摘要
对未知受体结构的药物设计其主导方法CoMFA来说,柔性目标分子的多种构象造成了问题的复杂性.本文介绍交叉验证参数R^2(q^2)引导的构象选择CoMFA方法,选择化合物的最佳构象.将一组47个HIV-1 RT抑制剂进行有、无构象选择的CoMFA分析来作评价.根据化合物的活性、毒性、选择性指数(毒性/活性比)等实验数据构建得到的模型,其交叉验证参数q^2为0.7以上,非交叉验证的相应参数为0.94以上,最后,还经过试验集化合物验证该模型的预测能力,置信度(1-α)>0.99.
Me comparative molecular field analysis (CoMFA) as a main type of drug design methods in the case of unknown 3D structures of the receptor is now facing complexity due, to the existence of many conformations for flexible target molecules. A cross-validated R-2(q(2)) guided conformation selection approach of CoMFA studies was proposed. As an example, a set of 47 HfV-1 RT inhibitors was investigated by this. approach. In comparison with. the conventional CoMFA, the result of the conformation selection approach for prediction of activity, toxicity and selective index gave much better results with q(2) > 0.7 and the non-cross-validated R-2 > 0.94. At last, this model was tested with a test set of 12 compounds showing good prediction ability with confidence level (1 - alpha) > 0.99.
出处
《化学学报》
SCIE
CAS
CSCD
北大核心
2003年第7期1121-1128,共8页
Acta Chimica Sinica
基金
中国科学院2002年度创新工程资助项目。
