the direction of association and identify genetic mediators between depression and the immune system, and investigate the effects of genetics on changes in immune system from depression treatment. The future impact of these results could include the development of biologically-informed treatment options, diagnostic panels, and phenotypic subtyping.