两种参数估计方法在ELISA检测结果标准化中的比较  

Comparison between Two Methods of Parameter Estimation Applied in Standardization of ELISA Testing Results

作　　者：屠小明[1] 罗建平[1] 吴海玮[2] 

机构地区：[1]南京医科大学基础医学院,南京210029 [2]美国罗德岛医院国际卫生研究中心

出　　处：《数理医药学杂志》2013年第6期660-662,共3页Journal of Mathematical Medicine

摘　　要：探讨了在ELISA法检测结果标准化过程中,不同参数估计方法对标准化结果的影响。标准化方法采用改良光密度值标准化变换(I-STOD),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中的高斯牛顿迭代法。结果显示线性最小二乘估计所得到的标准曲线主要由浓度较低的几个点所决定,而非线性最小二乘所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。非线性最小二乘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。The impact of different parameter estimation methods on standardized results was explored during standardization of testing results from ELISA. ＂Improved Standardization method for Optical Density＂ （I-STOD） was applied to realize standardization. Linear least square （LLS） and Gauss-Newton iteration from nonlinear least square （NLS） were used to estimate parameter in I-STOD formula. Results indicated that cal- ibration curve generated by LLS was lagrely depended on the several initial points. Calibration curve genera- ted by NLS was more close to real data. It was a better representation of what reacted on experimental plates. Therefore, NLS was more appropriately for parameter estimation in standardization of ELISA testing results and variation between plates was effectively decreased.

关 键 词：参数估计 标准化 拟合优度 非线性最小二乘 

分 类 号：R311[医药卫生—基础医学]