摘要
目的考察SPSS和Excel对血清乙型肝炎病毒(HBV)大蛋白(-LP)浓度标准曲线的拟合效果。方法HBV-LP标准品吸光度检测采用酶联免疫吸附试验,分别用SPSS软件曲线估计和Excel软件规划求解对HBV-LP标准品浓度与吸光度进行线性模型、对数线性模型、二次多项式模型、三次多项式模型的曲线拟合,比较两种拟合方法的各回归模型拟合效果的一致性,并根据各回归模型决定系数的大小来优选血清HBV-LP浓度标准曲线回归模型。结果HBV-LP标准品浓度与吸光度的散点图呈非线性趋势;两种拟合方法的线性模型、对数线性模型、二次多项式模型、三次多项式模型的回归方程均有意义(P<0.001),其决定系数小数点后四位是一致的;其中二次多项式模型、三次多项式模型的拟合精度较高,决定系数均大于0.95。结论Excel软件规划求解拟合血清HBV-LP浓度标准曲线的效果与SPSS高度一致,是一种临床实验室定量标准曲线拟合与优选的有效方法。
Objective To understand the fitting effects of SPSS and Excel tor concentration standard curve of serum hepatitis B virus large surface protein （HBV-LP）. Methods Enzyme-linked immunosorbent assay （ELISA）was used to measure the absorbance of standard preparation of HBV-LP. Concentration and absorbance of standard preparation of HBV-LP was carried out curve fitting with linear, log-linear, quadratic polynomial and cubic polynomial models by curve estimation of SPSS and program solution of Excel, respectively. It was compared to concordance of fitting effects between SPSS and Excel. The regression model of concentration standard curve of serum HBV-LP was optimizated with its coefficient of determination. Results The scatterplot of standard preparation of HBV-LP subnfited non-linear tendency. There were all significance to regression equation of linear model, log-linear model, quadratic polynomial model and cubic polynomial model（P〈0.001）. Those coefficient of determination was concordant to 4 decimal places with SPSS and Excel. It was well to precision of fitting of quadratic polynomial model and cubic polynomial model. Its coefficient of determination all exceeded 0.95. Conclusion Program solution of Excel is a modus operandi to fit and optimizate concentration standard curve of serum HBV-LP in clinical laboratory, and its fitting effects is supreme coincidence with that of SPSS.
出处
《实验与检验医学》
CAS
2009年第3期228-231,共4页
Experimental and Laboratory Medicine
