MS01 - MFBM-13 Part 1 of 4

Modern methods in the data-driven modeling of biological systems (Part 1)

Monday, July 14 at 10:20am

SMB2025 Follow

Note: this minisymposia has multiple sessions. The other sessions are: Part-2 Part-3 and Part-4.

Cody FitzGerald (Northwestern University), Rainey Lyons (CU Boulder), Nora Heitzman-Breen (CU Boulder), Susan Rogowski (NCSU)

Description:

Due to recent developments in laboratory technology and data collection techniques, there is an abundance of large and complex datasets resulting from a vast array of biological experiments. This surge of data demands the development of novel data-driven techniques to generate robust, interpretable, and generalizable models of biological systems. The purpose of this minisymposium is to present modern advances in data-driven methods for modeling biological dynamics in the areas of parameter estimation, scientific machine learning, algorithmic model selection, and weak form methods. This minisymposium also aims to discuss common challenges which appear in the context of data-driven modeling, such as sparse data, unobserved states, noisy data, structural and practical identifiability issues, and incorporating multiple biological scales. Applications for such methods will span many active areas of biological research, including cell migration, physiology, neuroscience, epidemiology, and ecology.

Ruth Baker

University of Oxford

"Combining models and data to unravel cell-cell interactions and provide a link between genotype and phenotype."

In vitro cell biology assays are a cornerstone of biomedical science, from basic research that aims to provide fundamental new understanding of development, disease and repair, through to pharmaceutical applications that include drug discovery and cytotoxicity testing. A widely adopted strategy is to conduct a particular in vitro assay under a range of genetic perturbation conditions and observe how cellular phenotypes change. The key challenge is then to determine the mechanisms by which the genetic perturbations give rise to the phenotypes. I will talk about recent work that combines mathematical modelling, data analysis techniques and computational statistics to characterise cell-cell interactions and provide a link between genotype and phenotype.

Cody FitzGerald

Northwestern University

"Discovering universal temperature regulation dynamics in animals"

Hibernation is an adaptation to extreme environmental seasonality that has been studied for almost 200 years, but our mechanistic understanding of the underlying physiological system remains lacking due to the partially observed nature of the system. During hibernation, small mammals, such as the Arctic ground squirrel, exhibit dramatic oscillations in body temperature, typically one of the only physiological states measured, of up to 40 $^{circ}$C. These spikes are known as interbout arousals and typically occur 10-20 times throughout hibernation. The physiological mechanism that drives interbout arousals is unknown, but two distinct mechanisms have been hypothesized. Using model selection for partially observed systems, we are able to differentiate between these two mechanistic hypotheses using only body temperature data recorded from a free-ranging Arctic ground squirrel. We then modify our discovered physiological model of Arctic ground squirrel to include environmental information and find that we can qualitatively match body temperature data recorded from a wide range of species, including a bird, a shrew, and a bear, which also dynamically modulate body temperature. Our results suggest that a universal, environmentally sensitive mechanism could regulate body temperature across a diverse range of species---a mechanistic restructuring of our current understanding of the physiological organization across species. While the findings presented here are applicable to thermophysiology, the general modeling procedure is applicable to time series data collected from partially observed biological, chemical, physical, mechanical, and cosmic systems for which the goal is to elucidate the underlying mechanism or control structure.

Joshua Macdonald

Johns Hopkins

"Toward Trait-Based Plankton Experiments: Invasion Margins and Automated Tools for Data-Driven Modeling"

We develop a size‐structured NPZD (nutrient–phytoplankton–zooplankton–detritus) framework that couples discrete size classes, concave nutrient uptake, generalized (including sigmoidal) mortality, and seasonal forcing to capture plankton blooms, coexistence, oscillations, and hysteresis within a single model. Under mild smoothness and monotonicity assumptions, we prove uniform dissipativity, characterize all boundary and interior equilibria via classical and effective invasion thresholds, and apply uniform‐persistence theory to derive necessary and sufficient conditions for the long‐term survival of each size class. We introduce net invasion margins (Delta R_P), (Delta R_Z) and define a balance index [ B ;=;frac{Delta R_P - Delta R_Z}{1 + Delta R_P}, ] which distinguishes bottom‐up ((B>0)), mixed ((Bapprox0)), and top‐down or bistable ((B<0)) regimes. Local stability analysis identifies only a phytoplankton‐only transcritical bifurcation and Hopf bifurcations at zooplankton invasion‐margin crossings, while sigmoidal mortality generates boundary‐bistability and hysteresis loops. Our framework makes it possible to estimate (Delta R_P) and (Delta R_Z) from standard dilution or mesocosm assays and (B) from field biomass ratios, thereby guiding the design of future experiments. We provide automated Julia workflows for time-series simulation, invasion-margin assays, and bifurcation detection to support future data-driven modeling and experimental planning. This work lays the groundwork for trait-based plankton experiments and modern data-driven modeling in biodiversity-rich marine systems.

William Lavery

Uppsala University

"Biologically-Informed Neural Networks for cell models with minimal prior assumptions"

We use a neural network architecture with a physics-informed cost function to formulate a mathematical model that describes the spatiotemporal dynamics of cells, as seen in, for example, microscopy images and MRIs. In particular, we use a supervised learning framework on cell coordinate data while enforcing a gen- eralised reaction-diffusion partial differential equation (PDE) in two dimensions. By representing the diffusion and reaction terms as multilayer perceptrons, the method learns their forms with minimal prior assumptions (informed by fun- damental biological considerations) while simultaneously solving the governing PDE. We have evaluated the method in two scenarios: (1) numerically simulated cell density data obtained by forward-solving the original PDE and (2) agent- based simulations that converge to the continuous cell density case in the limit of infinite cells and infinitesimal time steps. Our next step is to apply the frame- work to experimental in vitro data to uncover underlying dynamics. The overall approach can be extended to other governing PDEs (e.g., incorporating additional terms on the right-hand side of the reaction-diffusion equation) and broader particle interaction models.

#SMB2025 Follow

Annual Meeting for the Society for Mathematical Biology, 2025.