Data Science Research

数据科学研究

• 批准号: 9108711
• 负责人: PAUL M THOMPSON
• 金额: $ 179.22万
• 依托单位: UNIVERSITY OF SOUTHERN CALIFORNIA
• 依托单位国家: 美国
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-06-01 至 2018-09-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/9108711
• 关键词: AIDS/HIV problem; Affect; Algorithms; Attention deficit hyperactivity disorder; Autistic Disorder; Big Data; Biological Markers; Bipolar Disorder; Brain; Brain Diseases; Brain Mapping; Brain imaging; Clinical; Collaborations; Combinatorics; Computer software; Computers; Country; Data; Data Science; Data Set; Diagnosis; Diagnostic and Statistical Manual of Mental Disorders; Diffusion; Disease; Economics; Genes; Genetic; Genetic Markers; Genetic screening method; Genome; Genomics; Goals; HIV; Hippocampus (Brain); Image; Institution; Lead; Learning; Link; MRI Scans; Machine Learning; Major Depressive Disorder; Maps; Mathematics; Measures; Medicine; Mental Depression; Meta-Analysis; Mining; Modeling; Movement; Multimodal Imaging; Pattern; Pharmaceutical Preparations; Research; Resources; Savings; Schizophrenia; Science; Site; Testing; Variant; Work; addiction; biomarker discovery; brain tissue; brain volume; chromosome 22q deletion syndrome; clinical biomarkers; cohort; computer infrastructure; connectome; cost effective; disease diagnosis; disorder risk; gene discovery; genetic variant; genome wide association study; genome-wide analysis; genomic biomarker; imaging modality; innovation; lifestyle factors; multitask; neuroimaging; new therapeutic target; novel; outcome forecast; predictive modeling; programs; risk variant; screening; search engine; tool; working group

Our comprehensive Data Science Research program is organized into 4 complementary Algorithm Research Cores: (1) Imaging Genomics, (2) Connectomics, (3) Machine Learning & Clinical Prediction, and (4) ENIGMA Disease Working Groups. World leaders in each field lead each Core, tackling computational questions on a scale not previously imagined or attempted. BigData tools will fuel ENIGMA'S worldwide scientific discoveries. We combine innovations in mathematics, machine learning, genomics, consortium science, and expertise from >20 countries and >125 institutions. ENIGMA is not a project, it is a scientific movement of rapidly and constantly Interacting collaborations that support each other. ENIGMA cohorts boost each other's power with gigantic datasets, tools and expertise to maximally exploit each other's data, performing some of the world's largest disease studies, beyond what any one site could perform alone. ENIGMA is distributed computation at its best, drawing on gigantic datasets and expertise. We create massive economic savings - drawing on worldwide computational and infrastructural resources vastly beyond what any one site in any one country would apply to a targeted biomedical goal. We bring BigData Science and the ENIGMA Consortium together to advance Worldwide Medicine. In Cores 1 and 2 we mine images and connectomes for genetic markers that re-wire the brain or boost brain tissue loss, using new mathematics to prioritize and organize trillions of computations, jointly searching images and genomes. In Core 3, we unleash multi-task sparse learning to predict diagnosis and prognosis from vast high-dimensional biomarker data in the largest neuroimaging genetics datasets ever. In Core 4 we use these tools in a massive distributed computation: a vast, mutually interacting set of 9 Worldwide Working Groups, led by experts in 9 major diseases of the brain - schizophrenia, bipolar, major depression, ADHD, autism, OCD, 22q deletion syndrome, HIV/AIDS, and addictions. We will discover what genes, medications, and lifestyle factors promote or resist brain disease worldwide. Our ENIGMA Center is a worldwide movement in mutually supportive discovery in medicine - spurred on by tools to perform gigantic computations never before Imagined.

我们的综合数据科学研究计划被组织为4个互补的算法研究 核心：(1)成像基因组学，(2)连接学，(3)机器学习和临床预测，以及(4)谜 疾病工作组。每个领域的世界领导者领导每个核心，在一个 以前没有想象过或尝试过的规模。BigData工具将推动Enigma在世界范围内的科学发现。 我们结合了数学、机器学习、基因组学、联盟科学方面的创新，以及来自 &gt；20个国家和125个机构。谜不是一个项目，它是一个迅速和 持续互动的相互支持的协作。谜团们通过以下方式提升彼此的力量 巨大的数据集、工具和专业知识，以最大限度地利用彼此的数据，执行世界上一些 规模最大的疾病研究，超出了任何一个网站可以单独完成的任务。Enigma是分布式计算，位于 它是最好的，利用了巨大的数据集和专业知识。我们创造了大量的经济节省-依靠 世界范围内的计算和基础设施资源远远超过任何一个国家的任何一个地点 将适用于有针对性的生物医学目标。我们将BigData Science和Enigma财团结合在一起 推动世界医学的发展。在核心1和2中，我们挖掘图像和连接以寻找遗传标记 重新连接大脑或增加脑组织损失，使用新的数学方法来优先处理和组织数万亿 计算，联合搜索图像和基因组。在核心3中，我们释放了多任务稀疏学习来 在最大的神经成像中从海量高维生物标志物数据预测诊断和预后 遗传学数据集。在Core4中，我们在大规模分布式计算中使用这些工具：巨大的、相互的 由9个全球工作组组成的互动小组，由9种主要脑部疾病-精神分裂症的专家领导， 躁郁症、抑郁症、多动症、自闭症、强迫症、22q缺失综合征、艾滋病毒/艾滋病和成瘾。我们会 发现哪些基因、药物和生活方式因素在全球范围内促进或抵抗脑部疾病。我们的 谜中心是一场由工具推动的相互支持的医学发现的全球运动 进行以前从未想象过的巨大计算。