摘要
新药物研发时间长、成本高,但成功率低,为了提高收益比,药物重定位即旧药新用受到了广泛关注。从临床和实验角度鉴定药物的新用途需要耗费大量人力和物力,从计算角度预测药物新用途成为研究热点;并且,随着药物和疾病相关的大量多层次组学数据积累,通过挖掘药物相关数据鉴定药物新用途成为可能。重点挖掘药物化学结构、药理性质、药物靶蛋白功能、疾病表型等数据得到相应特征,并将这些药物疾病特征进行整合,再将特征输入XG-BOOST模型进行预测。实验结果表明,该方法准确率达87.9%,较逻辑回归、随机森林具有更高的预测精度。
The development of new drugs is long and costly,but the success rate is low.Therefore,in order to improve the yield,drug relocation,that is,the new use of old drugs has received extensive attention.The clinical and experimental identification of new uses of drugs requires a lot of manpower and material resources,and predicting the new use of drugs from a computational perspective has become a research hotspot in recent years.On the other hand,in recent years,the rapid accumulation of a large number of multi-level omics data related to drug-related and disease has made it possible to identify new drug uses by mining drug-related data.In this paper,the characteristics of the chemical structure,pharmacological properties,drug target protein function,disease phenotype,etc.of the drug were obtained,and the characteristics of these drugs were integrated.Finally,the feature is input into the XG-BOOST model for prediction.The experimental results show that our method has higher prediction accuracy than logistic regression and random forest.
作者
李苗苗
LI Miao-miao(Business School,University of Shanghai for Science and Technology,Shanghai 200090,China)
出处
《软件导刊》
2020年第2期110-113,共4页
Software Guide
作者简介
李苗苗(1995-),女,上海理工大学管理学院硕士研究生,研究方向为系统生物学。
