FitSpace Explorer: An algorithm to evaluate multidimensional parameter space in fitting kinetic data
FitSpace Explorer: An algorithm to evaluate multidimensional parameter space in fitting kinetic data
复制标题
DOI：
10.1016/j.ab.2008.12.025
复制
发表时间：
2009-04-01
影响因子：
2.9
通讯作者：
Blom, Thomas
中科院分区：
文献类型：
--
作者：
Johnson, Kenneth A.;Simpson, Zachary B.;Blom, Thomas
直接将多组动力学数据拟合到基于数值积分的模型中，提供了在不依赖于获得速率方程解析解所需的简化假设的情况下提取动力学参数的最佳方法。然而，现代计算机程序使得输入一个过于复杂的模型变得过于容易，并且当一个系统约束不足时，标准误差分析会严重低估误差，并且无法揭示多个参数通过动力学数据中常见的复杂关系相互关联的全部程度。在此，我们描述了置信轮廓分析的应用，该分析是通过测量平方和误差对每一对参数的依赖性来获得的，同时允许所有其余参数在寻求最佳拟合时进行调整。置信轮廓揭示了参数之间的复杂关系，并清晰地勾勒出参数可以变化的空间（“拟合空间”）。通过对色氨酸合酶已发表数据的良好约束拟合以及寡核苷酸与核酶结合动力学的实例，说明了该方法的实用性。相比之下，对丙氨酸消旋酶的分析明确反驳了全局分析进程曲线可用于提取酶催化反应的自由能曲线的说法。（C）2008爱思唯尔公司。保留所有权利。
Fitting several sets of kinetic data directly to a model based on numerical integration provides the best method to extract kinetic parameters without relying on the simplifying assumptions required to achieve analytical Solutions of rate equations. However, modern computer programs make it too easy to enter an overly complex model, and standard error analysis grossly underestimates errors when a system is underconstrained and fails to reveal the full degree to which multiple parameters are linked through the complex relationships common in kinetic data. Here we describe the application of confidence contour analysis obtained by measuring the dependence of the sum square error on each pair of parameters while allowing all remaining parameters to be adjusted in seeking the best fit. The confidence contours reveal complex relationships between parameters and clearly outline the space over which parameters can vary (the "FitSpace"). The utility of the method is illustrated by examples of well-constrained fits to published data on tryptophan synthase and the kinetics of oligonucleotide binding to a ribozyme. In contrast, analysis of alanine racemase clearly refutes claims that global analysis of progress curves can be used to extract the free energy profiles of enzyme-catalyzed reactions. (C) 2008 Elsevier Inc. All rights reserved.