HIV-1蛋白酶抑制剂的结构活性关系研究
详细信息 查看官网全文
- 英文论文题名：STRUCTURE AND BIOACTIVITY RELATIONSHIP STUDY ON THE HIV-1 PROTEASE
- 论文作者：李杨 ; 阎爱侠
- 英文论文作者：Yang Li ; Aixia Yan ; Department of Pharmaceutical Engineering ; Beijing University of Chemical Technology
- 年：2016
- 作者机构：北京化工大学;
- 论文关键词：蛋白酶抑制剂 ; 支持向量机 ; 随机森林 ; 构效关系
- 会议召开时间：2016-07-01
- 会议录名称：中国化学会第30届学术年会摘要集-第二十五分会：化学信息学与化学计量学
- 语种：中文
- 分类号：TQ460.1
- 学会代码：ZGHY
- 会议名称：中国化学会第30届学术年会-第二十五分会 ; 化学信息学与化学计量学
- 会议地点：中国辽宁大连
- 主办单位：中国化学会
- 学会名称：中国化学会
- 页数：1
- 文件大小：225k
- 原文格式：D
- 会议级别：全国
摘要
HIV-1蛋白酶是一种逆转录天冬氨酰蛋白酶,对HIV-1病毒的生命周期至关重要,是治疗艾滋病药物的重要靶标之一~([1])。本研究主要利用计算化学及构效关系等方法建立了HIV-1蛋白酶抑制剂生物活性的高低活性预测模型。在高低活性预测模型的建立过程中,收集了740个HIV-1蛋白酶的抑制剂,并将抑制剂随机分为训练集和测试集。用ADRIANA.Code和MACCS fingerprints描述符,以支持向量机、随机森林的方法建立模型,最好的模型训练集正确率为98.78%,测试集正确率为90.69%。通过对建立模型的描述符从分析,我们认为抑制剂的手性结构和氢键可能对活性有较大影响。
HIV-1 protease is a retroviral aspartyl-protease which is essential for the life-cycle of HIV.HIV-1 protease inhibitors can decrease the activity of HIV-1.It is an important target of the AIDS drug treatment.We use the methods of computational chemistry and structure-activity relationship to build classification models.In the model of high-and-low prediction,we selected 740 HIV-1 protease inhibitors and divided the compounds into training set and test set randomly.We built classification models using ADRIANA.Code and MACCS fingerprints descriptors with support vector machine(SVM) and random forest(RF) method.The best model obtained an accuracy of 98.78%on the training set and 90.69%on the test set.By analyzing descriptors,we found that the molecular chirality and hydrogen bonding may play important roles in the process.
HIV-1 protease is a retroviral aspartyl-protease which is essential for the life-cycle of HIV.HIV-1 protease inhibitors can decrease the activity of HIV-1.It is an important target of the AIDS drug treatment.We use the methods of computational chemistry and structure-activity relationship to build classification models.In the model of high-and-low prediction,we selected 740 HIV-1 protease inhibitors and divided the compounds into training set and test set randomly.We built classification models using ADRIANA.Code and MACCS fingerprints descriptors with support vector machine(SVM) and random forest(RF) method.The best model obtained an accuracy of 98.78%on the training set and 90.69%on the test set.By analyzing descriptors,we found that the molecular chirality and hydrogen bonding may play important roles in the process.
引文
[1]Ali,A.;Kumar Reddy,G.S.K.;Nalam,M.N.L.;Anjum,S.G.;Cao,H.;.Schiffer,C.A.;Rana T.M..J.Med.Chem.2010(53):7699.