摘要
通常蛋白质的热变性是一个两态过程，理论上可用非线性方程来表示．在求其热变性的速度常数和变性反应的活化能中，一般将其转化成一个线性问题，用最小二乘法来解决．这样在转化中会对估计参数产生外加的偏差．用模式搜索法求多元函数极值，直接得到最佳拟合及相应参数．并在此基础上，用计算机模拟进行数值实验，为耐热性测定提供了一个更加精确而有效的实验方案．
The thermal denaturation of proteins is usually recognized as a two-state process which follows a nonlinear equation. The equation can be conventionally transformed into a linear model, and the velocity constant and activation energy of thermal denaturation were estimated using the principle of the least square method. However, it will bring some extra error on the parameter estimation. In this paper, the extreme value of the multivariate nonlinear function was computed by the pattern search method, and the best-fitted parameters were obtained directly. Experimental designs for an efficient and accurate measure of protein thermostability were optimized based on computer simulation studies.
出处
《复旦学报（自然科学版）》
CAS
CSCD
北大核心
1998年第4期462-468,共7页
Journal of Fudan University：Natural Science
基金
国家自然科学基金
关键词
耐热性
非线性
最小二乘法
蛋白质
数值计算
thermostability
nonlinear least square method
pattern search method
