【摘要】 用分子力学、模拟退火分子动力学、量子化学等方法对转化生长因子-β受体激酶(TβR-Ⅰ)的取代吡唑类抑制剂进行了结构优化.用遗传算法(GFA)并结合多元线性回归技术(MLR),对该类抑制剂进行定量构效关系研究,筛选出了影响抑制活性的主要因素,建立定量构效关系方程(QSAR).结果表明,分子的总偶极矩的Y分量μy、分子的最低空轨道(LUMO)与最高占据轨道(HOMO)的能量差?E、Jurs的分子中原子正电荷加权表面积分率FPSA3等是影响该类化合物抑制活性的主要因素.所得模型对该类化合物关于TβR-Ⅰ的抑制活性有较好的预测效果(R=0.898,R2cv=0.642,s=0.263,F=11.105).更多还原

【Abstract】 The geometry structure of pyrazole-based inhibitors of the transformation growth factor-β type Ⅰ receptor kinase domain（TβR-Ⅰ） was optimized through the methods of quantum chemistry,molecular mechanism and molecular anneal dynamics.The quantitative structure-activity relationship of these inhibitors,in regard to TβR-Ⅰ,was systematically studied using the genetic function approximation（GFA） and multiple linear regression（MLR）.Some main independent factors affecting the activity of the compounds were selected out,and then the QSAR equation was established.It was found that the y component of total dipolemoment of molecule(μ_y),the energy gap（⊿E） between the lowest unoccupied molecular orbital（LUMO） and the highest occupied orbital（HOMO）,Jurs descriptor FPSA3-Fractional atomic charge-weighted positive surface area etc.,are the main independent factors contributing to the inhibiting activity of the compounds.The fitting correlation coefficient R,the cross-validation R²_cv,the standard error of the regression model s and F test value F for the model established by this study are 0.898,0.642,0.263 and 11.105,respectively.These results suggest that this model has good predictability.更多还原