On a phenotype-structured Shigesada--Kawasaki--Teramoto model: Turing instability and pattern selection under fast phenotype switching

arXiv:2605.28976 · q-bio.PE · Submitted 2026-05-27 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Genomics Radio. Generated commentary on the latest computational biology and genomics papers.

Ines: I'm Ines.

Marcus: And I'm Marcus, here with Yuki.

Ines: We're looking at the paper: "On a phenotype-structured Shigesada--Kawasaki--Teramoto model".

Marcus: Here's the gist of the paper. The paper presents a generalized phenotype-structured formulation of the Shigesada–Kawasaki–Teramoto (SKT) model to study Turing instability and pattern selection under fast phenotype switching, which is important because it connects phenotype distributions to population-scale aggregate spatial dynamics Model Formulation The authors present a generalized phenotype-structured formulation of the SKT model (1) that accounts for phenotypic variability by structuring the competing populations across continuous phenotype state spaces Y1 and Y2 The model uses non-local reaction-cross-diffusion equations where population densities are integrated over their respective phenotype state spaces to obtain total densities The dynamics of the phenotype densities n1 and n2 are governed by coupled non-local reaction-cross-diffusion equations (2) which include terms for movement, competition, and phenotype switching Quasi-Invariant Regime Analysis The authors show how a form of the classical SKT model can formally be derived in the quasi-invariant regime of fast phenotype switching by writing parameters in terms of continuous weighted averages with weights given by the phenotype distributions In this asymptotic regime, the leading order terms for population densities satisfy a system where the effective parameters are continuous weighted averages of their phenotype-dependent functions This leads to a closed system for the densities ρ1 and ρ2 that resembles the unstructured SKT model (1) with effective parameters defined via weighted averages Pattern Formation Analysis The analysis involves Turing-type linear and weakly nonlinear analyses to investigate the onset and characteristics of spatial patterns produced by this model The authors formally find a Turing instability threshold and identify the most unstable mode In the asymptotic regime ε → 0, there is a unique coexistence uniform-in-space steady-state solution of the system (9) under certain conditions The stability to uniform-in-space perturbations is determined by the characteristic equation of an eigenvalue problem Jρe = λρ, e with J defined via (38) Phenotype Distributions as Control Parameters The results demonstrate that phenotype distributions can act as effective control parameters for Turing instability and pattern selection Numerical simulations confirm that spatial patterns emerge for ⟨α21⟩ > ⟨α21⟩c, where ⟨α21⟩c is given by (49) Furthermore, the shape of the spatial patterns depends on the critical mode kc given by (50) Changing the phenotype distribution can impact both the formation/suppression and the shape of spatial patterns Numerical Validation The numerical simulations validate the formal asymptotics and confirm the predictions of pattern formation analyses The results show that spatial patterns emerge for ⟨α21⟩ > ⟨α21⟩c, and pattern formation does not occur for ⟨α21⟩ < ⟨α21⟩c Agreement between the non-local problem (9)-(10) as ε → 0 and the limiting problem (22)-(23) is shown by plots in Figure 4 The results also show that phenotype distributions act as effective control parameters for Turing instability, activating or suppressing the onset of the instability by determining the value of phenotype-average effective cross-diffusion coefficient ⟨α21⟩ The study concludes that phenotype distributions can act as effective control parameters for diffusion-driven instability by affecting the onset of pattern formation, the instability region, the critical spatial mode, and the nonlinear character of the Turing-type bifurcation Future investigations may benefit from a focused ecological application to refine model assumptions The authors expect that in an extended model with space-time dependent phenotype switching kernels, the phenotype distributions will mirror spatiotemporal environmental changes The results point towards novel compelling research directions on phenotype-structured models of cross-diffusion competitive systems Appendix A The appendix presents the scheme employed to solve numerically the rescaled system of non-local reaction-cross-diffusion equations (9) subject to the homogeneous Neumann boundary conditions (10) The numerical method is based on finite differences for the differential terms and the Simpson 1/3 rule for the integral terms References Marzia Bisi, Davide Cusseddu, Ana Jacinta Soares, and Romina Travaglini.

Ines: Reaction–diffusion systems from kinetic models for bacterial communities on a leaf surface.

Marcus: Journal of Nonlinear Science, 36:23, 2026 Fiammetta Conforto and Laurent Desvillettes.

Ines: Rigorous passage to the limit in a system of reactiondiffusion equations towards a system including cross diffusions.

Title and authors: Ines: So this paper, "On a phenotype-structured Shigesada--Kawasaki--Teramoto model: Turing instability and pattern selection under fast phenotype switching," it tackles how you can make a classical model for spatial segregation, like the SKT model, account for when the populations aren't all identical.

Marcus: Right. The authors are looking at adding phenotypic variability—basically different types of individuals—to that classic reaction-cross-diffusion system and seeing what happens when those phenotypes switch quickly. It connects the actual distribution of phenotypes to how big spatial patterns emerge in the population dynamics, which is important because it moves beyond just looking at one fixed type of competitor.

Yuki: From a population genetics standpoint, I'm interested in how this structure might relate to real-world species where individuals have different traits that affect their survival and movement. If we can model how those trait distributions drive instability, that gives us more insight into why certain spatial patterns occur in nature compared to simple models.

Ines: Exactly. The paper presents a generalized formulation where the competing populations are structured across continuous phenotype state spaces, Y one and Y two and they can switch between these states while moving around <ref:2605.28976#pg2>. It shows how to get back to the original SKT model by writing the parameters as continuous weighted averages of those phenotype-dependent functions, using the phenotype distributions as weights.

Marcus: That's the core mathematical trick they use in what they call a quasi-invariant regime of fast phenotype switching. They essentially show that when you look at it asymptotically, the leading order terms for the densities behave like a standard SKT model but with these effective parameters being weighted averages of their phenotype-dependent functions.

Yuki: But I wonder what this means for real samples. If we have a population where some individuals are predisposed to move faster or compete differently based on their phenotype, this formulation should capture that heterogeneity driving the pattern formation, not just an average effect.

Ines: The paper then moves into pattern formation analysis using Turing-type linear and weakly nonlinear methods to find the instability threshold and which spatial mode is most unstable. They formally identify a Turing instability threshold in this structured model, and they pinpoint a unique uniform-in-space steady state solution under certain conditions.

Title and authors: Marcus: And they determine the stability of that uniform state by looking at an eigenvalue problem defined by some matrix J which depends on the system parameters, specifically involving term three eight. They show that this analysis leads to a specific characteristic equation for the eigenvalue problem rho e J equals lambda rho e, where J is defined via term thirty eight.

Yuki: So, what's the big picture here? If we can predict the threshold based on these structural averages, it means we can use phenotype distributions to control whether a pattern even forms in a system that otherwise might not be unstable.

Ines: The results show that phenotype distributions are actually acting as effective control parameters for Turing instability and pattern selection. They confirm that spatial patterns emerge when the weighted average of alpha two one is greater than some critical value, which they call alpha two one c, given by term forty nine in the paper.

Marcus: And they also found that the shape of these spatial patterns depends on a critical mode kc, which is given by term fifty in their work. This means changing how the phenotypes are distributed directly changes both whether a pattern forms and what shape that pattern takes.

Yuki: That’s compelling for understanding ecological outcomes because it suggests that tweaking an individual's trait distribution can tune the entire spatial structure of the community, not just whether one exists at all.

Ines: Numerically, they validate these formal asymptotics by showing that for small epsilon, n epsilon i T x y i is approximately rho epsilon i T x Mi y i for almost every time t greater than zero. This confirms that the structured model approaches the simpler limiting problem when you're dealing with very fast phenotype switching dynamics.

Marcus: And they also confirmed the predictions of their pattern formation analysis through numerical simulations, showing that spatial patterns emerge when alpha two one is greater than alpha two one c, and that no pattern forms when it's less than alpha two one c. They also showed agreement between the non-local problem defined by terms nine and ten as epsilon approaches zero and the limiting problem defined by terms twenty two and twenty three.

Title and authors: Yuki: It sounds like a lot of work confirming that the theoretical predictions hold up in practice, especially when you look at how those parameters—the phenotype distributions—act as direct controls for pattern selection.

Ines: The conclusion is that phenotype distributions can act as effective control parameters for diffusion-driven instability by affecting the onset of pattern formation, the instability region, the critical spatial mode, and the nonlinear character of the Turing-type bifurcation. This means we have a richer set of tools to predict complex spatial outcomes in systems with trait variability.

Marcus: For future work, they suggest that if you extend this model to include space-time dependent phenotype switching kernels, then those phenotype distributions will reflect spatiotemporal environmental changes. That's where things get interesting for modeling evolving ecosystems.

Yuki: I think that idea of spatiotemporal dependency is key because it links the internal structure of the population—the phenotype distribution—to how the environment itself is changing over time and space. That opens up ways to model adaptation in response to shifting conditions.

Ines: So, for anyone interested in how we can use these trait structures to predict spatial patterns under fast switching, this paper, "On a phenotype-structured Shigesada--Kawasaki--Teramoto model: Turing instability and pattern selection under fast phenotype switching," is definitely worth looking at.

Marcus: Yeah. It's about showing that when you structure the model by phenotypes, you get these effective parameters that control the whole instability process, which is a big step for applying these kinds of structured models to real ecological data.

Yuki: And it really shows how important those specific distributions are—the ones giving you the right critical mode kc—for predicting whether you see fine segregation or coarser structures in the resulting spatial patterns.

Ines: We'll keep an eye on this work, especially as they explore those space-time dependent kernels, because that’s where we might see how population structure responds to changing conditions over time and place.

The paper's summary: Ines: So, to summarize what this paper is really about, they're taking that classic Shigesada Kawasaki Teramoto model and adding phenotype structure—meaning different types of individuals—and seeing how that affects pattern formation when those phenotypes switch fast.

Marcus: Exactly. It shows they can write a generalized formulation for this system where you have these continuous phenotype state spaces, Y one and Y two, and the dynamics are governed by non-local reaction cross diffusion equations.

Yuki: From a population genetics view, what that means is that we're not just looking at one single species competing; we're looking at how the distribution of traits—the phenotypes—drives the spatial arrangement when those traits are changing quickly.

Ines: They do this by showing how you can formally derive a version of the classical model in a fast switching regime, which is called the quasi-invariant regime.

Marcus: That derivation uses continuous weighted averages for the parameters, where those weights come directly from your phenotype distributions. It's like they're finding an effective set of parameters that capture the whole system's behavior when switching is super quick.

Yuki: So, it suggests that you don’t have to run a complex, high-dimensional model if you can figure out these effective parameters based on the trait distributions first. That makes it much more tractable for population genetics.

Ines: Right. Then they use that to do Turing instability analysis—the linear and weakly nonlinear stuff—to find the threshold where spatial patterns start appearing.

Marcus: They formally identify a Turing instability threshold, and they figure out which spatial mode is the most unstable one based on an eigenvalue problem defined by some matrix J.

Yuki: And what's really interesting is how they show that the phenotype distributions themselves become control parameters for this instability. Changing those distributions directly shifts whether a pattern forms at all.

Ines: They confirm that patterns emerge when a specific weighted average of alpha two one is greater than some critical value, alpha two one c, which they calculate from the model structure.

Marcus: And they also show that the actual shape of those spatial patterns depends on another parameter called kc, which is derived from their analysis. So you can control both the existence and the look of a pattern just by adjusting how much you weight different phenotypes.

Yuki: That’s pretty powerful for understanding community structure because it suggests that the underlying trait distribution dictates not just if species segregate, but *how* they segregate spatially.

Ines: Numerically, they validate this by showing that when you look at very small switching rates, the results of their structured model match the simpler limiting problem almost perfectly.

Marcus: They also confirmed those predictions in simulations, showing clear pattern formation above alpha two one c and no pattern below it. They even showed that the non-local problem matches the limiting problem as epsilon goes to zero.

Yuki: So, while this is a mathematical model of competition, the implication is we can use trait distributions to predict ecological outcomes in terms of spatial segregation under dynamic conditions.

Ines: And for future work, they suggest that if you make those phenotype switching kernels depend on time and space, then the phenotype distributions will reflect actual changing environments over time and location.

Marcus: That would link the internal structure of the population to external environmental changes, which is a big step toward modeling evolving ecosystems.

Yuki: It opens up ways to model adaptation because you’d be modeling how trait distributions change in response to shifting conditions.

The paper's improvements: Ines: So, we just talked about how they use phenotype distributions to control pattern formation in this Shigesada Kawasaki Teramoto model under fast switching. Now, let's look at what they suggest as improvements for the research itself.

Marcus: They point out that the current analysis is heavily focused on the quasi-invariant regime of very fast switching, where you use those weighted averages to get an effective unstructured SKT model.

Yuki: That means their main finding is tied to how accurately you can approximate those complex, structured dynamics with simpler, effective parameters derived from the phenotype distributions.

Ines: They suggest that they need to look at systems where the switching isn't strictly fast, but maybe intermediate—where it's not instantaneous but still fast enough for these averages to hold up.

Marcus: The paper implies that their current method is a formal limit, so the next step would be testing how sensitive those effective parameters are when you move slightly away from that ideal limit.

Yuki: From a population genetics standpoint, this means they're suggesting we test if the observed spatial patterns in nature can actually be explained by these averaged parameters, even if the switching isn't perfectly fast.

Ines: They also flag that their current analysis focuses mostly on Turing instability—which is diffusion-driven instability—and they acknowledge that other types of instabilities might exist in this structured setting.

Marcus: So, they recommend extending the work to check for other types of instabilities, not just the ones where you get those classic spots and stripes we see in reaction-diffusion systems.

Yuki: That would be important because ecological systems don't just form simple patterns; they can have more complex spatial arrangements based on how those traits interact.

Ines: They also hint at extending the model to include things that happen over space and time, which is where the real complexity lies—if you make the switching rules change depending on where you are in space or when it is in time.

Marcus: That's a big step because if you can model those space-time dependent kernels, then your phenotype distributions wouldn't just be static inputs anymore; they would become dynamic based on the environment.

Yuki: That really connects the internal biology—the traits—to how the environment itself is changing over time and place, which is crucial for modeling evolution.

Ines: So, they are pushing to move from a static trait structure to a dynamic one that responds to environmental shifts across space and time.

Marcus: It means moving away from just looking at a single snapshot of the system toward something that captures the ongoing process of adaptation and pattern selection in real-time.

Yuki: That moves us closer to understanding how species might maintain or change their spatial structures when they are reacting to changing conditions over time.

Conclusion: Tom: So we’re wrapping up on this paper, "On a phenotype-structured Shigesada--Kawasaki--Teramoto model: Turing instability and pattern selection under fast phenotype switching." Basically, they showed how you can use trait distributions to control whether spatial patterns form in these competing species models.

Ines: Exactly. The main point is that the structure of the phenotypes—how varied they are—acts as a powerful lever for controlling diffusion-driven instability, which is what causes those patterns we see in biology.

Marcus: It’s really about showing that you don't need a perfectly uniform population to get interesting spatial dynamics; the heterogeneity itself drives things.

Yuki: From my perspective, this means the biological reality of species having different traits is more important than just looking at a single averaged model when trying to predict where and how those patterns will form in an ecosystem.

Ines: They confirmed that phenotype distributions control the onset, the instability region, and even the shape of those patterns based on parameters like alpha two one and kc.

Marcus: That’s a big deal because it means we can use trait data to predict not just if a pattern will exist, but what kind of structure we should expect to see.

Yuki: It supports the idea that the historical evolution of species might have been shaped by these trait distributions driving spatial segregation in their habitats.

Ines: They also pointed out that future work needs to look at how those switching rules change depending on space and time, which is where you start modeling dynamic adaptation.

Marcus: So, the paper sets the stage for moving from static trait structures to models that reflect real-world environmental changes over time.

Yuki: It’s a great direction because it moves us beyond simple competition to looking at how biological structure evolves under changing conditions.

Department of Mathematical Sciences G. L. Lagrange Politecnico di Torino Department of Mathematics and Computer Science University of Palermo

q-bio.PE

Submitted: 2026-05-27

Updated: 2026-11-04

Comments: 24 pages, 6 figures

Code: https://github.com/davidecusseddu/PS-SKT

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 80/100

The gist: The paper presents a generalized phenotype-structured formulation of the Shigesada–Kawasaki–Teramoto (SKT) model to study Turing instability and pattern selection under fast phenotype switching…

Key concepts

Phenotype-Structured Formulation
This method incorporates phenotypic variability by structuring competing populations across continuous phenotype state spaces. Instead of one simple model, it uses non-local reaction-cross-diffusion equations where population densities are integrated over these continuous phenotype spaces, capturing complex biological interactions.
Quasi-Invariant Regime Analysis
In the regime of fast phenotype switching (small epsilon), the complex structured model simplifies. Parameters become continuous weighted averages based on phenotype distributions. This allows researchers to reduce the problem to a simpler, unstructured SKT model with effective parameters derived from these averages.
Phenotype Distributions as Control Parameters
The distribution of phenotypes in a population is not just descriptive; it actively controls pattern formation. Numerical simulations confirm that changing the phenotype distribution can activate or suppress Turing instability, alter the region where patterns form, and change the characteristic spatial mode of those patterns.
Turing Instability Threshold
This concept identifies the critical point where a uniform state becomes unstable and begins to spontaneously form spatial patterns. The analysis determines this threshold based on effective parameters derived from phenotype distributions, showing that these distributions directly influence whether this instability occurs.

Terminology

Summary

The paper presents a generalized phenotype-structured formulation of the Shigesada–Kawasaki–Teramoto (SKT) model to study Turing instability and pattern selection under fast phenotype switching, which is important because it connects phenotype distributions to population-scale aggregate spatial dynamics

Model Formulation

The authors present a generalized phenotype-structured formulation of the SKT model (1) that accounts for phenotypic variability by structuring the competing populations across continuous phenotype state spaces Y1 and Y2 The model uses non-local reaction-cross-diffusion equations where population densities are integrated over their respective phenotype state spaces to obtain total densities The dynamics of the phenotype densities n1 and n2 are governed by coupled non-local reaction-cross-diffusion equations (2) which include terms for movement, competition, and phenotype switching

Quasi-Invariant Regime Analysis

The authors show how a form of the classical SKT model can formally be derived in the quasi-invariant regime of fast phenotype switching by writing parameters in terms of continuous weighted averages with weights given by the phenotype distributions In this asymptotic regime, the leading order terms for population densities satisfy a system where the effective parameters are continuous weighted averages of their phenotype-dependent functions This leads to a closed system for the densities ρ1 and ρ2 that resembles the unstructured SKT model (1) with effective parameters defined via weighted averages

Pattern Formation Analysis

The analysis involves Turing-type linear and weakly nonlinear analyses to investigate the onset and characteristics of spatial patterns produced by this model The authors formally find a Turing instability threshold and identify the most unstable mode In the asymptotic regime ε → 0, there is a unique coexistence uniform-in-space steady-state solution of the system (9) under certain conditions The stability to uniform-in-space perturbations is determined by the characteristic equation of an eigenvalue problem Jρe = λρ, e with J defined via (38)

Phenotype Distributions as Control Parameters

The results demonstrate that phenotype distributions can act as effective control parameters for Turing instability and pattern selection Numerical simulations confirm that spatial patterns emerge for ⟨α21⟩ > ⟨α21⟩c, where ⟨α21⟩c is given by (49) Furthermore, the shape of the spatial patterns depends on the critical mode kc given by (50) Changing the phenotype distribution can impact both the formation/suppression and the shape of spatial patterns

Numerical Validation

The numerical simulations validate the formal asymptotics and confirm the predictions of pattern formation analyses The results show that spatial patterns emerge for ⟨α21⟩ > ⟨α21⟩c, and pattern formation does not occur for ⟨α21⟩ < ⟨α21⟩c Agreement between the non-local problem (9)-(10) as ε → 0 and the limiting problem (22)-(23) is shown by plots in Figure 4 The results also show that phenotype distributions act as effective control parameters for Turing instability, activating or suppressing the onset of the instability by determining the value of phenotype-average effective cross-diffusion coefficient ⟨α21⟩

Conclusions and Future Directions

The study concludes that phenotype distributions can act as effective control parameters for diffusion-driven instability by affecting the onset of pattern formation, the instability region, the critical spatial mode, and the nonlinear character of the Turing-type bifurcation Future investigations may benefit from a focused ecological application to refine model assumptions The authors expect that in an extended model with space-time dependent phenotype switching kernels, the phenotype distributions will mirror spatiotemporal environmental changes The results point towards novel compelling research directions on phenotype-structured models of cross-diffusion competitive systems

Appendix A

The appendix presents the scheme employed to solve numerically the rescaled system of non-local reaction-cross-diffusion equations (9) subject to the homogeneous Neumann boundary conditions (10) The numerical method is based on finite differences for the differential terms and the Simpson 1/3 rule for the integral terms

References

Marzia Bisi, Davide Cusseddu, Ana Jacinta Soares, and Romina Travaglini. Reaction–diffusion systems from kinetic models for bacterial communities on a leaf surface. p. 1 Journal of Nonlinear Science, 36:23, 2026 Fiammetta Conforto and Laurent Desvillettes. p. 1 Rigorous passage to the limit in a system of reactiondiffusion equations towards a system including cross diffusions. p. 0 Communications in Mathematical Sciences, 12(3):457–472, 2014 Giancarlo Consolo, Guglielmo Inferrera, Edoardo Proverbio, and Cinzia Soresina. p. 1 An extended corrosive-passivating model with cross-diffusion for the initiation of corrosion patterns. p. 1 Physica D: Nonlinear Phenomena, page 134986, 2025 Esther S Daus, Laurent Desvillettes, and Ansgar Jüngel. p. 1 Cross-diffusion systems and fast-reaction limits. p. 3 Bulletin des Sciences Mathématiques, 159:102824, 2020 L Desvillettes. p. 1 About the triangular shigesada–kawasaki–teramoto reaction cross diffusion system. p. 0 Ricerche di Matematica, 73(Suppl 1):105–114, 2024 Laurent Desvillettes and Cinzia Soresina. p. 1 Non-triangular cross-diffusion systems with predator–prey reaction terms. p. 2 Ricerche di Matematica, 68(1):295–314, 2019 Lukas Eigentler, Margarita Kalamara, Graeme Ball, Cait E. p. 1 MacPhee, Nicola R. Stanley-Wall, and Fordyce A. Davidson. Founder cell configuration drives competitive outcome within colony biofilms. p. 1 The ISME Journal, 16(6):1512–1522, 2022 F. p. 1 Farivar, G. p. 1 Gambino, V. p. 0 Giunta, M.C. Lombardo, and M. Sammartino. Intraguild predation communities with anti-predator behavior. p. 0 SIAM Journal on Applied Dynamical Systems, 24(2):1110 – 1149, 2025 Francesco Giannino, Annalisa Iuorio, and Cinzia Soresina. p. 1 Beyond water limitation in vegetationautotoxicity patterning: a cross-diffusion model. p. 0 Physica D: Nonlinear Phenomena, 493:135230, 2026 Jia Li and Yuting Ding. p. 1 New mechanism for spatial heterogeneity pattern revealed by nonlocal competition and host-taxis in a 2d pine wilt disease model. p. 2 Journal of Mathematical Biology, 92:19, 2026 Ricardo Ruiz-Baier and Canrong Tian. p. 1 Mathematical analysis and numerical simulation of pattern formation under cross-diffusion. p. 1 Nonlinear Analysis: Real World Applications, 14(1):601–612, 2013 Kamuela E. p. 1 Yong, Yi Li, and Stephen D. p. 1 Hendrix. Habitat choice of multiple pollinators in almond trees and its potential effect on pollen movement and productivity: A theoretical approach using the Shigesada–Kawasaki–Teramoto model. p. 0 Journal of Theoretical Biology, 305:103–109, 2012 Nanako Shigesada, Kohkichi Kawasaki, and Ei Teramoto. p. 1 Spatial segregation of interacting species. p. 0 Journal of Theoretical Biology, 79(1):83–99, 1979 Leah Edelstein-Keshet. p. 1 Mathematical Models in Biology, volume 46 of Classics in Applied Mathematics. p. 0 Society for Industrial and Applied Mathematics (SIAM), 2005 James D. p. 0 Murray. Mathematical Biology: I. p. 0 An Introduction, volume 17 of Interdisciplinary Applied Mathematics. p. 0 Springer New York, New York, 2002 Alison R. p. 4 Hughes, Brian D. p. 1 Inouye, Matthew T. J. Johnson, N. Underwood, and Mark Vellend. Ecological consequences of genetic diversity. p. 0 Ecology Letters, 11(6):609–623, 2008 Daniel S. Maynard, Carlos A. p. 1 Serván, José A. Capitán, and Stefano Allesina. Phenotypic variability promotes diversity and stability in competitive communities. p. 0 Ecology Letters, 22(11):1776–1786, 2019 Manh Hong Duong, Fabian Spill, and Blaine Van Rensburg. p. 1 Eco-evolutionary dynamics of traitstructured, sexually reproducing interacting populations with time-dependent birth and mortality rates. p. 1 Journal of Dynamics and Differential Equations, 2026 Tommaso Lorenzi, Kevin J. p. 0 Painter, and Chiara Villa. p. 1 Phenotype structuring in collective cell migration: a tutorial of mathematical models and methods. p. 0 Journal of Mathematical Biology, 2025 Anton Arnold, Laurent Desvillettes, and Céline Prévost. p. 1 Existence of nontrivial steady states for populations structured with respect to space and a continuous trait. p. 0 Communications on Pure and Applied Analysis, 11(1):83–96, 2012 Guy Barles, Sepideh Mirrahimi, and Benoît Perthame. p. 1 Concentration in Lotka-Volterra parabolic or integral equations: a general convergence result. p. 1 Methods and Applications of Analysis, 16:321–340, 2009 Ángel Calsina and Sílvia Cuadrado. p. 1 Small mutation rate and evolutionarily stable strategies in infinite dimensional adaptive dynamics. p. 7 Journal of mathematical Biology, 48(2):135–159, 2004 Emanuele Bernardi, Tommaso Lorenzi, Mattia Sensi, and Andrea Tosin. p. 1 Heterogeneously structured compartmental models of epidemiological systems: From individual-level processes to populationscale dynamics. p. 1 Studies in Applied Mathematics, 155(2):e70091, 2025 Tommaso Lorenzi, Nadia Loy, and Chiara Villa. p. 0 Phenotype-structuring of non-local kinetic models of cell migration driven by environmental sensing. p. 2 Multiscale Modeling and Simulation, In press, 2026 Tommaso Lorenzi, Elisa Paparelli, and Andrea Tosin. p. 0 Modelling coevolutionary dynamics in heterogeneous SI epidemiological systems across scales. p. 0 Communications in Mathematical Sciences, 22:2131–2165, 2024 Alexander Lorz, Tommaso Lorenzi, Michael E. p. 0 Hochberg, Jean Clairambault, and Benoît Perthame. Populational adaptive evolution, chemotherapeutic resistance and multiple anti-cancer therapies. p. 5 ESAIM: Mathematical Modelling and Numerical Analysis, 47(2):377–399, 2013 Gaetana Gambino, Maria Carmela Lombardo, and Marco Sammartino. p. 0 Turing instability and traveling fronts for a nonlinear reaction–diffusion system with cross-diffusion. p. 1 Mathematics and Computers in Simulation, 82(6):1112–1132, 2012 Gaetana Gambino, Maria Carmela Lombardo, and Marco Sammartino. p. 0 Pattern formation driven by cross-diffusion in a 2D domain. p. 0 Nonlinear Analysis: Real World Applications, 14(3):1755–1779, 2013 Tommaso Lorenzi and Kevin J. p. 1 Painter. Pattern formation within phenotype-structured chemotactic populations. p. 0 Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 481(2324), 2025 Wesley J. p. 0M. Ridgway, Mohit P. p. 1 Dalwadi, Philip Pearce, and S. Jonathan Chapman. A weakly nonlinear theory for pattern formation in structured models with localized solutions. p. 1 arXiv preprint arXiv:2510. p. 020781, 2025 Stéphane Génieys, Vitaly Volpert, and Pierre Auger. p. 1 Adaptive dynamics: modelling Darwin’s divergence principle. p. 0 Comptes Rendus Biologies, 329(11):876–879, 2006 Benoît Perthame and Stéphane Génieys. Concentration in the nonlocal Fisher equation: the Hamilton-Jacobi limit. p. 3 Mathematical Modeling of Natural Phenomena, 2(4):135–151, 2007

Numerical Method

The appendix presents a scheme employed to solve numerically the rescaled system of non-local reaction-cross-diffusion equations (9) subject to the homogeneous Neumann boundary conditions (10) The scheme is based on finite differences for the differential terms and the Simpson 1/3 rule for the integral terms The system is discretized using temporal, spatial, and phenotypic steps ∆t, ∆x, ∆y > 0

The scheme is presented for the original system of non-local reaction-cross-diffusion equations (2) subject to the homogeneous Neumann boundary conditions (8) The discrete points are defined as tn = n∆t, xi = i∆x, yj = j∆y The system is approximated by linear systems using fixed-point iteration to linearize the nonlinearities

The limiting problem (22)-(23) is consistently solved by employing a similar approach: the same finite difference schemes are employed to discretise the differential terms and the same fixed-point iteration algorithm is used to linearise the nonlinearities The numerical schemes are coded in Python and the scripts are available at the link //github.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved AI system can perform formal asymptotic derivations for phenotype-structured SKT models in fast phenotype switching regimes to identify Turing instability thresholds and bifurcation types, as shown by we carry out, on the formal level, Turing-type linear and weakly nonlinear analyses to investigate the onset and characteristics of spatial patterns produced by this model.

  2. Improved AI system can analyze how phenotype distributions act as effective control parameters for pattern formation by determining their impact on the onset of pattern formation, the instability region, the critical spatial mode, and the nonlinear character of the Turing-type bifurcation.

  3. Improved AI system can predict that changing a parameter like mu2 in a phenotype switching kernel M2 can drive a system across a Turing-type bifurcation threshold by affecting the phenotype-average effective cross-diffusion coefficient ⟨α21⟩, thereby controlling the distance from the Turing-type bifurcation threshold without modifying the structure of the instability region.

  4. Improved AI system can use numerical simulations to validate formal asymptotics, confirming that for small epsilon, "n epsilon i(T, x, yi) approx rho epsilon i(T, x) Mi(yi) for almost every t > 0."

  5. Improved AI system can analyze the dependence of spatial patterns on the critical mode kc by predicting that a phenotype distribution corresponding to larger values of kc leads to the emergence of spatially finer segregation structures, whereas a phenotype distribution corresponding to smaller values of kc brings about spatially coarser segregation structures.

Sources

Related papers