基于文献的中医经方靶点预测模型研究

Literature-based target prediction model of classical prescriptions in traditional Chinese medicine

由于经方由药材构成,药材又由成分组成,成分和靶点在靶点数据库中可以建立直接联系。因此先提取经方配方和经方成分;然后进行靶点提取,借助靶点数据库构建靶点筛选模型;最后将不同来源的靶点汇总,计算靶点置信度,绘制经方靶点预测。根据经方到靶点的不同路径建立从经方到靶点之间的联系,构建经方靶点预测模型,并以芍药甘草汤和大黄黄连泄心汤为例进行综合实验验证预测模型的有效性。预测模型利用自然语言处理技术,借助外部数据库,从文献中发现经方靶点,可以减轻科研人员文献调研的负担。

The direct association between ingredients of medicinal materials and their targets can be established in targets database since the classical prescriptions are formulated according to the medicinal materials and the medicinal materials are composed of their ingredients.The ingredients of medicinal formulae and classical prescriptions were extracted followed by extracting the targets of medicinal materials and establishing the model for screening their targets by making use of targets database.The targets selected from different sources were collected with their confidence calculated.The target prediction table of classical prescriptions was worked out.The target prediction model of classical prescriptions was established by establishing the association between classical prescriptions and their targets based on the different paths from classical prescriptions to their targets.Its effectiveness was verified in comprehensive experiment of Peony and Licorice Decoction,Radix et Rhizoma Rhei and Coptis Decoction to remove the heart-fire.The target prediction model of classical prescriptions helps the scientific researchers to discover the targets from literature by making use of its natural language technique and the external databases,and can thus reduce their pressure in literature survey.