Profiling the Functional Genetics of Health and Disease using BrainGENIE: The Brain Gene Expression and Network Imputation Engine

摘要 大脑疾病的病理生理学仍然未知，因为我们不能在活体中研究相关组织。 人类实验对象死后的脑组织是有用的，但昂贵，罕见，严重混淆， 死前和濒死因素这两个事实激发了对替代策略的探索，例如 替代标记的使用，如血液基因表达。为了提高这个广泛使用的 在这种方法中，我们开发了一种新的计算方法，称为大脑基因表达和网络 利用血液和大脑基因之间的生物可比性的插补引擎（BrainGENIE） 基于血液基因表达谱预测脑组织的转录组谱。 BrainGENIE从根本上不同于其他转录组估算方法， 在大脑转录组中变异的比例和比例更大。BrainGENIE能够 预测大约9-57%的大脑转录组，这产生了大约1.8倍的增加， 覆盖率相对于“金标准”方法PrediXcan，这大大提高了我们的统计能力， 检测与疾病相关的基因和途径。我们的建议包括三个具体目标，以改善我们的 方法和揭示神经精神疾病的生物学途径。目标1：优化 BrainGENIE来捕获我们的方法目前无法很好预测的其他基因。目标2：应用 BrainGENIE将我们收集的公开和内部数据用于预测大脑区域特异性基因 超过8,000个活着的人的表达谱，并发现特定区域的基因表达模式 与神经精神障碍和神经退行性疾病相关。目标3：传播BrainGENIE 作为独立软件供其他研究人员免费使用。在最近的遗传学和基因组研究的指导下， 假设在神经精神疾病中发现类似的基因失调模式 在涉及先天免疫、染色质重塑、神经发育和神经传递的途径中。 在我们的分析中纳入神经退行性疾病将使我们能够确定基因表达是否 在更广泛的脑部疾病中也有相同的模式。我们还希望识别出特定的疾病， 脑区特异性转录组学关联。我们的项目将使生物学的新的研究方向成为可能。 这些变化出现在活人的大脑中，并创造了改善诊断的机会， 干预和治疗。

Abstract The pathophysiology of brain disorders remains unknown because we cannot study the relevant tissue in living human subjects. Postmortem brain tissue is useful, but expensive, rare, and critically confounded by antemortem and agonal factors. These two facts have inspired the search for alternative strategies, such as the use of surrogate markers like blood-based gene expression. To improve the rigor of this widely used approach, we developed a novel computational method called the Brain Gene Expression and Network Imputation Engine (BrainGENIE) that leverages biological comparability between blood and brain gene expression to predict transcriptome profiles for brain tissue based on blood gene-expression profiles. BrainGENIE is fundamentally different from other transcriptome-imputation methods, and captures a much larger proportion of the variance in—and larger fraction of—the brain transcriptome. BrainGENIE is capable of predicting approximately 9–57% of the brain transcriptome, which yields an approximate 1.8-fold increase in coverage relative to the “gold standard” method PrediXcan, and which greatly improves our statistical power to detect genes and pathways associated with disease. Our proposal contains three Specific Aims to improve our method and shed light on biological pathways underlying neuropsychiatric disorders. Aim 1: Refine BrainGENIE to capture additional genes that are not currently well predicted by our method. Aim 2: Apply BrainGENIE to our collection of publicly available and in-house data to predict brain-region-specific gene expression profiles for over 8,000 living persons, and discover region-specific gene-expression patterns associated with neuropsychiatric disorders and neurodegenerative diseases. Aim 3: Disseminate BrainGENIE as stand-alone software for other researchers to use freely. Guided by recent genetic and genomic studies, we hypothesize that comparable patterns of gene dysregulation will be found across neuropsychiatric disorders among pathways involving innate immunity, chromatin remodeling, neurodevelopment, and neurotransmission. Inclusion of neurodegenerative disorders in our analysis will allow us to determine whether gene expression patterns are shared across a broader range of brain disorders. We also expect to identify disorder-specific and brain-region-specific transcriptomic associations. Our project will enable new lines of inquiry into biological changes that emerge in the brains of living persons, and create opportunities to improve diagnostics, intervention, and treatment.