Plumbridge T W, Aarons L J, Brown J R
J Pharm Pharmacol. 1978 Feb;30(2):69-74. doi: 10.1111/j.2042-7158.1978.tb13164.x.
In the analysis of binding data, arbitrary transformations such as the Scatchard plot, may give misleading estimates of the binding parameters. The statistically correct approach is to determine values of K and n by non-linear regression of the actual dependent variable against the actual independent variable. In the case of the spectrophotometric titration method the dependent variable is the absorbance and the independent variable is the composition of the drug/macromolecule mixture. The method relies on an accurate estimate of the extinction coefficient of the bound drug and this is best treated as a parameter to be estimated in the regression analysis. In testing models by data fits alone it is emphasized that whilst a model may be rejected if it does not fit the data, a good fit does not ensure uniquieness and confirmatory, independent evidence must be sought.
在结合数据的分析中，诸如Scatchard图等任意变换可能会给出有误导性的结合参数估计值。统计学上正确的方法是通过将实际因变量与实际自变量进行非线性回归来确定K和n的值。对于分光光度滴定法，因变量是吸光度，自变量是药物/大分子混合物的组成。该方法依赖于对结合药物消光系数的准确估计，而这最好作为回归分析中要估计的一个参数来处理。仅通过数据拟合来测试模型时，需要强调的是，虽然一个模型如果不拟合数据可能会被拒绝，但拟合良好并不能确保唯一性，必须寻求确证性的独立证据。
