CDEV-MS08

Control theory for microbiology

Tuesday, June 15 at 11:30am (PDT)
Tuesday, June 15 at 07:30pm (BST)
Wednesday, June 16 03:30am (KST)

SMB2021 Follow Tuesday (Wednesday) during the "MS08" time block.

Robert Planqué (Vrije Universiteit Amsterdam, Netherlands), Diego Oyarzún (University of Edinburgh, United Kingdom), Mustafa Khammash (ETH Zurich, Switzerland)

Description:

In recent years, there has been much interest in the design and study of control theory in unicellular life. The aims are both fundamental and applied. We wish to understand how cells control themselves, by maintaining homeostatis, or by adapting to changing conditions. In synthetic biology, microbial cells such as bacteria and yeasts are to be controlled to produce compounds such as medicines or flavours. Also bioreactors need to be optimised to be cost-effective and give high yields. In this minisymposium, four speakers will present recent results, highlighting state of the art applications of control theory in cellular systems.

Elisa Franco

(University of California Los Angeles, United States)

"Ultrasensitive feedback controllers for quasi-integral feedback"

Biological organisms regulate many of their properties so they fall in a prescribed range, for example temperature, osmotic pressure, and glucose levels. The capacity to preserve a desired condition is enabled by feedback loops that adjust gene expression or metabolism in response to changes or perturbations in the environment. Theory developed in automation engineering indicates that the best way to reject perturbations in a feedback system is to include components that integrate (maintain memory) of past effects of the disturbances, and are known as integrators. While models of biological networks such as osmoregulation and chemotaxis are known to include integral feedback, a different question is how to build molecular integral control systems from the bottom up. With mathematical modeling I will describe how ultrasensitive components can be helpful within feedback loops to maintain a desired gene expression level. I will also discuss a particular ultrasensitive reaction network that combines molecular sequestration and an activation/deactivation cycle, and could be used not only for maintaining a steady state but also for setting a tunable reference. I will finally provide an overview of ongoing projects in our group focused on the role of ultrasensitivity in the context of molecular computation and non-equilibrium kinetics.

Jorge I. Poveda

(University of Colorado, Boulder, United States)

"High-Performance Online Optimization of Bioreactor Systems via Non-Smooth and Hybrid Extremizing Feedback Controllers"

It is well-known that bioreactor systems are highly nonlinear and difficult to model in a precise way by using first principles. Yet, substantial economic benefits can be achieved when the system operates at optimal points, which are difficult to calculate offline. Instead, to achieve this optimal operation in real time, different types of feedback controllers with online adaptation have been developed during the last decades. However, a persistent challenge in most existing approaches is the emergence of prohibitively slow rates of convergence and potentially small basins of attraction, which difficult the tuning of the controller in practical settings. To address these problems, in this talk we will explore a new class of extremizing control algorithms, grounded on ideas from non-smooth control and hybrid control theory, which can overcome some of the limitations of existing approaches based on smooth feedback control laws. We will illustrate our theoretical results via numerical examples in two different types of models of bioreactors.

Mustafa Khammash

(ETH Zurich, Switzerland)

"Beyond Perfect Adaptation: Biological Antithetic Controllers for Enhanced Transient Performance"

Proportional-Integral-Derivative (PID) feedback controllers have been the most widely used controllers in the industry for almost a century due to their simplicity and intuitive operation. Motivated by their success in various engineering disciplines, PID controllers are being explored for use in molecular biology. In this talk, we consider the mathematical realization of PID controllers using biomolecular interactions based on the antithetic controller motif. We propose a simple PID architecture based on a combination of feedback and incoherent feedforward and demonstrate its capability of enhancing the transient dynamics and reducing cell-to-cell variability.

Diego Oyarzún

(University of Edinburgh, United Kingdom)

"Multiobjective optmization of metabolic control systems"

Progress in genetic engineering now allows the construction of molecular circuits inside living cells. In this talk I will present our approach to design such systems using multiobjective optimization. We focus on feedback control circuits designed to steer cellular metabolism toward the production of high-value chemicals. Starting from two-timescale ODE model, we pose and solve cost-benefit optimisation problems for control systems built in the literature so far. The results reveal previously unknown trade-offs between optimality, performance and robustness of metabolic control systems. Our results lay the groundwork for the automated design of control circuits in synthetic biology, with applications in the food, energy and pharma sectors.

Hosted by SMB2021 Follow

Virtual conference of the Society for Mathematical Biology, 2021.