Session W11: Collective Behaviors in Biology III

Early Career Award for Biological Physics Research: Life in a tight spot: How bacteria navigate crowded spaces

Bacterial spreading through motility and growth plays a central role in agriculture, biotechnology, the environment, and medicine. These processes are typically studied in the lab in liquid cultures or on flat surfaces; however, many bacterial habitats—e.g., soils, sediments, and biological gels/tissues—are more complex and crowded 3D spaces. In this talk, I will summarize my group's work unravelling how confinement in a crowded 3D space changes how bacteria behave. We have developed the ability to (i) directly visualize bacteria from the scale of a single cell to that of an entire population, and (ii) 3D-print precisely structured multi-cellular communities, in crowded 3D porous media more akin to their natural habitats. Our experiments using this platform have revealed previously unknown ways in which crowding fundamentally alters how bacteria move and grow, both at the single cell and population scales. Guided by these findings, we have developed theoretical models to more accurately predict the motion and growth of bacterial populations, and other forms of "active matter", in complex environments. Taken together, these findings help to reveal new principles to predict and control the organization of bacteria, and active matter in general, in complex and crowded environments. They could also potentially help provide quantitative guidelines for the control of these dynamics in processes ranging from bioremediation and agriculture to drug delivery.   

Cell-cell collisions: Wetting, speed, and acceleration

Collective cell migration is crucial to many biological functions and is controlled by cell-cell interactions, often studied by colliding cell pairs. Geometry of contact can further affect these contact-based interactions. A recent study on colliding epithelial cells suggests that cells with smaller contact angles to the substrate and larger speeds are more likely to maintain their direction upon a collision (“win”) [1]. Can we predict collision outcomes by leveraging cell shape and speed? We investigate this question by simulating two-body cell collisions within the phase field model. Under two distinct polarity mechanisms, we reproduce the observed results of faster and flatter cells winning more consistently. Additionally, we identify summary variables that faithfully predict collision outcomes across a vast spectrum of cell attributes. If outcomes are predictable, can we discover an experimentally accessible predictor? Recent work suggests that the dynamics of epithelial cells can be captured by their acceleration maps [2]. Within our phase field framework, we investigate whether the cell’s acceleration can be a predictor of its collision dynamics.

[1] Jain, Cachoux, Narayana, et. al, Nature Physics 2020

[2] Bruckner, Arlt, Fink, et al, PNAS 2021   

A coarse-grained model for cytoplasmic streaming

Cytoplasmic streaming is a striking example of fluid-structure interactions within living cells. Egg cells are among the largest, and transporting by diffusion of the proteins necessary for their development is extremely slow. In the later stages of the developing fruit fly egg, a coherent circulatory flow emerges that spans the entire ~200 μm scale cell. This streaming flow is driven by the motion of nanometric motors transporting subcellular cargo along stiff biopolymers (microtubules) anchored at the cell wall. Streaming is crucial for the organism’s development, but exactly what functions are fulfilled remains unclear. Here we theoretically investigate the transition to streaming and the consequent transport and mixing. For this, we use a coarse-grained continuum theory that captures the collective response of microtubules and motors that drive the internal flows. This model has the form of a boundary force field fully coupled to an internal Stokesian flow. In particular, we study how the flow line topology is influenced by microtubule density and the geometry of the egg cell, as well as first-passage times from sources to sinks.

    

Diffusion maps for collective dynamics

The emergence of coarse-grain dynamics is a prevalent feature in systems characterized by a separation of spatio-temporal scales. A classical example in Physics is thermodynamics, in which the evolution of the statistical moments of the particle ensembles offer a simpler description than the equations of motion of the particles themselves. In general, identifying coarse-grain variables allows us to simplify the description of a complex system and gain insights from analytical models. Finding the correct coarse-grain description is however challenging, particularly for living systems, since most of the intuition we have developed for classical physical systems does not apply. Here we show that a coarse-grain description of a high-dimensional dynamical system can be found by defining a measure of similarity between the different possible states that the system can achieve. In that sense, data mining techniques become suitable candidates for a systematic approach to find coarse grain descriptors. Here we show that the diffusion map technique is successful at finding the order parameters, identifying phase transitions and the mean-field descriptions of two model systems, the Ising model of ferromagnetism and the Vicsek model of flocking. We apply the same idea to biological data and identify collective states.   

AI-enabled Analysis of collective cell dynamics driven by electric fields

Collective cell motion is critical for the development and healing of organisms. Electric fields are known to impact this collective behavior and may impact both individual cells and their interactions. Particle image velocimetry (PIV) has been widely used to shed light on collective rearrangements. In PIV-based analysis, dynamic information is collected based on small segments of the image, regardless of the local cell density. We present an AI-based cell segmentation and tracking workflow that can identify the motion of individual cells in a group from phase contrast images, and demonstrate its capability of providing detailed cell scale information on the link between individual and collective behavior. AI-based cell segmentation and tracking show that a weak electric field is sufficient to guide the direction of motion of cells without changing their speed, whereas a strong electric field also increases the speed of a subset of cells in the sheet.  The analysis also reveals a significant delay of the cell sheet in its response to a switching of the electric field, when the field is weak.  

    

Unveil microscopic mechanism from collective behaviors

In many-body systems, the interplay between single-particle dynamics and interparticle interaction gives rise to the emergence of collective behavior. For physical systems, condensed matter theory allows us to derive such collective behavior from microscopic details. But for biological systems, microscopic mechanisms are often unknown. In this talk, we will present a generic neural network architecture to solve the inverse problem---revealing dynamical mechanisms of individual cells from their collective behaviors that are easily accessible in experiment. It employs graph neural networks to learn complex intercellular interactions and normalizing flows to capture the intrinsic fluctuations of cells. We show that by training on a few experimental videos, this machine learning (ML) model can accurately predict the stochastic motion of epithelial cells, the deterministic growth of a fly wing, and the wave propagation of ERK signaling. In contrast to traditional ML methods that make deterministic predictions, our probabilistic design makes it possible to reveal the stochasticity of a biological system including the correlation between the cells. Our method paves the way to the data-driven study of biological many-body systems.

    

Learning Mechanisms for Collective Multicellular Behavior with Differentiable Molecular Dynamics Simulations

Living cells display a remarkable ability to self organize into increasingly complex structures - from "symmetry breaking" of identical cells in the embryo to the formation of organoids in vitro. During development, cells can undergo a complex sequence of intercellular interactions and movements to create spatiotemporal organization. However, uncovering the developmental programs that orchestrate this multicellular behavior has proven to be a challenge. In this work, we leverage advances in machine learning technologies to optimize over physics and biology based molecular dynamics (MD) simulations of individual cells, to learn mechanisms that can drive collective behavior. We apply this framework to recover cell-based rules to drive (i) homogeneous tissue growth, (ii) homeostasis between cell types and (iiiI) elongation and sorting of a symmetric cluster of cells. Our work opens new avenues for designing cellular interactions to program complex multicellular behavior, as well as to learn mechanisms from experimental data of developmental trajectories.   

Collective behavior of model swimmers at intermediate Reynolds numbers

While most studies on the collective dynamics of swimming organisms have focused either on the microscale (Stokes regime) at low Reynolds (Re) numbers or the macroscale (Eulerian regime) at high Re wherein viscous and inertial force respectively dominate, little is known about such behaviors in the intermediate regime where the two forces concurrently play a role. Using computational fluid dynamics approaches, we examine how the collective behaviors emerge in a mixture of a simple mesoscale model swimmer immersed in a viscous incompressible fluid under the increasing effect of inertia. Specifically, we observe a range of nontrivial dynamical swimming patterns as Re increases, including different kinds of network structures and swarming. Our results suggest that inertia is one of the important factors and a possible knob to control the collective behaviors of active mesoscale particles immersed in fluids at intermediate Re regime.

    

Chemotactic motility-induced phase separation

In the past decade, extensive research has elucidated the mechanism and generalized thermodynamics of motility-induced phase separation (MIPS), where randomly oriented and self-propelled agents known as active Brownian particles separate into dilute and dense phases. However, it is unclear how chemotaxis, directed motion along a chemical gradient, can affect MIPS—despite its ubiquity in many biological systems such as microbes, eukaryotic cells, and even enzymes, as well as synthetic forms of active matter such as chemically-responsive colloids and robots. Here, we combine continuum models for MIPS and chemotaxis and use linear stability analysis and numerical simulations to study chemotactic MIPS. We find that chemotaxis can dramatically suppress MIPS, arrest Ostwald ripening, and lead to the formation of a fascinating array of oscillatory patterns. These results therefore expand our understanding of the rich phenomenology of MIPS.   

Anomalous collective dynamics of auto-chemotactic populations

Various non-equilibrium systems that are subject to local interactions are known to exhibit collective behavior in the form of scale invariance. A prominent example are absorbing state phase transitions, which often fall into the universality class of directed percolation. Inspired by insights from equilibrium phase transitions, where long-ranged interactions tend to change the collective behavior, we want to study their influence on non-equilibrium systems. For this purpose we investigate a system of reproducing agents that are subject to limited resources and long-ranged chemotactic interactions in the form of Keller-Segel type dynamics. Close to the extinction threshold, a perturbative renormalization group analysis reveals distinct universality classes for attractive and repulsive interactions. We present the adapted phase diagram as well as a a novel nonlinear mechanism that could stabilize the continuous transition against a pattern forming instability.   

Toward ideal gas law for crowds with large pressures

Human collective behavior exhibits a wide range of phenomena. The humans (particles) in the crowds are active, and one may not expect the speed distribution to follow the 2D Maxwell-Boltzmann distribution. However, in many of the cases, the measured/simulated human crowd does follow the Maxwell-Boltzmann distribution. Our study reveals the criterion for crowd speed (corresponds to fluctuation velocity) to follow the Maxwell-Boltzmann distribution. We reveal that the speed PDF becomes more similar to the Maxwell-Boltzmann distribution as the pressure increases, and there is a power law relationship between the mean square error of fit and pressure. To explain the possible reason, we claim that it is the relative magnitude between the relaxation time of individuals and the collision time scale. We reveal that our result is general regardless of the boundary conditions of the crowd. With the Maxwell-Boltzmann speed distribution, one can directly imply that high-pressure crowds follow the ideal gas law. This result will mark a clear pathway to study crowds with high pressures.   

Stochastic bounds of aggregation dynamics distinguish near-wild-type from wild-type strains in social bacteria

Merrill E. Asp, Eduardo Caro, Roy D. Welch, Alison E. Patteson

Living systems constantly navigate stochasticity to survive. This is especially true in the case of multicellular development, where the interwoven networks of signal transduction, gene expression, and physical response produce emergent phenotypes from the collective behavior of many cells. The genotype-phenotype problem for multicellular development asks how genetic inputs control these collective phenotypes. To study this problem, we use the model organism Myxococcus xanthus, a bacterial species that exhibits a rudimentary but robust form of development in the form of fruiting body aggregation, wherein cells in a low-nutrient environment spontaneously organize from a homogeneous, flat swarm into a discrete number of mound-shaped aggregates harboring thousands of cells.

We present a high-throughput imaging setup and analysis pipeline to produce a dataset of these aggregation events that is sufficiently large to use data-driven statistics to quantify the bounds of wild-type behavior. This approach catches subtle phenotypic distinctions between wild-type and mutant strains of M. xanthus, and can be used as a platform for gene annotation and mechanobiological studies into the genotype-phenotype problem.