Leveraging Heterogeneity in Preclinical Traumatic Brain Injury to Drive Discovery and Reproducibility

利用临床前创伤性脑损伤的异质性来推动发现和重现性

• 批准号: 10212363
• 负责人: Austin C Chou
• 金额: $ 4.52万
• 依托单位: UNIVERSITY OF CALIFORNIA, SAN FRANCISCO
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-07-01 至 2022-02-14
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10212363
• 关键词: Address; Affect; Animal Model; Big Data; Biological; Biological Markers; Biomechanics; Brain region; Chronic; Clinic; Clinical; Closed head injuries; Common Data Element; Complex; Data; Data Analyses; Data Collection; Data Commons; Data Element; Data Science; Data Set; Development; Foundations; Goals; Heterogeneity; Incidence; Individual; Inflammation; Informatics; Injury; Institutes; Machine Learning; Measures; Methods; Modeling; Multivariate Analysis; National Institute of Neurological Disorders and Stroke; Outcome; Pathway interactions; Pattern; Pharmacologic Substance; Population; Positioning Attribute; Pre-Clinical Model; Principal Component Analysis; Publishing; Reproducibility; Research; Severities; Standardization; Synaptic plasticity; Therapeutic; TimeLine; Translating; Translational Research; Translations; Traumatic Brain Injury; behavioral outcome; bench to bedside; biomarker discovery; controlled cortical impact; data curation; data framework; data harmonization; data sharing; disability; experimental study; functional outcomes; genetic manipulation; improved; insight; multidimensional data; multiple datasets; nerve injury; nervous system disorder; neuroinflammation; open data; patient subsets; pre-clinical; pre-clinical research; precision medicine; repository; response; response to injury

Traumatic brain injury (TBI) is a leading cause of neurological disorders and affects over 2.5 million people each year, yet no treatment has successfully translated from bench to clinic. TBI is a broad term and encompasses an extremely heterogeneous set of injuries differing by cause, severity, biomechanics, and the varied, complex secondary injury responses that collectively result in chronic disabilities. Current preclinical research circumvents the issue of TBI heterogeneity by relying on specific preclinical animal models that mimic subpopulations of patients and particular secondary injury mechanisms with each study focusing on limited, individual pathways. This proposal instead aims to tackle TBI heterogeneity by approaching TBI as a “big data” problem and aggregating and analyzing the multidimensional data collectively. A framework for data harmonization and curation will be developed, and datasets from a consortium of preclinical labs employing a variety of preclinical TBI models will be collected and curated into an open data commons (ODC-TBI). Utilizing machine learning and multidimensional analytics, the proposed research will directly leverage TBI heterogeneity in the merged dataset to identify persistent features of TBI to empower translational research. By creating a preclinical TBI ODC and applying machine learning to integrate the heterogeneity of preclinical TBI models, the project will reveal multidimensional features of TBI across heterogeneous injuries and characterize how diverse secondary injury mechanisms interact and ultimately affect injury outcome. Throughout the project's timeline, new datasets will continue to be harmonized into the ODC-TBI according to the established framework. The ODC-TBI will be the first open multicenter, multi-model repository of preclinical TBI data and will enable the application of data science to the field of TBI. Furthermore, the ODC-TBI and the methods implemented throughout the project will be openly shared to improve reproducibility of TBI research. Together with the multidimensional analysis that will provide quantitative and qualitative understanding of TBI heterogeneity, the project aims to ultimately accelerate data- driven discovery and precision medicine for TBI.

创伤性脑损伤（TBI）是神经系统疾病的主要原因，并且每个创伤性脑损伤影响超过250万人。 然而，没有一种治疗方法成功地从实验室转移到诊所。TBI是一个广泛的术语，包括 一组因原因、严重程度、生物力学和各种复杂的 二次伤害反应，共同导致慢性残疾。目前的临床前研究规避了 TBI异质性的问题依赖于特定的临床前动物模型， 患者和特定的继发性损伤机制，每项研究都侧重于有限的个体途径。 相反，该提案旨在通过将TBI视为“大数据”问题来解决TBI异质性， 集中地聚集和分析多维数据。数据统一框架， 将开发策展，并使用各种临床前实验室的临床前实验室联盟的数据集， TBI模型将被收集并整理到开放数据共享（ODC-TBI）中。利用机器学习和 多维分析，拟议的研究将直接利用合并数据集中的TBI异质性 识别TBI的持续特征，以增强转化研究。通过创建临床前TBI ODC， 应用机器学习来整合临床前TBI模型的异质性，该项目将揭示 TBI在异质性损伤中的多维特征，并描述继发性损伤的多样性 机制相互作用并最终影响损伤结果。在整个项目的时间轴中，新的数据集将 根据既定框架，继续将其统一纳入臭氧消耗物质技术性基础设施。ODC-TBI将是 首个开放的多中心、多模型临床前TBI数据库，将实现数据科学的应用 TBI领域。此外，ODC-TBI和整个项目实施的方法将公开 共享以提高TBI研究的可重复性。再加上多方面的分析， TBI异质性的定量和定性理解，该项目旨在最终加速数据- 驱动的发现和精确的治疗。

项目成果

Expert-augmented automated machine learning optimizes hemodynamic predictors of spinal cord injury outcome.

• DOI: 10.1371/journal.pone.0265254
• 发表时间: 2022
• 期刊: PloS one
• 影响因子: 3.7
• 作者: Chou A;Torres-Espin A;Kyritsis N;Huie JR;Khatry S;Funk J;Hay J;Lofgreen A;Shah R;McCann C;Pascual LU;Amorim E;Weinstein PR;Manley GT;Dhall SS;Pan JZ;Bresnahan JC;Beattie MS;Whetstone WD;Ferguson AR;TRACK-SCI Investigators
• 通讯作者: TRACK-SCI Investigators