雌激素类衍生物对雌激素受体结合活性的QSAR和分子对接研究

发布时间：2018-01-13 19:31

本文关键词：雌激素类衍生物对雌激素受体结合活性的QSAR和分子对接研究　出处：《西北农林科技大学》2012年硕士论文　论文类型：学位论文

【摘要】：雌激素是由人体产生的一类能够和雌激素受体相互作用的配体分子，雌激素的存在对人体的生殖系统、骨骼系统和神经系统等都有显著影响。雌激素相关的疾病包括生殖系统癌症、神经系统紊乱、肥胖等。这些疾病的发生是由雌激素和相应的受体相互作用引发的。雌激素受体结合雌激素后与其他蛋白质相互作用，进而参与或调节不同的生物通路，因此选择合适的雌激素类化合物和特定受体结合是相关疾病治疗和预防的一个重要途径，并且以雌激素受体为靶标的药物设计在化学和制药行业一直都是一个热点。本研究选取82个雌激素受体配体分子构建QSAR模型和分子对接学习，以期能够为雌激素药物设计和筛选提供良好的工具。我们共采用MLR、PLSR、和BRNN三种方法分别对ERα、ERβ的结合活性以及配体对两者间的结合选择性建立QSAR模型，其中对ERα和ERβ活性，MLR和PLSR模型的结果要优于BRNN模型（ERα：MLR的统计结果为Rtr~2=0.72, Qte~2=0.63； PLSR模型的统计结果为Rtr~2=0.92,Qte~2=0.84。 ERβ: MLR统计结果为Rtr~2=0.75, Qte~2=0.75；PLSR统计结果为Rtr~2=0.98, Qte~2=0.80）。对于ER受体选择性来说，MLR模型也是表现最好的Rtr~2=0.74，Qte~2=0.80。 Surflex-dock技术被用来进行配体结合蛋白质受体的对接学习，结果表明这类分子的8位取代基和母核平面间的夹角对分子结合有很大影响。通过使用多个蛋白质结晶进行对接学习，，我们探讨了一些对接中出现的问题，结果表明对接得分和结合活性间线性关系极弱。而且蛋白质结晶的不同以及分子处理方式的不同对分子对接有很大的影响。这在使用分子对接时应该给予充分重视。
[Abstract]:Estrogen is a kind of ligand produced by human body which can interact with estrogen receptor. The existence of estrogen affects the reproductive system of human body. The skeletal system and nervous system have significant effects. Estrogen-related diseases include cancer of the reproductive system and disorders of the nervous system. Obesity and so on. These diseases are caused by the interaction between estrogen and corresponding receptors. Estrogen receptors interact with other proteins after binding estrogen, and then participate in or regulate different biological pathways. Therefore, the selection of appropriate estrogen compounds and specific receptor binding is an important approach to the treatment and prevention of related diseases. Drug design targeting estrogen receptor has been a hot topic in chemical and pharmaceutical industries. 82 estrogen receptor ligands were selected to construct QSAR model and molecular docking learning. In order to provide a good tool for the design and screening of estrogen drugs, we used MLR PLSR and BRNN respectively to detect ER 伪. The binding activity of ER 尾 and the selectivity of ligand binding to ER 尾 were used to establish QSAR model, in which ER 伪 and ER 尾 activity were determined. The results of MLR and PLSR model are better than that of BRNN model ER 伪: MLR. The statistical results of the PLSR model were Rtrtrine 20.92% Qtei 20.84.The ER 尾: MLR statistic result was Rtr~2=0.75. Qteo 2: 0.75; The results of PLSR were Rtrl2 + 0.98 and QT 2 + 0. 80%. MLR model was also the best Rtr~2=0.74 for ER receptor selectivity. Qtean 0.80. Surflex-dock technology is used to study the docking of ligand binding protein receptors. The results show that the angle between the 8-position substituent group and the parent nuclear plane has a great influence on the molecular binding. By using multiple protein crystals for docking learning, we discuss some problems in the docking process. The results show that the linear relationship between docking score and binding activity is very weak, and the difference of protein crystallization and molecular treatment has great influence on molecular docking. This should be given sufficient weight when using molecular docking. Look at.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R341

【共引文献】