Skip to main content
QUICK REVIEW

[论文解读] Implementing measurement error models with mechanistic mathematical models in a likelihood-based framework for estimation, identifiability analysis, and prediction in the life sciences

Ryan J. Murphy, Oliver J. Maclaren|arXiv (Cornell University)|Jul 4, 2023
Gene Regulatory Network AnalysisBiochemistry, Genetics and Molecular Biology被引用 3
一句话总结

本文提出了一种基于似然的框架,用于将多种测量误差模型(如加法高斯、乘法对数正态和泊松)整合到系统生物学与数学生物学中的机制性常微分方程(ODE)和偏微分方程(PDE)模型中。该框架展示了如何通过轮廓似然分析实现稳健的参数估计、可辨识性评估以及在不同误差假设下的预测,配套提供了开源的 Julia 代码以支持可复现性。

ABSTRACT

Throughout the life sciences we routinely seek to interpret measurements and observations using parameterised mechanistic mathematical models. A fundamental and often overlooked choice in this approach involves relating the solution of a mathematical model with noisy and incomplete measurement data. This is often achieved by assuming that the data are noisy measurements of the solution of a deterministic mathematical model, and that measurement errors are additive and normally distributed. While this assumption of additive Gaussian noise is extremely common and simple to implement and interpret, it is often unjustified and can lead to poor parameter estimates and non-physical predictions. One way to overcome this challenge is to implement a different measurement error model. In this review, we demonstrate how to implement a range of measurement error models in a likelihood-based framework for estimation, identifiability analysis, and prediction, called Profile-Wise Analysis. This frequentist approach to uncertainty quantification for mechanistic models leverages the profile likelihood for targeting parameters and understanding their influence on predictions. Case studies, motivated by simple caricature models routinely used in systems biology and mathematical biology literature, illustrate how the same ideas apply to different types of mathematical models. Open-source Julia code to reproduce results is available on GitHub.

研究动机与目标

  • 解决生命科学建模中广泛存在但往往缺乏依据的加法高斯测量误差模型的使用问题。
  • 证明替代性测量误差模型(如乘法型、泊松型)可提供更准确且更符合物理实际的参数估计与预测结果。
  • 提供一种实用且统一的框架,用于在基于似然的推断方法下实现和比较不同的测量误差模型。
  • 通过轮廓似然方法,实现对具有多样化误差结构的机制性模型的结构可辨识性与实际可辨识性分析。
  • 通过提供开源的 Julia 代码,支持研究的可复现性,涵盖所有模型实现与分析。

提出的方法

  • 构建一个似然函数,通过多种测量误差模型(如加法高斯、乘法对数正态、泊松等)将机制性 ODE/PDE 模型的解与观测数据关联起来。
  • 利用 Julia 中的 Distributions 包提供的对数似然函数,计算在不同误差假设下的模型拟合度。
  • 采用数值优化方法(Nelder-Mead)最大化似然函数,以估计模型参数。
  • 进行轮廓对数似然分析,以评估实际可辨识性并计算参数的置信区间。
  • 使用卡方分布的临界值来定义基于似然比检验的近似置信区间。
  • 使用 Julia 中的 DifferentialEquations.jl 包对 ODE/PDE 系统进行数值求解。
Figure 1: Implementing a variety of measurement error models in a profile likelihood-based framework for parameter estimation, identifiability analysis, and prediction. (a) Synthetic data (circles). (b) The framework is applicable to a range of mathematical models and measurement error models. Schem
Figure 1: Implementing a variety of measurement error models in a profile likelihood-based framework for parameter estimation, identifiability analysis, and prediction. (a) Synthetic data (circles). (b) The framework is applicable to a range of mathematical models and measurement error models. Schem

实验结果

研究问题

  • RQ1测量误差模型的选择如何影响生物系统机制性模型中的参数估计与预测精度?
  • RQ2当数据呈现异方差性或为计数数据时,乘法对数正态或泊松误差模型是否能比标准的加法高斯假设更好地捕捉生物变异性?
  • RQ3在使用非高斯测量误差模型时,如何利用轮廓似然分析评估实际可辨识性?
  • RQ4在系统生物学应用中,使用离散型或异方差型误差模型会带来哪些计算与统计影响?
  • RQ5该似然框架在多大程度上可推广至不同类型的机制性模型(ODE、PDE)及不同的误差结构?

主要发现

  • 当真实数据生成过程不符合加法高斯误差假设时,加法高斯误差模型常导致参数估计不佳且产生非物理的预测结果。
  • 当数据呈现异方差性或为计数型数据时,乘法对数正态和泊松误差模型能提供更合适的拟合效果。
  • 轮廓似然分析即使在复杂误差结构下也能成功识别不可辨识参数,从而提升模型的可靠性。
  • 该框架通过基于似然比检验的阈值方法,实现了参数的精确置信区间估计,具体提供了卡方分布分位数(如 χ²(ν) 用于 90%、95%、99%、99.9% 置信区间的临界值)。
  • 该方法计算高效且可扩展,支持多种分布(如伽马分布、β分布、负二项分布)及不同的误差结构。
  • 开源的 Julia 代码已发布于 GitHub,支持完整可复现性,并可轻松扩展至新的生物模型与误差模型。
Figure 2: Caricature ODE model with linear reactions (Eq ( 16 )). (a) Synthetic data (circles) at sixteen equally–spaced time points from $t=0.0$ to $t=2.0$ are generated by simulating Eq ( 16 ), the additive Gaussian measurement error model, known model parameters $\theta=(r_{1},r_{2},\sigma_{\math
Figure 2: Caricature ODE model with linear reactions (Eq ( 16 )). (a) Synthetic data (circles) at sixteen equally–spaced time points from $t=0.0$ to $t=2.0$ are generated by simulating Eq ( 16 ), the additive Gaussian measurement error model, known model parameters $\theta=(r_{1},r_{2},\sigma_{\math

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。