定量药物设计最佳回归方程计算中剔除样本的逐步回归迭代法

Rejection of Samples by Repeated Replacement Method in the Computation of Optimum Regressive Equations in Quantitative Drug Design

在新药物研究中,用定量药物设计的方法来缩短合成和筛选的过程,已经得到越来越多的研究工作者的赞同。在进行定量药物设计时。

In a quantitative drug design of 1, 7, 8-trisubstituted 6-fluoro-1, 4-dihydro-4-oxo-3-quinolinecarboxylie acid antibacterials, 216 compounds (samples) having in vitro inhibiting E. coli NIHJ JC-2 data [minimum inhibiting (100%) concentration, MIC] from the literature were used. The MIC′s were determined by different laboratories and there may be considerable errors. In order to reje(?)t those samples with larger error, a progressive regression analysis was carried out with those 216 samples and 72 indices (71 physico-chemical parameters and a biological parameter MIC, Y) regressive equation was established. Then the values of concerned physieo-chemical parmeters of each the 216 samples were put into the equation. A critical value α was chosen to reject the samples having large difference in computed and determined Y values. Again another progressive regression analysis was carried out with the unrejceted samples and the 72 indices. A new optimum regressive equation was established. The values of concerned physico-chemical parameters of each of 216 samples were put into this new equation and the samples having larger difference in computed and determined Y values (than the same critical value α) were rejceted. This same procedure was carried out again till the samples rejected by two last processes were completely identical. Then the process of rejecting samples was ended.