Cheminformatics analysis of natural products and indication distribution prediction  

天然产物的化学信息学分析及骨架适应症预测(英文)

作　　者：Xin Zhou Yibo Li Chuanyu Lv Zhenming Liu Liangren Zhang 周鑫;李亦博;吕传宇;刘振明;张亮仁(北京大学医学部药学院天然药物与仿生药物国家重点实验室,北京100191)

机构地区：Department of Medicinal Chemistry, School of Pharmaceutical Sciences, Peking University Health Science Center, Beijing 100191, China

出　　处：《Journal of Chinese Pharmaceutical Sciences》2017年第9期635-641,共7页中国药学（英文版）

基　　金：The National Natural Science Foundation of China(Grant No.21572010B020601)

摘　　要：As a valuable resource for drug discovery, natural products remain largely unexplored. The cheminformatics analysis of natural product databases could help us know better about natural products, providing valuable information in drug design. In this study, we collected an in-home natural product library consisting of more than 220 000 molecules. The results showed that natural products were distributed more diversely than synthetic compounds and approved drugs in chemical space, and natural products still possessed better scaffold diversity. Besides, natural product scaffolds had more potential in some specific indications, such as antiarthritic, antihypertensive, antiallergic and analgesic. However, the utilization rate of natural product scaffolds is relatively low, especially in terms of potential indications. Therefore, we recommend the greater use of natural products while designing lead libraries.天然产物是药物发现的重要来源,对天然产物库进行信息学分析有助于加深我们在整体层面上对天然产物的认识,并起到辅助药物设计的作用。本研究收集了较为完整的天然产物库,包含超过220 000个天然产物分子,对其进行分析结果表明天然产物在化学空间上的分布比合成化合物和药物更广,天然产物库的骨架多样性也优于其他化合物库。另外,通过适应症预测模型发现,天然产物骨架在某些适应症例如关节炎、高血压、抗过敏、镇痛上分布较广,然而现有的天然产物来源药物对其利用度并不高,因此我们建议在先导库设计中更多地利用天然产物骨架结构。

关 键 词：Natural products Feature analysis Indication prediction 

分 类 号：R914[医药卫生—药物化学]