Gabetta Matteo, Larizza Cristiana, Bellazzi Riccardo
Department of Industrial and Information Engineering, University of Pavia, Italy.
Stud Health Technol Inform. 2013;192:412-6.
Literature-Based Discovery (LBD) is a technique that can be used in translational research to connect the very sparse and huge information available in scientific publications in order to extract new knowledge. This paper presents an LBD system based on the open discovery paradigm exploiting NLP techniques and UMLS medical concepts mapping, to provide a set of tools useful to discover unknown relationships. The system has been evaluated on the problem of discovering new candidate genes potentially related to dilated cardiomyopathies (DCM), and can be used in any medical context to connect different type of concepts. The validation of the system involves reproducing the discovery of genes currently associated to DCM. Validation showed that the system is able to discover many gene-disease associations by using the literature available before their first publication in a scientific article.
基于文献的发现（LBD）是一种可用于转化研究的技术，用于连接科学出版物中非常稀疏且海量的信息，以提取新知识。本文提出了一种基于开放发现范式的LBD系统，该系统利用自然语言处理（NLP）技术和统一医学语言系统（UMLS）医学概念映射，以提供一套有助于发现未知关系的工具。该系统已针对发现可能与扩张型心肌病（DCM）相关的新候选基因的问题进行了评估，并且可用于任何医学背景中以连接不同类型的概念。该系统的验证涉及重现当前与DCM相关的基因的发现。验证表明，该系统能够通过使用在科学文章中首次发表之前的现有文献发现许多基因与疾病的关联。
