两种参数估计方法在ELISA检测结果标准化中的应用(英文)

Application of two methods of parameter estimation in the standardization of ELISA testing results

目的探讨在运用改良光密度值标准化方法进行的ELISA检测结果标准化过程中,两种不同参数估计方法对标准化结果造成的差异。方法标准化公式中的参数估计分别采用线性最小二乘法和非线性最小二乘法中的高斯牛顿迭代法。结果通过线性最小二乘估计得到的标准曲线很大程度上由浓度较低的几个点所决定,而非线性最小二乘法确定的标准曲线总体上更接近实验数据,真实体现了不同实验板的反应动力学特性。结论非线性最小二乘法更适用于标准化过程中的参数估计,有效降低了检测结果在板间的变异。

Objective Investigate differences between two methods of parameter estimation employed in the standardization of ELISA testing results with the improved standardization method for optical density（I-STOD）. Methods Linear least square（LLS） and Gauss-Newton iteration from nonlinear least square（NLS） were employed for the parameter estimation of the formulas. Results Calibration curve derived through the LLS is plotted by several points which present relatively low antibody level to a great extent. In contrast, the curve derived through the NLS was more conformed to the real data as a whole and better represented what reacted on experimental plates. Conclusion NLS is more eligible for the parameter estimation during the standardization of ELISA testing results and effectively reduced the variation between plates.