Sensitivity Measures for Oscillating Systems: Application to Mammalian Circadian Gene Network

Stephanie R. Taylor, Rudiyanto Gunawan, Linda Petzold, Francis DoyleIEEE Transactions on Circuits and Systems I Fundamental Theory and Applications 2007 · DOI

TL;DR

This paper introduces a novel method, the parametric impulse phase response curve (pIPRC), to analyze the robustness of oscillating biological systems, specifically the mammalian circadian clock. By extending traditional sensitivity analysis, the pIPRC provides a powerful tool to predict how changes in system parameters affect the timing (phase) of these biological rhythms, even for large perturbations.

What Problem Does This Paper Address?

Many biological processes, from cell division to sleep-wake cycles, are governed by internal biological clocks called circadian oscillators. A key feature of these clocks is their ability to synchronize with the external environment, a process called entrainment. This synchronization is crucial for maintaining health and involves the clock's ability to adjust its internal timing (phase) in response to environmental cues like light.

While mathematical models (systems of ordinary differential equations) are used to understand these complex biological networks, traditional methods for analyzing system robustness, like classical sensitivity analysis, are not well-suited for studying the phase behavior of oscillating systems. Specifically, these methods struggle to quantify how small changes in parameters affect the timing of the oscillations. This gap makes it difficult to fully understand the robustness of biological clocks and how they maintain their rhythmic behavior despite internal and external variations.

Methods

The authors propose a novel sensitivity measure called the parametric impulse phase response curve (pIPRC). This method is designed to quantify the phase shift incurred by an infinitesimal impulse perturbation to a parameter.

Here's a breakdown of their approach:

  • Classical Sensitivity Analysis: They start by defining classical sensitivity coefficients, which measure the change in a state variable at a specific time due to a parameter perturbation at an earlier time.
  • Phase Definition: They define phase as a measure of time evolution along the limit cycle of an oscillating system.
  • State Impulse Phase Response Curve (sIPRC): They introduce the sIPRC to measure the phase shift caused by an infinitesimal perturbation to a state variable. This is derived using adjoint Green's functions.
  • Parametric Phase Sensitivity: They extend this to parametric perturbations by using the chain rule, relating the change in phase to changes in state variables and parameters. This leads to the concept of cumulative phase sensitivity.
  • Parametric Impulse Phase Response Curve (pIPRC): The pIPRC is derived as the time derivative of the cumulative phase sensitivity. It essentially captures the instantaneous rate of change of phase with respect to a parameter. This is the core novel contribution of the paper.
  • Analytical Phase Response Curve (APRC): They also define an APRC for square pulse perturbations, which is related to the cumulative phase sensitivity.
  • Model: They apply these methods to a well-established mathematical model of the mammalian circadian clock, which is a system of ordinary differential equations exhibiting stable limit cycle behavior.

Datasets

Not discussed in the paper. The study uses a simulated mathematical model of the mammalian circadian clock.

Key Insights and Ideas

  • Phase is Key for Circadian Clocks: The paper emphasizes that for circadian clocks, maintaining the correct phase (internal timing) is as important, if not more so, than maintaining a precise period. The ability to adjust phase is what allows for entrainment to the environment.
  • Limitations of Traditional Sensitivity Analysis: Standard sensitivity tools, which focus on state variables at specific times, are insufficient for understanding the phase dynamics of oscillators. They don't directly capture how parameter changes affect the timing of the rhythm.
  • pIPRC as a Powerful Predictor: The pIPRC, though derived from local sensitivity analysis, proves to be a surprisingly powerful predictor of phase shifts, even for large perturbations. This is a significant finding, as it offers a computationally efficient way to assess robustness.
  • Isochrons are Fundamental: The concept of isochrons (level sets of phase) is crucial for understanding how perturbations lead to phase shifts. The paper leverages this concept to develop its phase-based sensitivity measures.
  • Unified Framework: The proposed sensitivity measures (sIPRC, pIPRC, APRC) offer a unified framework for analyzing the timing properties of oscillating systems, relating phase sensitivity to period sensitivity and response to various signal types.

Key Results

  • The paper successfully derives and demonstrates the utility of the parametric impulse phase response curve (pIPRC).
  • They show that the pIPRC can accurately predict the phase shift incurred by parametric perturbations, even when these perturbations are significant.
  • Numerical experiments using a mammalian circadian clock model demonstrate that the pIPRC predictions align well with simulated experimental results for various light pulse durations and strengths. For example, for 1-circadian hour pulses, predictions were within 0.01 to 0.5 circadian hours of the numerical experimental PRC.
  • The study highlights a direct relationship between the area under the pIPRC and the period sensitivity, confirming theoretical postulates about circadian rhythm dynamics.

Limitations and Future Work

  • Local Measure, Global Prediction: While the pIPRC is a local sensitivity measure, its effectiveness in predicting responses to large perturbations is a key finding. However, the authors acknowledge that for very large perturbations, direct numerical simulation (the "direct method") remains the most accurate approach due to its ability to capture all non-linearities.
  • Assumption of Return to Nominal Flow: The pIPRC is computed assuming the system returns to its nominal vector field immediately after the perturbation. If the perturbation is sustained or changes, the predictions might be less accurate, though still qualitatively correct.
  • Future Work: The authors suggest extending the use of pIPRCs to study collections of coupled oscillators that communicate via intercellular signals, which is crucial for understanding synchronization phenomena in biological systems like the mammalian circadian clock. They also envision using pIPRCs to study the response to arbitrary signals, moving beyond simple square pulses.

Summary updated

Abstract

Vital physiological behaviors exhibited daily by bacteria, plants, and animals are governed by endogenous oscillators called circadian clocks. The most salient feature of the circadian clock is its ability to change its internal time (phase) to match that of the external environment. The circadian clock, like many oscillators in nature, is regulated at the cellular level by a complex network of interacting components. As a complementary approach to traditional biological investigation, we utilize mathematical models and systems-theoretic tools to elucidate these mechanisms. The models are systems of ordinary differential equations exhibiting stable limit cycle behavior. To study the robustness of circadian phase behavior, we use sensitivity analysis. As the standard set of sensitivity tools are not suitable for the study of phase behavior, we introduce a novel tool, the parametric impulse phase response curve.
BibTeX
@article{taylor2007sensitivity,
  author = {Stephanie R. Taylor and Rudiyanto Gunawan and Linda Petzold and Francis Doyle},
  title = {{Sensitivity Measures for Oscillating Systems: Application to Mammalian Circadian Gene Network}},
  journal = {IEEE Transactions on Circuits and Systems I Fundamental Theory and Applications},
  year = {2007},
  doi = {10.1109/tcsi.2007.911364},
}

Loading related content...

Figure