论文部分内容阅读
利用遗传算法,并结合线性回归和交叉验证方法,对一系列43个苯并呋喃/噻吩联二苯类PTP1B抑制剂作了二维定量构效关系的研究.计算得到了一组效果较好的定量构效关系模型.模型不仅具有好的回归能力,而且还具有很好的预测能力.同时,通过分析在遗传优化过程中参数在精华种群中所占的比例,还得到了可能对活性影响较大的成分.计算结果表明,分子的4个参数:lgP(分配系数)、Area(表面积)、MW(分子量)以及Dip(偶极距)是影响化合物活性的最重要的参数,这对抑制剂的设计和改造提供了指导.
A series of 43 benzofuran / thiophene-biphenyl PTP1B inhibitors were studied by means of genetic algorithms, combined with linear regression and cross-validation methods. The results of two-dimensional quantitative structure-activity relationships Quantitative structure-activity relationship model.The model not only has good regression ability, but also has good predictive ability.At the same time, by analyzing the proportion of the parameters in the elite population in the process of genetic optimization, The results show that the four parameters of the molecule: lgP (partition coefficient), Area (surface area), MW (molecular weight) and Dip (dipole moment) are the most important parameters affecting the activity of a compound, The design and remodeling provided guidance.