Preliminary Exploration of Improving Predictive Capability of Three Dimensional Quantitative Structure Activity Relationship Models
-
摘要:
为了解决使用三维定量构效关系（three-dimensional quantitative structure-activity relationship，3D-QSAR）模型预测新化合物生物活性效果不理想的问题，建立了2种新的一致性模型.模型一是由多元线性回归（multiple linear regression，MLR）方法构建的加权一致性模型（weighted consensus modeling，WCM），该模型为每个子模型添加了各自的权重系数.模型二通过计算多个子模型预测值的平均值来构建平均一致性模型（average consensus modeling，ACM）.研究结果表明，当交叉验证相关系数0.5＜q2≤0.8时，一致性模型可以提高预测能力，而在q2＞0.8时不能提高3D-QSAR模型的预测能力.该方法可为提高模型预测能力和设计新型高活性抑制剂提供帮助.
Abstract:To solve the problem that the three-dimensional quantitative structure-activity relationship (3D-QSAR) model is not ideal when using the model to predict the biological activity of the new compounds, two new consensus models were established to improve the prediction ability of the model. A different weight to each submodule (named weighted consensus model, WCM) was added to one of the consensus models. In order to construct WCM, multiple linear regression (MLR) methods were used to calculate different weight coefficients for each submodule. Another consensus model was constructed from the average of the predicted values for each sub-model obtained in the literature (named average consensus model, ACM). Results show that the consensus model can improve the prediction ability when 0.5 < q2 ≤ 0.8, but it can't improve the 3D-QSAR model's prediction ability when q2 > 0.8. This result can help to improve the prediction of the model and the design of new high activity inhibitors.
-
Keywords:
- 3D-QSAR /
- consensus model /
- weighted consensus modeling /
- average consensus modeling
-
三维定量构效关系(three-dimensional quantitative structure-activity relationship，3D-QSAR)是一种可用于描述分子生物活性与其结构之间定量关系的方法，该方法可以预测药物分子和生物大分子之间的相互作用[1-2]，并评估抑制剂对受体分子的抑制效果.比较分子场分析(comparative molecular field analysis，CoMFA)方法与比较分子相似因子分析(comparative molecular similarity induces analysis，CoMSIA)方法是2种常用的3D-QSAR方法. 3D-QSAR模型广泛应用于生物学、医学、化学、环境科学等众多领域，研究内容包括分子的生物活性[3-5]、化合物毒性、药代动力学参数和抑制剂的抑制机理等[6-7].使用3D-QSAR方法研究人类免疫缺陷病毒(human immunodeficiency virus，HIV)抑制剂的技术目前已经比较成熟[8]，但是仍需提高模型的预测准确性.
现有的提高模型的预测准确性方法之一是建立一致性模型.一致性模型与单个模型相比具有许多优点，如一致性模型比单一模型更稳定、具有更强的泛化能力.一致性模型可以更充分地描述整个数据集中的分子特征，并从整个数据集中获得更全面的分子结构信息. 2013年，Helguera等[3]发表了关于预测人类单胺氧化酶抑制活性和选择性的一致性模型的文章，首先，使用包括Dragon、MOE和TOPS-MODE模型在内的多种3D-QSAR模型，计算了21组数据并按顺序进行编号，然后将几个数据集组合成一个一致性模型，再验证这个新模型对预测能力是否有提升.这一研究证明一致性模型在一定程度上确实可以提高模型的预测能力.
1. 原理与方法
1.1 一致性模型
本文使用多元线性回归和均值计算这2种经典统计方法，分别建立一致性模型，然后比较这2种一致性模型的预测能力.这个实验的过程如图 1所示.
模型一是由多元线性回归(multiple linear regression，MLR)方法构建的加权一致性模型(weighted consensus modeling，WCM)，这一模型为每个子模型添加了各自的权重系数.多元线性回归是研究一组独立变量如何直接影响因变量的方法，其最大的优势是可以根据2个变量之间的线性关系清晰地分析它们的物理意义[9].模型二是通过计算多个子模型预测值的平均值来构建平均一致性模型(average consensus modeling，ACM).最后使用配对t检验来验证ACM和WCM的预测能力是否比单个模型更高.
1.2 数据库的构建
为了构建数据库，本课题组查阅了数百篇文献，并从2006—2013年发表的80篇文章中收集了3D-QSAR模型的数据作为原始实验数据[1, 8, 10-86].本文使用IC50(最大抑制浓度的50%)的负对数，即pIC50作为化合物分子生物活性的指标.
1.3 一致性模型的构建
下面以Lu等[11]的论文为例详细介绍模型构建的方法.在其论文中共构建了48个化合物分子，其中随机挑选12个作为测试集化合物，在其编号后加角标t表示，其余化合物作为训练集.训练集用来构建一致性模型，测试集用来验证一致性模型.这些化合物的相关数据见表 1.表中pIC50, CoMFA和pIC50, CoMSIA分别是从文献中获得的用CoMFA、CoMSIA两种模型计算得到的化合物活性的pIC50预测值.
表 1 文献中化合物的pIC50实验值和单一模型预测值Table 1. pIC50 experimental value and predicted value by single model of compounds in the literature化合物编号 实验值 pIC50, CoMFA pIC50, CoMSIA 1 7.30 7.26 7.23 2 5.79 5.73 5.73 3t 6.10 6.90 6.91 4 6.39 6.42 6.40 5 6.30 6.43 6.31 6t 5.97 6.56 6.02 7 5.93 5.92 5.87 8 6.14 6.20 6.11 9 6.43 6.45 6.44 10t 6.60 6.92 7.08 11 7.16 6.98 7.18 12 5.64 5.71 5.77 13 7.08 7.13 7.21 14 7.51 7.52 7.39 15t 7.26 7.09 7.25 16 7.59 7.62 7.58 17 7.19 7.17 7.20 18 7.50 7.60 7.51 19t 7.42 7.11 7.53 20 7.46 7.40 7.51 21 6.94 6.90 6.84 22 6.67 7.13 6.98 23t 7.68 7.46 7.52 24 7.11 7.00 7.00 25 7.36 7.33 7.31 26 7.62 7.51 7.53 27t 7.08 7.64 7.46 28 7.60 7.48 7.44 29 7.47 7.50 7.46 30 7.92 8.08 8.00 31t 7.37 7.45 7.18 32 7.39 7.33 7.33 33 6.17 6.08 6.20 34 7.30 7.22 7.34 35t 8.05 8.01 7.84 36 7.83 7.86 7.83 37 7.42 7.32 7.47 38 8.05 8.01 7.99 39t 8.09 8.14 8.13 40 8.09 8.06 8.12 41 8.22 8.15 8.25 42 8.25 8.27 8.32 43t 8.01 7.94 7.94 44 8.24 8.25 8.17 45 8.14 8.14 8.22 46 7.84 7.83 7.85 47t 8.24 8.41 8.24 48 8.17 8.22 8.10 注：“t”为测试集化合物. 构建一致性模型要将表 1中训练集的pIC50, CoMFA、pIC50, CoMSIA和实验值输入到SPSS软件，用多元线性回归分析方法，把pIC50, CoMFA和pIC50, CoMSIA作为自变量，实验值作为因变量.非标准化系数B即WCM方程的各项系数，由此得到WCM方程.另一个一致性模型是ACM的构建方法，即使用SPSS软件转换菜单中的计算变量来求取pIC50, CoMFA和pIC50, CoMSIA的平均值，即ACM的pIC50, A预测值.
1.4 模型的评价方法
从文献中引入2个参数q2和r2，q2是交叉验证相关系数，r2是非交叉验证相关系数. q2值和r2值分别由公式
$${q^2} = 1 - \frac{{\sum {{{({y_{{\rm{pred}}}} - {y_{{\rm{obs}}}})}^2}} }}{{\sum {{{({y_{{\rm{pred}}}} - {y_{{\rm{mean}}}})}^2}} }}$$ (1) $${r^2} = 1 - \frac{{\sum {{{({y_{{\rm{obs}}}} - {y_{{\rm{pred}}}})}^2}} }}{{\sum {{{({y_{{\rm{obs}}}} - {y_{{\rm{mean}}}})}^2}} }}$$ (2) 得出.式中：yobs表示测试集中化合物的pIC50的实验值；ypred是测试集中化合物的pIC50的预测值；ymean表示测试集所有分子pIC50值的均值.
将测试集中化合物的单一模型预测值pIC50, CoMFA和pIC50, CoMSIA代入一致性模型方程，便可得到pIC50, W或pIC50, A(pIC50, W和pIC50, A分别是使用WCM和ACM计算的测试集化合物的pIC50预测值).选择SPSS软件分析菜单中的回归线性，把pIC50, W或pIC50, A作为自变量，实验值作为因变量，点击确定进行计算得到rpred2.生物统计学领域认为，一个有效的模型应同时符合[10]
$${q^2} > 0.5$$ (3) $${r^2} > 6$$ (4) 不满足这2个条件的值应该被剔除.
2. 结果与讨论
2.1 模型构建
以Lu等[11]的工作为例，用多元线性回归方法计算WCM各项系数，由此得到WCM方程
$${\rm{pI}}{{\rm{C}}_{50,{\rm{W}}}} = 0.604{\rm{pI}}{{\rm{C}}_{50,{\rm{CoMFA}}}} + 0.367{\rm{pI}}{{\rm{C}}_{50,{\rm{CoMSIA}}}} + 0.131$$ (5) 使用ACM方法构建的一致性方程，即
$${\rm{pI}}{{\rm{C}}_{50,{\rm{A}}}} = \frac{1}{2}({\rm{pI}}{{\rm{C}}_{50,{\rm{CoMFA}}}} + {\rm{pI}}{{\rm{C}}_{50,{\rm{CoMSIA}}}})$$ (6) 由一致性方程得到pIC50, W和pIC50, A预测值，汇总至表 2中.
表 2 WCM和ACM计算得到的pIC50预测值Table 2. Predicted value of pIC50 calculated by WCM and ACM化合物编号 pIC50, W pIC50, A 3t 6.91 6.91 6t 6.14 6.30 10t 7.05 7.00 15t 7.22 7.17 19t 7.44 7.32 23t 7.51 7.49 27t 7.50 7.55 31t 7.24 7.32 35t 7.88 7.93 39t 8.14 8.14 43t 7.95 7.94 47t 8.28 8.33 注：“t”为测试集化合物. 2.2 一致性模型rpred2的直观分析
2.2.1 WCM
以文献[11]为例，计算得到的WCM的rpred2值为0.867，高于文献中给出的CoMFA模型的rpred2(0.810)和CoMSIA模型的rpred2(0.860)，说明使用WCM可以提高预测能力. 图 2是使用WCM计算出的80篇文章中预测值rpred2.
由图 2(a)可以看出，当CoMFA模型的rpred2＜0.850时，大部分WCM的rpred2比单一模型高；当单一模型的rpred2＞0.850时，大部分WCM的rpred2比单一模型低.由图 2(b)可以看出，当CoMSIA模型的rpred2＜0.750时，大部分WCM的rpred2比单一模型高；当单一模型的rpred2 满足0.750＜rpred2＜0.800时，WCM的部分rpred2大于单一模型；当单一模型的rpred2＞0.800时，大部分WCM方法的rpred2低于单一模型.所以WCM方法只能提高部分模型的预测能力.
2.2.2 ACM
以文献[11]为例，ACM的rpred2值为0.874，大于CoMFA模型rpred2值(0.810)和CoMSIA模型rpred2值(0.860)，说明使用ACM的预测能力更高.
将CoMFA和CoMSIA模型测试集中化合物分子的pIC50预测值代入公式(6)，计算这些化合物预测值的平均值pIC50, A，后续的验证步骤与WCM相同. 图 3是使用ACM计算出的80篇文章中的预测值rpred2.
由图 3(a)可以看出，在CoMFA模型的rpred2＜0.850时，大部分ACM的rpred2比单一模型高；在单一模型的rpred2＞0.850时，ACM的rpred2大部分没有明显提高.由图 3(b)可以看出，在CoMSIA模型的rpred2＜0.750时，本文所用的ACM的rpred2比单一模型高；在单一模型的rpred2＞0.750时一部分ACM的rpred2低于单一模型.所以ACM方法只能提高部分模型的预测能力.
2.3 单侧配对t检验
2.3.1 WCM
进一步将79篇文献中的q2和rpred2两个值按照q2的大小分成以下4组：0.5＜q2≤0.6, 0.6＜q2≤0.7, 0.7＜q2≤0.8和q2＞0.8.然后对WCM的rpred2和文献中的CoMFA或CoMSIA的rpred2值进行配对t检验.
用配对t检验验证WCM相对于CoMFA模型预测能力的提高效果，比较的结果见表 3.由表 3可以看出，当0.5＜q2≤0.6时，t的绝对值为2.588，大于t界值1.729(自由度df=19)，差异具有统计学意义.因为WCM的rpred2均值(0.777)大于CoMFA的rpred2均值(0.715)，所以WCM可以提高3D-QSAR模型的预测能力.当0.6＜q2≤0.7和0.7＜q2≤0.8时，差异具有统计学意义，且WCM的rpred2均值更高，所以结论与第一组相同，即WCM可以提高3D-QSAR模型的预测能力.但是当q2＞0.8时，t的绝对值为0.974，比t界值表中的统计值1.895(df = 7)小，差异不具有统计学意义.从整体上分析，结果见表 3最后一行，t值的绝对值大于t界值，差异具有统计学意义，且WCM的rpred2均值(0.820)大于CoMFA的rpred2均值(0.757)，这说明WCM提高了模型的预测能力.
表 3 WCM与CoMFA模型的比较结果Table 3. Comparison of WCM and CoMFA modelsq2 样本数 t值 t界值a 统计学意义 CoMFA $\bar r_{{\rm{pred}}}^2$b $\bar r_{{\rm{pred}}}^2$c 0.5~0.6 20 -2.588 1.729 有 0.715 0.777 0.6~0.7 23 -2.341 1.717 有 0.753 0.796 0.7~0.8 19 -1.904 1.734 有 0.773 0.819 ＞0.8 8 -0.974 1.895 无 0.840 0.871 ＞0.5 70 -4.120 1.667 有 0.757 0.820 注：a表示t界值的显著水平为95%，下同；b表示使用CoMFA模型计算出的rpred2的均值，记为CoMFA $\bar r_{{\rm{pred}}}^2$，下同；c表示使用WCM模型计算出的rpred2的均值，记为WCM $\bar r_{{\rm{pred}}}^2$，下同. 用配对t检验验证WCM相对于CoMSIA模型预测能力的提高效果，比较的结果见表 4，其结果与CoMFA模型相同.
表 4 WCM与CoMSIA模型的比较结果Table 4. Comparison of WCM and CoMSIA modelsq2 样本数 t值 t界值 统计学意义 CoMSIA $\bar r_{{\rm{pred}}}^2$a $\bar r_{{\rm{pred}}}^2$ 0.5~0.6 16 -3.090 1.753 有 0.655 0.772 0.6~0.7 23 -3.324 1.717 有 0.686 0.784 0.7~0.8 22 -2.143 1.721 有 0.780 0.811 ＞0.8 8 -0.152 1.895 无 0.911 0.916 ＞0.5 69 -4.729 1.667 有 0.735 0.806 注：a表示使用CoMSIA模型计算出的rpred2的均值，记为CoMSIA $\bar r_{{\rm{pred}}}^2$，下同. 2.3.2 ACM
用配对t检验验证ACM相对于CoMFA模型预测能力的提高效果，比较的结果见表 5.当0.5＜q2≤0.6时，t的绝对值为3.263，大于t界值1.729(df=19)，认为差异具有统计学意义.因为ACM的rpred2均值(0.787)大于CoMFA的rpred2均值(0.715)，证明ACM可以提高3D-QSAR模型的预测能力.当0.6＜q2≤0.7和0.7＜q2≤0.8时差异具有统计学意义，且ACM的rpred2均值更高，所以结论与第一组相同.但是当q2＞0.8时，t的绝对值为0.835，比t界值表中的统计值1.895(df=7)小，差异不具有统计学意义.从整体上分析，结果见表 5最后一行，t值的绝对值大于t界值，认为差异具有统计学意义，而且ACM的rpred2均值(0.815)大于CoMFA的rpred2均值(0.764)，可以说明ACM提高了模型的预测能力.
表 5 ACM与CoMFA模型的比较结果Table 5. Comparison of ACM and CoMFA modelsq2 样本数 t值 t界值 统计学意义 CoMFA $\bar r_{{\rm{pred}}}^2$ ACM $\bar r_{{\rm{pred}}}^2$a 0.5~0.6 20 -3.263 1.729 有 0.715 0.787 0.6~0.7 21 -2.499 1.725 有 0.764 0.814 0.7~0.8 21 -1.997 1.725 有 0.781 0.823 ＞0.8 8 -0.835 1.895 无 0.840 0.869 ＞0.5 70 -4.543 1.667 有 0.764 0.815 注：a表示使用ACM模型计算出的rpred2的均值，记为ACM $\bar r_{{\rm{pred}}}^2$，下同. 使用相同方法验证ACM相对于CoMSIA模型预测能力的提高效果，验证结果见表 6，结论与CoMFA模型相同.
表 6 ACM与CoMSIA模型的比较结果Table 6. Comparison of ACM and CoMSIA modelsq2 样本数 t值 t界值 统计学意义 CoMSIA $\bar r_{{\rm{pred}}}^2$ ACM $\bar r_{{\rm{pred}}}^2$ 0.5~0.6 15 -2.956 1.761 有 0.657 0.769 0.6~0.7 23 -4.629 1.717 有 0.686 0.799 0.7~0.8 22 -2.232 1.721 有 0.796 0.836 ＞0.8 9 -1.158 1.860 无 0.891 0.919 ＞0.5 69 -5.684 1.667 有 0.743 0.818 3. 结论
1) 配对t检验的结果表明：当使用WCM方法时，在0.5＜q2≤0.8的条件下，WCM提高了模型的预测能力；当q2＞0.8时，此方法不能提高3D-QSAR模型的预测能力.但是从整体上分析，可以认为WCM提高了模型的预测能力.
2) 当使用ACM方法时，配对t检验的结果与WCM配对t检验的结果相同.
3) 在q2≤0.8时，建立一致性模型可以提高原始模型的预测能力，而在q2＞0.8时不能提高3D-QSAR模型的预测能力.这说明本实验建立的WCM和ACM模型在一定条件下可以有效提高化合物活性的预测能力.这一结果可以为提高模型预测能力的研究和新型高活性抑制剂的设计提供帮助.
-
表 1 文献中化合物的pIC50实验值和单一模型预测值
Table 1 pIC50 experimental value and predicted value by single model of compounds in the literature
化合物编号 实验值 pIC50, CoMFA pIC50, CoMSIA 1 7.30 7.26 7.23 2 5.79 5.73 5.73 3t 6.10 6.90 6.91 4 6.39 6.42 6.40 5 6.30 6.43 6.31 6t 5.97 6.56 6.02 7 5.93 5.92 5.87 8 6.14 6.20 6.11 9 6.43 6.45 6.44 10t 6.60 6.92 7.08 11 7.16 6.98 7.18 12 5.64 5.71 5.77 13 7.08 7.13 7.21 14 7.51 7.52 7.39 15t 7.26 7.09 7.25 16 7.59 7.62 7.58 17 7.19 7.17 7.20 18 7.50 7.60 7.51 19t 7.42 7.11 7.53 20 7.46 7.40 7.51 21 6.94 6.90 6.84 22 6.67 7.13 6.98 23t 7.68 7.46 7.52 24 7.11 7.00 7.00 25 7.36 7.33 7.31 26 7.62 7.51 7.53 27t 7.08 7.64 7.46 28 7.60 7.48 7.44 29 7.47 7.50 7.46 30 7.92 8.08 8.00 31t 7.37 7.45 7.18 32 7.39 7.33 7.33 33 6.17 6.08 6.20 34 7.30 7.22 7.34 35t 8.05 8.01 7.84 36 7.83 7.86 7.83 37 7.42 7.32 7.47 38 8.05 8.01 7.99 39t 8.09 8.14 8.13 40 8.09 8.06 8.12 41 8.22 8.15 8.25 42 8.25 8.27 8.32 43t 8.01 7.94 7.94 44 8.24 8.25 8.17 45 8.14 8.14 8.22 46 7.84 7.83 7.85 47t 8.24 8.41 8.24 48 8.17 8.22 8.10 注：“t”为测试集化合物. 
下载: 导出CSV
表 2 WCM和ACM计算得到的pIC50预测值
Table 2 Predicted value of pIC50 calculated by WCM and ACM
化合物编号 pIC50, W pIC50, A 3t 6.91 6.91 6t 6.14 6.30 10t 7.05 7.00 15t 7.22 7.17 19t 7.44 7.32 23t 7.51 7.49 27t 7.50 7.55 31t 7.24 7.32 35t 7.88 7.93 39t 8.14 8.14 43t 7.95 7.94 47t 8.28 8.33 注：“t”为测试集化合物.
下载: 导出CSV
表 3 WCM与CoMFA模型的比较结果
Table 3 Comparison of WCM and CoMFA models
q2 样本数 t值 t界值a 统计学意义 CoMFA $\bar r_{{\rm{pred}}}^2$b $\bar r_{{\rm{pred}}}^2$c 0.5~0.6 20 -2.588 1.729 有 0.715 0.777 0.6~0.7 23 -2.341 1.717 有 0.753 0.796 0.7~0.8 19 -1.904 1.734 有 0.773 0.819 ＞0.8 8 -0.974 1.895 无 0.840 0.871 ＞0.5 70 -4.120 1.667 有 0.757 0.820 注：a表示t界值的显著水平为95%，下同；b表示使用CoMFA模型计算出的rpred2的均值，记为CoMFA $\bar r_{{\rm{pred}}}^2$，下同；c表示使用WCM模型计算出的rpred2的均值，记为WCM $\bar r_{{\rm{pred}}}^2$，下同.
下载: 导出CSV
表 4 WCM与CoMSIA模型的比较结果
Table 4 Comparison of WCM and CoMSIA models
q2 样本数 t值 t界值 统计学意义 CoMSIA $\bar r_{{\rm{pred}}}^2$a $\bar r_{{\rm{pred}}}^2$ 0.5~0.6 16 -3.090 1.753 有 0.655 0.772 0.6~0.7 23 -3.324 1.717 有 0.686 0.784 0.7~0.8 22 -2.143 1.721 有 0.780 0.811 ＞0.8 8 -0.152 1.895 无 0.911 0.916 ＞0.5 69 -4.729 1.667 有 0.735 0.806 注：a表示使用CoMSIA模型计算出的rpred2的均值，记为CoMSIA $\bar r_{{\rm{pred}}}^2$，下同.
下载: 导出CSV
表 5 ACM与CoMFA模型的比较结果
Table 5 Comparison of ACM and CoMFA models
q2 样本数 t值 t界值 统计学意义 CoMFA $\bar r_{{\rm{pred}}}^2$ ACM $\bar r_{{\rm{pred}}}^2$a 0.5~0.6 20 -3.263 1.729 有 0.715 0.787 0.6~0.7 21 -2.499 1.725 有 0.764 0.814 0.7~0.8 21 -1.997 1.725 有 0.781 0.823 ＞0.8 8 -0.835 1.895 无 0.840 0.869 ＞0.5 70 -4.543 1.667 有 0.764 0.815 注：a表示使用ACM模型计算出的rpred2的均值，记为ACM $\bar r_{{\rm{pred}}}^2$，下同.
下载: 导出CSV
表 6 ACM与CoMSIA模型的比较结果
Table 6 Comparison of ACM and CoMSIA models
q2 样本数 t值 t界值 统计学意义 CoMSIA $\bar r_{{\rm{pred}}}^2$ ACM $\bar r_{{\rm{pred}}}^2$ 0.5~0.6 15 -2.956 1.761 有 0.657 0.769 0.6~0.7 23 -4.629 1.717 有 0.686 0.799 0.7~0.8 22 -2.232 1.721 有 0.796 0.836 ＞0.8 9 -1.158 1.860 无 0.891 0.919 ＞0.5 69 -5.684 1.667 有 0.743 0.818
下载: 导出CSV
-
[1] GADHE C G, KOTHANDAN G, CHO S J. Binding site exploration of CCR5 using in silico methodologies:a 3D-QSAR approach[J]. Archives of Pharmacal Research, 2013, 36(1):6-31. doi: 10.1007/s12272-013-0001-1
[2] VILAR S, COSTANZI S. Predicting the biological activities through QSAR analysis and docking-based scoring[J]. Methods Mol Biol, 2012, 914(914):271-284. doi: 10.1007/978-1-62703-023-6_16
[3] HELGUERA A M, PÉREZ-GARRIDO A, GASPAR A, et al. Combining QSAR classification models for predictive modeling of human monoamine oxidase inhibitors[J]. European Journal of Medicinal Chemistry, 2013, 59:75-90. doi: 10.1016/j.ejmech.2012.10.035
[4] SHAHLAEI M, FASSIHI A, SAGHAIE L, et al. Computational evaluation of some indenopyrazole derivatives as anticancer compounds; application of QSAR and docking methodologies[J]. Journal of Enzyme Inhibition & Medicinal Chemistry, 2011, 28(1):16-32. http://www.oalib.com/references/4659790
[5] WIGGERS H J, ROCHA J R, CHELESKI J, et al. Integration of ligand-and target-based virtual screening for the discovery of cruzain inhibitors[J]. Qsar & Combinatorial Science, 2011, 30(6/7):565-578. http://www.producao.usp.br/handle/BDPI/31739
[6] ZAKHAROV A V, PEACH M L, SITZMANN M, et al. Computational tools and resources for metabolism-related property predictions. 2. Application to prediction of half-life time in human liver microsomes[J]. Future Medicinal Chemistry, 2012, 4(15):1933-1944. doi: 10.4155/fmc.12.152
[7] TETKO I V, SUSHKO I, PANDEY A K, et al. Critical assessment of QSAR models of environmental toxicity against Tetrahymena pyriformis:focusing on applicability domain and overfitting by variable selection[J]. Journal of Chemical Information & Modeling, 2008, 48(9):1733-1746. doi: 10.1021/ci800151m?src=recsys
[8] HU R J, BARBAULT F, DELAMAR M, et al. Receptor-and ligand-based 3D-QSAR study for a series of non-nucleoside HIV-1 reverse transcriptase inhibitors[J]. Bioorganic & Medicinal Chemistry, 2009, 17(6):2400-2409. https://www.sciencedirect.com/science/article/pii/S0968089609001400
[9] FENG L P, GUO Z Y, LIANG J, et al. Research progress and application in the several QSAR modeling method[J]. Journal of Agro-Environment Science, 2007, 26:651-655.
[10] SUN X H, GUAN J Q, TAN J J, et al. 3D-QSAR studies of quinoline ring derivatives as HIV-1 integrase inhibitors[J]. Sar & Qsar in Environmental Research, 2012, 23(7/8):683-703. doi: 10.1080/1062936X.2012.717541?queryID=
[11] LU P, WEI X, ZHANG R. CoMFA and CoMSIA 3D-QSAR studies on quionolone caroxylic acid derivatives inhibitors of HIV-1 integrase[J]. European Journal of Medicinal Chemistry, 2010, 45(8):3413-3419. doi: 10.1016/j.ejmech.2010.04.030
[12] GADHE C G, KOTHANDAN G, MADHAVAN T, et al. Molecular modeling study of HIV-1 gp120 attachment inhibitors[J]. Medicinal Chemistry Research, 2011, 21(8):1892-1904. doi: 10.1007/s00044-011-9711-4.pdf
[13] RAVICHANDRAN V, SANKAR S, AGRAWAL R K. Predicting anti-HIV activity of 1, 1, 3-trioxo[1, 2, 4] -thiadiazine (TTD) derivatives:3D QSAR approach[J]. Medicinal Chemistry Research, 2009, 18(7):511-522. doi: 10.1007/s00044-008-9145-9
[14] TEIXEIRA C, SERRADJI N, MAUREL F, et al. Docking and 3D-QSAR studies of BMS-806 analogs as HIV-1 gp120 entry inhibitors[J]. European Journal of Medicinal Chemistry, 2009, 44(9):3524-3532. doi: 10.1016/j.ejmech.2009.03.028
[15] VYAS V K, PARIKH H, GHATE M. 3D QSAR studies on 5-(2-methylbenzimidazol-1-yl)-N -alkylthiophene-2-carboxamide derivatives as P. falciparum, dihydroorotate dehydrogenase (Pf DHODH) inhibitors[J]. Medicinal Chemistry Research, 2013, 22(5):2235-2243. doi: 10.1007/s00044-012-0216-6
[16] AHER Y D, AGRAWAL A, BHARATAM P V, et al. 3D-QSAR studies of substituted 1-(3, 3-diphenylpropyl)-piperidinyl amides and ureas as CCR5 receptor antagonists[J]. Journal of Molecular Modeling, 2007, 13(4):519-529. doi: 10.1007/s00894-007-0173-z
[17] KIM M H, RYU J S, HAH J M. 3D-QSAR studies of 1, 2-diaryl-1H-benzimidazole derivatives as JNK3 inhibitors with protective effects in neuronal cells[J]. Bioorganic & Medicinal Chemistry Letters, 2013, 23(6):1639-1642. https://www.sciencedirect.com/science/article/pii/S0960894X13001157
[18] KOTHANDAN G, MADHAVAN T, GADHE C G, et al. A combined 3D QSAR and pharmacophore-based virtual screening for the identification of potent p38 MAP kinase inhibitors:an in silico approach[J]. Medicinal Chemistry Research, 2013, 22(4):1773-1787. doi: 10.1007/s00044-012-0179-7
[19] GHASEMI J B, AGHAEE E, JABBARI A. Docking, CoMFA and CoMSIA studies of a series of N-benzoylated phenoxazines and phenothiazines derivatives as antiproliferative agents[J]. Bulletin of the Korean Chemical Society, 2013, 34(3):899-906. doi: 10.5012/bkcs.2013.34.3.899
[20] CHEN J, YU R, SHEN B Z, et al. Docking-based 3D-QSAR modeling of the inhibitors of IMP metallo-β-lactamase[J]. Medicinal Chemistry Research, 2013, 22(4):1730-1739. doi: 10.1007/s00044-012-0172-1
[21] SUN J Y, HU M. Binding site analysis, 3D-QSAR studies, and molecular design of flavonoids derivatives as potent neuraminidase inhibitors[J]. Medicinal Chemistry Research, 2013, 22(2):606-614. doi: 10.1007/s00044-012-0054-6
[22] YONG D H, BAEK H S, CHO H, et al. 3D-QSAR study of adamantyl N-benzylbenzamides as melanogenesis inhibitors[J]. Bioorganic & Medicinal Chemistry Letters, 2014, 24(2):667-673. https://www.sciencedirect.com/science/article/pii/S0960894X13013474
[23] VYAS V K, GUPTA N, GHATE M. CoMFA and CoMSIA analysis of protein kinase B (PKB β) inhibitors using various alignment methods[J]. Medicinal Chemistry Research, 2013, 22(12):6046-6062. doi: 10.1007/s00044-013-0593-5
[24] HE Y W, NIU C W, LI H, et al. Experimental and computational correlation and prediction on herbicide resistance for acetohydroxyacid synthase mutants to Bispyribac[J]. Science China Chemistry, 2013, 56(3):286-295. doi: 10.1007/s11426-013-4841-9
[25] LIANG T G, REN L H, LI Q S. 3D-QSAR studies of tetraoxanes derivatives as antimalarial agents using CoMFA and CoMSIA approaches[J]. Bulletin of the Korean Chemical Society, 2013, 34(6):1823-1828. doi: 10.5012/bkcs.2013.34.6.1823
[26] ZENG G H, FANG D Q, WU W J, et al. Binding conformations, QSAR, and molecular design of Alkene-3-quinolinecarbonitriles as Src inhibitors[J]. International Journal of Quantum Chemistry, 2013, 113(10):1467-1478. doi: 10.1002/qua.v113.10
[27] CHO J E, KIM J T, JUNG S H, et al. Characterization of binding mode for human coagulation factor XI (FXI) inhibitors[J]. Bulletin of the Korean Chemical Society, 2013, 34(4):1212-1220. doi: 10.5012/bkcs.2013.34.4.1212
[28] ROFOUIE M K, SALAHINEJAD M, GHASEMI J B, et al. Comparative molecular field analysis and comparative molecular similarity index analysis studies on 1H NMR chemical shift of NH group of diaryl triazene derivatives[J]. Magnetic Resonance in Chemistry, 2013, 51(5):269-274. doi: 10.1002/mrc.v51.5
[29] UDDIN R, NAZ A, AKHTAR N, et al. Development of robust QSAR model using rapid overlay of crystal structures (ROCS) based alignment:a test case of Tubulin inhibitors[J]. Medicinal Chemistry Research, 2013, 22(7):3229-3241. doi: 10.1007/s00044-012-0327-0
[30] MADHAVAN T, GADHE C G, KOTHANDAN G, et al. Enhancement of P-gylcoprotein modulators of arylmethylaminephenyl derivatives:an integrative modeling approach[J]. Medicinal Chemistry Research, 2012, 22(22):2511-2523. doi: 10.1007%2Fs00044-012-0246-0.pdf
[31] WU X Y, WAN S H, ZHANG J J. Three dimensional quantitative structure-activity relationship of 5H-Pyrido[4, 3-b]indol-4-carboxamide JAK2 inhibitors[J]. International Journal of Molecular Sciences, 2013, 14(6):12037-12053. doi: 10.3390/ijms140612037
[32] JIMÉNEZ VILLALOBOS T P, GAITÁN I R, MONTALVO ACOSTA J J. 2D, 3D-QSAR and molecular docking of 4(1H)-quinolones analogues with antimalarial activities[J]. Journal of Molecular Graphics & Modelling, 2013, 46:105-124. https://www.sciencedirect.com/science/article/pii/S002228601731102X
[33] LI X L, FU J, SHI W, et al. 3D-QSAR and molecular docking studies on benzotriazoles as antiproliferative agents and histone deacetylase inhibitors[J]. Bulletin of the Korean Chemical Society, 2013, 34(8):2387-2393. doi: 10.5012/bkcs.2013.34.8.2387
[34] WANG J H, TANG K, HOU Q Q, et al. 3D-QSAR studies on C24-monoalkylated vitamin D 3, 26, 23-lactones and their C2α-modified derivatives with inhibitory activity to vitamin D receptor[J]. Molecular Informatics, 2010, 29(8/9):621-632. doi: 10.1021/jm060797q
[35] SRIVASTAVA V, GUPTA S P, SIDDIQI M I, et al. 3D-QSAR studies on quinazoline antifolate thymidylate synthase inhibitors by CoMFA and CoMSIA models[J]. European Journal of Medicinal Chemistry, 2010, 45(4):1560-1571. doi: 10.1016/j.ejmech.2009.12.065
[36] ZENG H H, ZHANG H B. Combined 3D-QSAR modeling and molecular docking study on 1, 4-dihydroindeno[1, 2-c]pyrazoles as VEGFR-2 kinase inhibitors.[J]. Journal of Molecular Graphics & Modelling, 2010, 29(1):54-71. https://www.sciencedirect.com/science/article/pii/S1093326310000562
[37] UL H Z, UDDIN R, WAI L K, et al. Docking and 3D-QSAR modeling of cyclin-dependent kinase 5/p25 inhibitors[J]. Journal of Molecular Modeling, 2011, 17(5):1149-1161. doi: 10.1007/s00894-010-0817-2
[38] HUANG X Y, SHAN Z J, ZHAI H L, et al. Molecular design of anticancer drug leads based on three-dimensional quantitative structure-activity relationship[J]. Journal of Chemical Information & Modeling, 2011, 51(8):1999-2006.
[39] LI P Z, TIAN Y L, ZHAI H L, et al. Study on the activity of non-purine xanthine oxidase inhibitor by 3D-QSAR modeling and molecular docking[J]. Journal of Molecular Structure, 2013, 1051(44):56-65. https://www.sciencedirect.com/science/article/pii/S0022286013006595
[40] MANVAR A T, PISSURLENKAR R R, VIRSODIA V R, et al. Synthesis, in vitro antitubercular activity and 3D-QSAR study of 1, 4-dihydropyridines[J]. Molecular Diversity, 2010, 14(2):285-305. doi: 10.1007/s11030-009-9162-8
[41] RAO G W, WANG C, WANG J, et al. Synthesis, structure analysis, antitumor evaluation and 3D-QSAR studies of 3, 6-disubstituted-dihydro-1, 2, 4, 5-tetrazine derivatives[J]. Bioorganic & Medicinal Chemistry Letters, 2013, 23(23):6474-6480. https://www.sciencedirect.com/science/article/pii/S0960894X13011104
[42] MORDE V A, SHAIKH M S, PISSURLENKAR R R S, et al. Molecular modeling studies, synthesis, and biological evaluation of Plasmodium falciparum, enoyl-acyl carrier protein reductase (Pf ENR) inhibitors[J]. Molecular Diversity, 2009, 13(4):501-517. doi: 10.1007/s11030-009-9141-0
[43] DU J, LEI B L, QIN J, et al. Molecular modeling studies of vascular endothelial growth factor receptor tyrosine kinase inhibitors using QSAR and docking[J]. Journal of Molecular Graphics & Modelling, 2009, 27(5):642-654. https://www.sciencedirect.com/science/article/pii/S1093326308001459
[44] NAYANA R S, BOMMISETTY S K, SINGH K, et al. Structural analysis of carboline derivatives as inhibitors of MAPKAP K2 using 3D QSAR and docking studies[J]. Journal of Chemical Information & Modeling, 2009, 49(1):53-67. http://www.oalib.com/paper/157157
[45] MUDDASSAR M, PASHA F A, NEAZ M M, et al. Elucidation of binding mode and three dimensional quantitative structure-activity relationship studies of a novel series of protein kinase B/Akt inhibitors[J]. Journal of Molecular Modeling, 2009, 15(2):183-192. doi: 10.1007/s00894-008-0416-7
[46] PRASANNA S, DAGA P R, XIE A, et al. Glycogen synthase kinase-3 inhibition by 3-anilino-4-phenylmaleimides:insights from 3D-QSAR and docking[J]. Journal of Computer-Aided Molecular Design, 2009, 23(2):113-127. doi: 10.1007/s10822-008-9244-1
[47] NEAZ M M, PASHA F A, MUDDASSAR M, et al. Pharmacophore based 3D-QSAR study of VEGFR-2 inhibitors[J]. Medicinal Chemistry Research, 2009, 18(2):127-142. doi: 10.1007/s00044-008-9113-4
[48] LU P, WEI X, ZHANG R S. CoMFA and CoMSIA studies on HIV-1 attachment inhibitors[J]. European Journal of Medicinal Chemistry, 2010, 45(5):1792-1798. doi: 10.1016/j.ejmech.2010.01.011
[49] CHAUDHAERY S S, ROY K K, SAXENA A K. Consensus superiority of the pharmacophore-based alignment, over maximum common substructure (MCS):3D-QSAR studies on carbamates as acetylcholinesterase inhibitors[J]. Journal of Chemical Information & Modeling, 2009, 49(6):1590-1601. doi: 10.1007/s00044-012-0465-4
[50] FREITAS G B L D, SILVA L L D, ROMEIRO N C, et al. Development of CoMFA and CoMSIA models of affinity and selectivity for indole ligands of cannabinoid CB1 and CB2 receptors[J]. European Journal of Medicinal Chemistry, 2009, 44(6):2482-2496. doi: 10.1016/j.ejmech.2009.01.026
[51] BASU A, JASU K, JAYAPRAKASH V, et al. Development of CoMFA and CoMSIA models of cytotoxicity data of anti-HIV-1-phenylamino-1H-imidazole derivatives[J]. European Journal of Medicinal Chemistry, 2009, 44(6):2400-2407. doi: 10.1016/j.ejmech.2008.09.043
[52] WEBER A, BÖHM M, SUPURAN C T, et al. 3D QSAR selectivity analyses of carbonic anhydrase inhibitors:insights for the design of isozyme selective inhibitors[J]. Journal of Chemical Information & Modeling, 2006, 46(6):2737-2760. http://www.academia.edu/13734752/Homology_Modeling_and_Receptor-Based_3D-QSAR_Study_of_Carbonic_Anhydrase_IX
[53] ZHENG M Y, YU K Q, LIU H, et al. QSAR analyses on avian influenza virus neuraminidase inhibitors using CoMFA, CoMSIA, and HQSAR[J]. Journal of Computer-Aided Molecular Design, 2006, 20(9):549-566. doi: 10.1007/s10822-006-9080-0
[54] LI W, TANG Y, XIE Q, et al. 3D-QSAR studies of orvinol analogs as κ-opioid agonists[J]. Journal of Molecular Modeling, 2006, 12(6):877-884. doi: 10.1007/s00894-005-0084-9
[55] PATEL M R, DIMMOCK J R, TALELE T T. CoMFA and CoMSIA studies on 1, 3-Bis(benzylidene)-3, 4-dihydro-1H-naphthalen-2-one, 2, 6-Bis(benzylidene)cyclohexanone, and 3, 5-Bis(benzylidene)-4-piperidone series of cytotoxic compounds[J]. Journal of Chemical Information & Modeling, 2007, 47(6):2110-2123.
[56] HE Y Z, LI Y X, ZHU X L, et al. Rational design based on bioactive conformation analysis of pyrimidinylbenzoates as acetohydroxyacid synthase inhibitors by integrating molecular docking, CoMFA, CoMSIA, and DFT calculations[J]. Journal of Chemical Information & Modeling, 2007, 47(6):2335-2344. doi: 10.1021/ci7002297
[57] PISSURLENKAR R R S, SHAIKH M S, COUTINHO E C. 3D-QSAR studies of Dipeptidyl peptidase IV inhibitors using a docking based alignment[J]. Journal of Molecular Modeling, 2007, 13(10):1047-1071. doi: 10.1007/s00894-007-0227-2
[58] LEE J Y, DODDAREDDY M R, CHO Y S, et al. Comparative QSAR studies on peptide deformylase inhibitors[J]. Journal of Molecular Modeling, 2007, 13(5):543-558. doi: 10.1007/s00894-007-0175-x
[59] JUAN A A S, CHO S J. 3D-QSAR study of microsomal prostaglandin E 2, synthase(mPGES-1) inhibitors[J]. Journal of Molecular Modeling, 2007, 13(5):601-610. doi: 10.1007/s00894-007-0172-0
[60] ZAHEERULHAQ, UDDIN R, YUAN H, et al. Receptor-based modeling and 3D-QSAR for a quantitative production of the butyrylcholinesterase inhibitors based on genetic algorithm[J]. Journal of Chemical Information & Modeling, 2008, 48(5):1092-1103. doi: 10.1021/ci8000056
[61] LUO X Y, SHU M, WANG Y Q, et al. 3D-QSAR studies of dihydropyrazole and dihydropyrrole derivatives as inhibitors of human mitotic kinesin Eg5 based on molecular docking[J]. Molecules, 2012, 17(2):2015-2029. https://www.sciencedirect.com/science/article/pii/S1476927115302644
[62] ZHANG Y M, LIU H C, JIAO Y, et al. De novo design of N-(pyridin-4-ylmethyl)aniline derivatives as KDR inhibitors:3D-QSAR, molecular fragment replacement, protein-ligand interaction fingerprint, and ADMET prediction.[J]. Molecular Diversity, 2012, 16(4):787-802. doi: 10.1007/s11030-012-9405-y
[63] JING P, ZHAO S J, JIAN W J, et al. Quantitative studies on structure-DPPH· scavenging activity relationships of food phenolic acids[J]. Molecules, 2012, 17(11):12910-12924. http://www.oalib.com/paper/164674
[64] GAO J, CHENG Y H, CUI W, et al. 3D-QSAR and molecular docking studies of hydroxamic acids as peptide deformylase inhibitors[J]. Medicinal Chemistry Research, 2011, 21(8):1597-1610. doi: 10.1007/s00044-011-9672-7.pdf
[65] UL-HAQ Z, KHAN W, ZIA S R, et al. Structure-based 3D-QSAR models and dynamics analysis of novel N-benzyl pyridinone as p38α MAP kinase inhibitors for anticytokine activity[J]. Journal of Molecular Graphics & Modelling, 2012, 36:48-61. https://www.sciencedirect.com/science/article/pii/S1093326312000204
[66] BHATT H G, PATEL P K. Pharmacophore modeling, virtual screening and 3D-QSAR studies of 5-tetrahydroquinolinylidine aminoguanidine derivatives as sodium hydrogen exchanger inhibitors[J]. Bioorganic & Medicinal Chemistry Letters, 2012, 22(11):3758-3765. https://www.sciencedirect.com/science/article/pii/S0960894X12004519
[67] WU B J, WANG X Q, ZHANG S X. Accurate prediction of glucuronidation of structurally diverse phenolics by human UGT1A9 using combined experimental and in silico approaches[J]. Pharmaceutical Research, 2012, 29(6):1544-1561. doi: 10.1007/s11095-012-0666-z
[68] MADHAVAN T, KOTHANDAN G, GADHE C G, et al. QSAR analysis on PfPK7 inhibitors using HQSAR, CoMFA, and CoMSIA[J]. Medicinal Chemistry Research, 2011, 21(6):681-693. doi: 10.1007/s00044-011-9572-x.pdf
[69] XIA J, LI J, SUN H. Insights into ET(A) subtype selectivity of benzodiazepine endothelin receptor antagonists by 3D-QSAR approaches[J]. Journal of Molecular Modeling, 2012, 18(4):1299-1311. doi: 10.1007/s00894-011-1153-x
[70] LI Y, HAO M, REN H, et al. Exploring the structure requirement for PKCθ inhibitory activity of pyridinecarbonitrile derivatives:an in silico analysis[J]. Journal of Molecular Graphics & Modelling, 2012, 34(34):76-88. https://www.sciencedirect.com/science/article/pii/S1093326311001859
[71] WU X Y, WU S G, CHEN W H. Molecular docking and 3D-QSAR study on 4-(1 H -indazol-4-yl) phenylamino and aminopyrazolopyridine urea derivatives as kinase insert domain receptor (KDR) inhibitors[J]. Journal of Molecular Modeling, 2012, 18(3):1207-1218. doi: 10.1007/s00894-011-1146-9
[72] CHEN N, LIU C K, ZHAO L Z, et al. 3D-QSAR study of multi-target-directed AchE inhibitors based on autodocking[J]. Medicinal Chemistry Research, 2012, 21(2):245-256. doi: 10.1007/s00044-010-9516-x
[73] VERMA S M, RAZDAN B K, SASMAL D. 3D-QSAR study of 8-azabicyclo[3.2.1] octane analogs antagonists of cholinergic receptor[J]. Bioorganic & Medicinal Chemistry Letters, 2009, 19(11):3108-3112. https://www.sciencedirect.com/science/article/pii/S0960894X09004843
[74] LIU H Y, LIU S S, QIN L T, et al. CoMFA and CoMSIA analysis of 2, 4-thiazolidinediones derivatives as aldose reductase inhibitors[J]. Journal of Molecular Modeling, 2009, 15(7):837-845. doi: 10.1007/s00894-008-0439-0
[75] QIN J, LIU H X, LI J Z, et al. 3D-QSAR studies on the inhibitors of AP-1 and NF-κB mediated transcriptional activation[J]. European Journal of Medicinal Chemistry, 2009, 44(7):2888-2895. doi: 10.1016/j.ejmech.2008.12.006
[76] CHEN Y D, LI H F, TANG W Q, et al. 3D-QSAR studies of HDACs inhibitors using pharmacophore-based alignment[J]. European Journal of Medicinal Chemistry, 2009, 44(7):2868-2876. doi: 10.1016/j.ejmech.2008.12.008
[77] PASHA F A, MUDDASSAR M, NEAZ M M, et al. Pharmacophore and docking-based combined in-silico study of KDR inhibitors[J]. Journal of Molecular Graphics & Modelling, 2009, 28(1):54-61. https://www.sciencedirect.com/science/article/pii/S1093326309000448
[78] PASHAA F A, CHOA S J, BEG Y. 3D and quantum QSAR of non-benzodiazepine compounds[J]. European Journal of Medicinal Chemistry, 2008, 43(11):2361-2372. doi: 10.1016/j.ejmech.2008.01.030
[79] DU J, QIN J, LIU H X, et al. 3D-QSAR and molecular docking studies of selective agonists for the thyroid hormone receptor β[J]. Journal of Molecular Graphics & Modelling, 2008, 27(2):95-104. https://www.sciencedirect.com/science/article/pii/S1093326308000405
[80] ROY K K, DIXIT A, SAXENA A K. An investigation of structurally diverse carbamates for acetylcholinesterase (AChE) inhibition using 3D-QSAR analysis[J]. Journal of Molecular Graphics & Modelling, 2008, 27(2):197-208. doi: 10.1007/s00894-015-2797-8
[81] LEI B L, DU J, LI S Y, et al. Comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) of thiazolone derivatives as hepatitis C virus NS5B polymerase allosteric inhibitors[J]. Journal of Computer-Aided Molecular Design, 2008, 22(10):711-725. doi: 10.1007/s10822-008-9230-7
[82] NAYANA M R S, SEKHAR Y N, NANDYALA H, et al. Insight into the structural requirements of proton pump inhibitors based on CoMFA and CoMSIA studies[J]. Journal of Molecular Graphics & Modelling, 2008, 27(3):233-243. https://www.sciencedirect.com/science/article/pii/S1093326308000569
[83] SEKHAR Y N, NAYANA M R S, SIVAKUMARI N, et al. 3D-QSAR and molecular docking studies of 1, 3, 5-triazene-2, 4-diamine derivatives against r-RNA:novel bacterial translation inhibitors[J]. Journal of Molecular Graphics & Modelling, 2008, 26(8):1338-1352. https://www.sciencedirect.com/science/article/pii/S1093326308000156
[84] ZHI J S, HONG L Z, XIAO Y H, et al. Molecular design of new aggrecanases-2 inhibitors[J]. Bioorganic & Medicinal Chemistry Letters, 2013, 23(19):5339-5350. https://www.sciencedirect.com/science/article/pii/S0960894X13009050
[85] LI P Z, TIAN Y L, ZHAI H L, et al. Study on the activity of non-purine xanthine oxidase inhibitor by 3D-QSAR modeling and molecular docking[J]. Journal of Molecular Structure, 2013, 1051(44):56-65. https://www.sciencedirect.com/science/article/pii/S0022286013006595
[86] QIAN C W, ZHENG J X, XIAO G K, et al. 3D-QSAR studies on thiazolidin-4-one S1P1 receptor agonists by CoMFA and CoMSIA[J]. International Journal of Molecular Sciences, 2011, 12(10):6502-6516. https://core.ac.uk/display/8639720
计量
- 文章访问数: 181
- HTML全文浏览量: 1
- PDF下载量: 41
