A phylogeny of biological patterns formed by nonlocal advection

Valeria Giunta, Thomas Hillen, Mark A Lewis, Jonathan R Potts · 2025

Read A phylogeny of biological patterns formed by nonlocal advection by Valeria Giunta, Thomas Hillen, Mark A Lewis, Jonathan R Potts (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-00489.txt.

Introduction

Valeria Giunta, Thomas Hillen, Mark Lewis, and Jonathan Potts map the patterns that can form when living things respond to distant neighbours, in a 2025 paper. Such nonlocal effects drive tumour spread, cell sorting, and animal territories. Maths models that capture them are well studied on a line, but far less on a flat surface.

The team ran detailed simulations on two-dimensional domains and varied the model settings in a planned way. Patterns that emerged included separate clusters, stripes, volcano shapes, and mosaic tiles, many of which appear in nature. By linking each pattern to the movement rule behind it, they build a family tree of patterns. The authors argue that matching a real pattern to this tree can hint at the hidden rule that made it, in fields from ecology to cancer study.

Catalog record

CREATORValeria Giunta, Thomas Hillen, Mark A Lewis, Jonathan R Potts
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE56 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-00489

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A phylogeny of biological patterns formed by nonlocal advection Valeria Giunta, Thomas Hillen, Mark A Lewis, Jonathan R Potts Abstract From tumour invasion to cell sorting and animal territoriality, many biological systems rely on nonlocal interactions that drive complex spatial organisation. Partial differential equations (PDEs) with nonlocal advection are increasingly recognised as powerful tools for capturing such phenomena. However, most research has focused on one-dimensional domains, leaving their two-dimensional behaviour largely unexplored. Here, we present a detailed numerical study of the patterns formed by these systems on 2D domains. Depending on the underlying mechanisms, a wide variety of spatial patterns can emerge – including segregated clusters, stripes, volcanos, and polygonal mosaics – many of which have been observed in natural systems. By systematically varying model parameters, we classify the links between emergent patterns and their underlying movement mechanisms. In comparing these patterns with empirical observations, we show how this modelling framework can help reveal possible mechanisms of self-organisation in various situations within the life sciences, from ecology and developmental biology to cancer research. 1 Introduction The phenomenon of pattern formation has long fascinated scientists from many disciplines [ 1 ] . From the coordinated movements of flocks of birds and schools of fish to the formation of honeycomb lattices in beehives, nature is full of striking examples of order emerging from disorder. More remarkably, similar patterns occur in very different systems–for example, the sorting of cells into segregated patterns during tissue growth [ 2 ] and the segregation of animal populations into spatially-distinct territories [ 3 ] –raising the question of whether this similarity suggests common features in underlying mechanism. Indeed, despite the diversity of the agents involved, these structures all emerge from the same fundamental principle: at the microscopic level, individuals (particles, cells or animals) interact according to simple local rules (for example, attraction or repulsion); then, when these interactions occur over many individuals, they give rise to macroscopic emergent patterns. Consequently, self-organisation has become a pillar of mathematical biology, underpinning many models used to understand pattern formation and collective dynamics [ 4 ] . Recently, nonlocal advection-diffusion models have emerged as a powerful framework for capturing the spatial organisation of interacting populations [ 5 , 6 , 7 , 8 , 9 ] . These formulations take into account the effect of spatially nonlocal sensing of other organisms on organism movement , which could be mediated through sight, smell, hearing, or touch. Biological examples of nonlocal interactions are abundant. In ecology, they range from the relatively proximate sensing of swarming insects and flocking birds [ 10 ] to longer-range interactions between vertebrate species [ 11 , 12 ] . It is even possible to place indirect local interactions into the nonlocal advection-diffusion framework, for example those mediated by memory or marks on the landscape [ 7 , 13 ] . In cell biology, by extending filopodia (essentially tentacle-like protrusions), cells can sense the presence of others at distances of several cell diameters [ 14 , 15 ] . Such non-local sensing is believed to be a key process behind various observed phenomena such as cell sorting, cellular aggregations, and chase-and-run dynamics [ 16 , 17 ] . Mathematically, nonlocal sensing of all these kinds can be captured by a nonlocal advection term in a partial differential equation. These nonlocal advection terms can drive spontaneous pattern formation in both single species [ 18 ] and multi-species [ 19 ] systems. Existing studies have documented a wide range of patterns that emerge from such systems, including stationary aggregations, segregation patterns, time-oscillating aggregates, travelling waves, period-doubling routes to chaos, and chase-and-run dynamics [ 19 , 20 , 21 , 22 ] . These have been uncovered using analytic techniques such as linear stability [ 19 , 21 ] , weakly non-linear analysis [ 23 ] , Crandall-Rabinowitz bifurcation theory [ 9 , 24 ] , and energy approaches [ 25 , 26 , 27 ] , as well as through numerical simulations [ 28 , 22 ] . In one spatial dimension, these approaches have revealed quite rich pattern formation properties from systems with only two species [ 19 , 26 , 23 , 22 ] , involving various configurations of aggregation amongst and between species, partial and full segregation between species, as well as perpetual oscillatory and chase-and-run dynamics. However, most of these studies have focused on one spatial dimension, with the effect of higher dimensions being less explored, despite having qualitatively different patterning properties [ 21 ] . Here, we analyse numerically the patterns that form from a multi-species system of advection-diffusion equations in two spatial dimensions, where the nonlocal advection terms may model either attraction or repulsion, whether between individuals of the same population (self-interaction) or between those of different populations (mutual interaction). We investigate the range of spatial patterns that can form within this system, focusing particularly on a two-species system but also investigating certain patterns for more than two species. Our aim is to make steps toward a categorisation of these emergent patterns and how the underlying processes drive changes in emergent patterns. The resulting categorisation takes a form analogous to a phylogenetic tree from evolutionary biology, but in our case the trunk is the spatially-homogeneous steady state (i.e. no patterns) and the branching points represent bifurcations from one pattern to another, driven by changes in one or more of the parameter values, rather than by changes in genes. Having constructed this ‘phylogeny’, we examine which of the observed patterns correspond to those found in natural systems. The aim here is to demonstrate how our model might help elucidate the collective motion mechanisms behind the formation of observed patterns. In situations where we were unable to identify biological patterns similar to those discovered numerically, this gives rise to the question of why not: has biology evolved to avoid these mechanisms or have we simply not observed them, despite such patterns perhaps being present in the empirical world? Overall, our study helps shed light on the variety of patterns that can form purely as a result of the type of non-local advection processes that are ubiquitous across the biological sciences. 2 The Model Let u_{1}(x,t),\dots,u_{N}(x,t) each denote the density of a population of organisms at time t . The purpose of our study is to investigate numerically the following system \frac{\partial u_{i}}{\partial t}=D_{i}\Delta u_{i}+\nabla\cdot\left(u_{i}\sum_{j=1}^{N}\gamma_{ij}\nabla(K_{ij}\ast u_{j})\right),\qquad i=1,\dots,N, (1) defined on a 2 D spatial domain. The constant D_{i}>0 is the diffusion coefficient of species i , for i=1,\dots,N . The functions K_{ij} are zero-mean probability density functions describing the nonlocal detection of species j by species i . We assume that the support of K_{ij} is {\cal{B}}^{2}_{R_{ij}} , where {\cal{B}}^{2}_{R_{ij}} denotes the 2 -dimensional sphere with radius R_{ij} . From a biological point of view, the constant R_{ij} denotes the detection interval within which the species i is able, or willing, to interact with species j . To ensure that the solution remains positive and exists for all time, we assume that K_{ij} is a twice-differentiable, with \nabla K_{ij}\in L^{\infty} [ 29 ] . The function K_{ij}\ast u_{j} denotes a convolution operator defined as K_{ij}\ast u_{j}(x)=\int_{\Omega}K_{ij}(x-y)u_{j}(y)dy. For simplicity with nonlocal boundary conditions, we consider solutions on a torus.

Read the complete text