.png){kind=icon}
返回
Linked Entity Attribute Pair (LEAP): A Harmonization Framework for Data Pooling
.png){kind=icon}
OA
AI
DOI：10.1200/CCI.20.00037.png)
.png){kind=icon}
.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
PURPOSE As data-sharing projects become increasingly frequent, so does the need to map data elements between multiple classification systems. A generic, robust, shareable architecture will result in increased efficiency and transparency of the mapping process, while upholding the integrity of the data. MATERIALS AND METHODS The American Association for Cancer Research's Genomics Evidence Neoplasia Information Exchange (GENIE) collects clinical and genomic data for precision cancer medicine. As part of its commitment to open science, GENIE has partnered with the National Cancer Institute's Genomic Data Commons (GDC) as a secondary repository. After initial efforts to submit data from GENIE to GDC failed, we realized the need for a solution to allow for the iterative mapping of data elements between dynamic classification systems. We developed the Linked Entity Attribute Pair (LEAP) database framework to store and manage the term mappings used to submit data from GENIE to GDC. RESULTS After creating and populating the LEAP framework, we identified 195 mappings from GENIE to GDC requiring remediation and observed a 28% reduction in effort to resolve these issues, as well as a reduction in inadvertent errors. These results led to a decrease in the time to map between OncoTree, the cancer type ontology used by GENIE, and International Classification of Disease for Oncology, 3rd Edition, used by GDC, from several months to less than 1 week. CONCLUSION The LEAP framework provides a streamlined mapping process among various classification systems and allows for reusability so that efforts to create or adjust mappings are straightforward. The ability of the framework to track changes over time streamlines the process to map data elements across various dynamic classification systems. (c) 2020 by American Society of Clinical Oncology
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
J
IF:
2.8
论文数:
669
被引数:
2.6K
机构
查看更多机构
引用论文
AACR Project GENIE: Powering Precision Medicine through an International ConsortiumAACR项目精灵: 通过国际财团为精密医学提供动力.png)
CANCER DISCOVERYOAAI
CANCER DISCOVERYIF33.3
OAAI分享收藏分享收藏