Institute of Geology and Geophysics, 100041, Khodzhibaev St. 49, Tashkent, Uzbekistan.
Eur J Med Chem. 2010 Apr;45(4):1387-94. doi: 10.1016/j.ejmech.2009.12.037. Epub 2010 Jan 6.
The International Chemical Identifier (InChI) has been used to construct InChI-based optimal descriptors to model the binding affinity for fullerene[C60]-based inhibitors of human immunodeficiency virus type 1 aspartic protease (HIV-1 PR). Statistical characteristics of the one-variable model obtained by the balance of correlations are as follows: n=8, r(2)=0.9769, q(2)(LOO)=0.9646, s=0.099, F=254 (subtraining set); n=7, r(2)=0.7616, s=0.681, F=16 (calibration set); n=5, r(2)=0.9724, s=0.271, F=106, R(m)(2)=0.9495 (test set). Predictability of this approach has been checked with three random splits of the data: into the subtraining set, calibration set, and test set.
国际化学标识符 (InChI) 已被用于构建基于 InChI 的最优描述符,以模拟对人类免疫缺陷病毒 1 天冬氨酸蛋白酶 (HIV-1 PR) 的富勒烯 [C60] 基抑制剂的结合亲和力。通过相关性平衡获得的单变量模型的统计特征如下:n=8,r(2)=0.9769,q(2)(LOO)=0.9646,s=0.099,F=254(子训练集);n=7,r(2)=0.7616,s=0.681,F=16(校准集);n=5,r(2)=0.9724,s=0.271,F=106,R(m)(2)=0.9495(测试集)。通过将数据随机分为子训练集、校准集和测试集三种方式来检查该方法的可预测性。