Integrating Multiple Omics to Illuminate Gene Networks Underlying Cigarette Smoking and Opioids.

项目总结/摘要 这项研究的目标是发现神经生物学上可解释的成瘾基因网络， 特别是香烟和阿片类药物为了实现这一目标，我们将应用一种新的多组学、多方法 框架-基因网络识别和整合（GNetII）-识别相关的基因网络 独特的香烟或阿片类药物的结果和网络共享这些成瘾。GNetII包括 全基因组上位性、可解释人工智能、基因网络构建和证据线 方法.这些基础方法将使大规模全基因组关联研究的整合成为可能 （GWAS）数据，死后人脑数据（RNA测序，DNA甲基化， 染色质免疫沉淀测序和变异基因型） 死者（已故的个人）和公共组学数据。吸烟和阿片类药物的结果是 基因相关，我们有两个高度相关的平行GWAS和多组学大脑数据 背外侧前额叶皮层和背外侧前额叶核，这两种成瘾。 在美国，香烟和阿片类药物是可预防的发病率和死亡率的主要原因。这些 成瘾影响数百万美国成年人和青少年并且是高度遗传的（例如，阿片类药物为54%和71% 成瘾和尼古丁依赖）。GWAS分析已经在全基因组范围内确定了300多个位点， 吸烟的意义。阿片类药物的GWAS在样本量上远远落后，但全基因组显著 正在出现。已知基因座的神经生物学效应在很大程度上是未知的，更多的基因座和连接 仍有待发现。我们假设，应用新的大数据科学方法， 脑组织中大规模的GWAS和基因调控数据将揭示以前未发现的关系 (such作为基因间的上位相互作用），并增加了成瘾背后的神经生物学知识。 我们提出以下具体目标： 目的1：整合多组学数据，发现与癌症相关的基因网络。 目的2：整合多组学数据，发现阿片相关基因网络。 目的3：整合多组学数据，寻找一般成瘾相关基因网络。 对于香烟，目标1和3将利用GWAS（N= 528，259）和多组学数据在死后人体中进行研究。 来自活跃吸烟者和非吸烟者死亡者的脑（N=262）。对于阿片类药物，目标2和3将使用GWAS （N= 49，178）和来自阿片类药物过量病例和对照死亡者（N=147）的多组学脑数据。分析 将在世界上最快的超级计算机Summit上执行，这将大大提高 神经生物学意义的发现我们的研究将捕捉整个基因组的复杂网络， 以前未知的基因，以及帮助解释神经生物学基础的越来越多的 与香烟或阿片类药物相关的基因位点。

PROJECT SUMMARY/ABSTRACT The goal of the proposed research is to discover neurobiologically interpretable gene networks for addiction, specifically for cigarettes and opioids. To achieve this goal, we will apply a new multi-omics, multi-method framework—Gene Network Identification and Integration (GNetII)—to identify gene networks associated uniquely with cigarette or opioid outcomes and networks shared across these addictions. GNetII includes genome-wide epistasis, Explainable Artificial Intelligence, gene network construction, and Lines-of-Evidence methods. These cornerstone methods will enable integration of large-scale genome-wide association study (GWAS) data in living subjects, postmortem human brain data (RNA-sequencing, DNA methylation, chromatin immunoprecipitation sequencing, and variant genotypes) from addiction case and control decedents (deceased individuals), and public omics data. Cigarette smoking and opioid outcomes are genetically correlated, and we have parallel GWAS and multi-omics brain data available in two highly relevant tissues, dorsolateral prefrontal cortex and nucleus accumbens, for both of these addictions. Cigarettes and opioids are leading causes of preventable morbidity and mortality in the United States. These addictions affect millions of U.S. adults and youths and are highly heritable (e.g., 54% and 71% for opioid addiction and nicotine dependence, respectively). GWAS analyses have identified 300+ loci at genome-wide significance for smoking. GWAS for opioids are farther behind in sample size, but genome-wide significant loci are emerging. Neurobiological effects of known loci are largely unknown, and more loci and connections among the loci are still to be discovered. We hypothesize that applying new big data science methods to large-scale GWAS and gene regulation data in brain tissue will reveal previously undetected relationships (such as epistatic interactions between genes) and add knowledge of the neurobiology underlying addiction. We propose the following specific aims: Aim 1: Integrate multi-omics data to discover cigarette-associated gene networks. Aim 2: Integrate multi-omics data to discover opioid-associated gene networks. Aim 3: Integrate multi-omics data to find general addiction-associated gene networks. For cigarettes, Aims 1 and 3 will leverage GWAS (N=528,259) and multi-omics data in postmortem human brain from active smoker and nonsmoker decedents (N=262). For opioids, Aims 2 and 3 will use GWAS (N=49,178) and multi-omics brain data from opioid overdose case and control decedents (N=147). Analyses will be performed on Summit, the world's fastest supercomputer, which will greatly improve the likelihood of neurobiologically meaningful discoveries. Our study will capture complex networks across the genome to find previously unknown genes, as well as help explain the neurobiological underpinnings for the growing number of genetic loci associated with cigarette or opioid outcomes.