Weber Lutz, Böhme Timo, Irmer Matthias
OntoChem GmbH, Heinrich-Damerow-Str. 4, D-06120, Germany.
Pharm Pat Anal. 2013 Jan;2(1):39-54. doi: 10.4155/ppa.12.76.
Ontology-based semantic text analysis methods allow to automatically extract knowledge relationships and data from text documents. In this review, we have applied these technologies for the systematic analysis of pharmaceutical patents. Hierarchical concepts from the knowledge domains of chemical compounds, diseases and proteins were used to annotate full-text US patent applications that deal with pharmacological activities of chemical compounds and filed in the years 2001-2010. Compounds claimed in these applications have been classified into their respective compound classes to review the distribution of scaffold types or general compound classes such as natural products in a time-dependent manner. Similarly, the target proteins and claimed utility of the compounds have been classified and the most relevant were extracted. The method presented allows the discovery of the main areas of innovation as well as emerging fields of patenting activities - providing a broad statistical basis for competitor analysis and decision-making efforts.
基于本体的语义文本分析方法能够从文本文件中自动提取知识关系和数据。在本综述中，我们已将这些技术应用于药物专利的系统分析。来自化合物、疾病和蛋白质知识领域的层次概念被用于注释2001年至2010年期间提交的涉及化合物药理活性的美国专利全文申请。这些申请中所要求保护的化合物已被分类到各自的化合物类别中，以便以时间依赖的方式审查支架类型或一般化合物类别（如天然产物）的分布情况。同样，已对化合物的靶蛋白和所要求保护的用途进行了分类，并提取了最相关的信息。所提出的方法能够发现创新的主要领域以及专利活动的新兴领域——为竞争对手分析和决策工作提供广泛的统计基础。
