Emergent frequency-dependent selection predicts mutation outcomes in complex ecological communities
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: We're looking at the paper: "Emergent frequency-dependent selection predicts mutation outcomes in complex ecological communities".
Paper summary: Ines: So to recap this paper, they are tackling how community structure affects evolution. The central claim is that ecological interactions generate emergent frequency-dependent selection between parents and mutants, which is quantified by a single parameter measuring the strength of those feedbacks >
Marcus: They achieve this by using dynamical mean-field theory to integrate community ecology into classical population genetics models, showing that this result generalizes classical population genetics models to highly diverse communities >
Ines: The main result they highlight is that when ecological feedback is strong, fixation probabilities for moderately beneficial mutants are suppressed compared to what Kimura's formula predicts >
Marcus: This suppression depends on several things: the effective population size, Neff, the correlation between parents and mutants, rho, and the strength of those ecological feedbacks >
Yuki: It matters because it shows that parent-mutant coexistence is not just a fluke in complex settings; it's a generic feature of population genetics when ecology is involved >
Ines: Exactly. They show that this coexistence region alters fixation and extinction times, resulting in exponentially long absorption times for moderately beneficial mutants when the ecological effects are strong >
Marcus: So, what this means for the field is that we can now use these classical population genetics models with frequency-dependent selection even when we're dealing with complex communities >
Yuki: It suggests that traditional metrics that ignore ecology might be under- or over-estimate extinction risks in conservation genetics because they miss this whole dynamic >
Ines: That’s the big picture. This paper gives us a principled way to predict mutation outcomes even in these highly diverse ecological settings by making it explicit how the community context influences selection >
Marcus: And they're providing a tool, that parameter eta, which allows us to measure that strength of ecological feedback directly from strain dynamics alone >
Yuki: It’s really about connecting the evolutionary trajectory to the community structure in a way we couldn't do before >
Conclusion: Ines: Looking at the title, "Emergent frequency-dependent selection predicts mutation outcomes in complex ecological communities," it tells you exactly what they are doing: they are showing that community structure generates a form of selection we didn't have to account for in simpler models >
Marcus: The authors, Li, Feng, Goyal, and Mehta at MIT and Boston University, are using dynamical mean-field theory to show how this emergent selection works across different ecological models >
Yuki: From my view, the implication is that we need to stop treating population genetics as if it happens in a vacuum; it needs to account for the community context because that context actually dictates the evolutionary fate of a mutation >
Ines: Exactly. It’s not just adding an environmental correction; it’s showing how the selection itself becomes frequency-dependent because of those interactions, which is captured by parameter eta >
Marcus: So, in plain terms, this work suggests that for any complex ecological system you study, you can predict mutation outcomes better if you measure the strength of those community feedbacks and include that into your genetic model >
Yuki: That moves us toward a more realistic way to build evolutionary models for real-world biological systems where interactions are the norm rather than the exception >
Ines: It’s about moving from just predicting what happens in isolation to understanding how interactions drive the selection process itself, and this paper does that by showing that emergent frequency-dependent selection is a direct consequence of community structure >
MIT Center for Theoretical Physics - a Leinweber Institute · Department of Physics, Boston University Department of Physics, International Centre for Theoretical Sciences Tata Institute of Fundamental Research Faculty of Computing and Data Sciences Boston University
q-bio.PE, cond-mat.dis-nn, cond-mat.stat-mech
Submitted: 2025-09-28
Updated: 2026-11-17
Comments: 14 pages, 5 figures + SI Appendices; v3: new data analysis on human gut microbiome
Code: https://github.com/Emergent-Behaviors-in-Biology/Pop-Gen-Ecology
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 87/100
The gist: Emergent frequency-dependent selection predicts mutation outcomes in complex ecological communities by integrating community ecology into classical population genetics models to predict mutant…
Key concepts
- Emergent frequency-dependent selection
- Ecological interactions between parents and mutants create a form of natural selection where the selective advantage changes based on the mutant's frequency within the community. This effect arises generically from community feedbacks and is characterized by a single parameter quantifying its strength.
- Community-mediated feedback parameter ($\eta$)
- This new parameter quantifies how strongly ecological interactions, like niche packing, influence the dynamics of a mutant. It captures the combined effect of the entire community on mutant frequency changes, allowing researchers to simplify complex ecological models into a single evolutionary term.
- Parent-mutant coexistence
- When ecological effects are strong, parents and mutants can coexist for very long periods instead of one eliminating the other quickly. This prolonged coexistence significantly alters fixation and extinction times, leading to exponentially longer mean absorption times for moderately beneficial mutations.
- Suppression of fixation probabilities
- Strong ecological feedback suppresses the probability that a moderately beneficial mutant will eventually fix in the population compared to predictions from classical population genetics. This suppression is dependent on factors like effective population size and parent-mutant correlations.
Terminology
Summary
Emergent frequency-dependent selection predicts mutation outcomes in complex ecological communities by integrating community ecology into classical population genetics models to predict mutant fixation probabilities in highly diverse settings
Bridging Population Genetics and Community Ecology
The study bridges the gap between classical population genetics, which largely neglects interactions from a species-rich community, and ecological context by using dynamical mean-field theory to integrate community ecology into classical population genetics models Here, the authors show that ecological interactions result in emergent frequency-dependent selection between parents and mutants, characterized by a single parameter measuring the strength of ecological feedbacks This result generalizes classical population genetics models to highly diverse communities and enables predictions of mutation outcomes in these eco-evolutionary settings They derive an analytic expression for fixation probability that extends Kimura’s formula and reveals that ecological interactions strongly suppress the fixation of moderately beneficial mutations
Emergent Selection Mechanism
Ecological feedbacks manifest as emergent frequency-dependent selection between parents and mutants This frequency-dependence emerges as a generic consequence of community-mediated ecological feedbacks on the parent and mutant The community’s combined effect can be captured by only a single new parameter that quantifies the strength of community-mediated feedbacks on mutant dynamics This parameter can be computed from a community’s statistical properties such as its nichepacking The resulting equation for mutant frequency is given by df/dt = (sinv - ηf) f(1 - f) + s/f(1 - f) Neff ξ(t), where sinv measures the invasion fitness of the mutant in the community
Suppression of Fixation Probabilities
The study demonstrates that when ecological feedback is strong, the fixation probabilities for moderately beneficial mutants are suppressed compared to Kimura’s predictions This suppression increases with effective population size (Neff) and weaker parent-mutant correlations (ρ) The deviation from Kimura’s formula arises from prolonged parent-mutant coexistence arising from ecological interactions Specifically, when ecological effects are strong (α >> 1), fixation probabilities for moderately beneficial mutants are strongly suppressed relative to Kimura’s prediction
Coexistence and Time Scales
Parent-mutant coexistence alters fixation and extinction times, which is a second important quantity in population genetics When ecological effects are strong (α >> 1), the mean times to absorption (either extinction or fixation) become exponentially long for moderately beneficial mutants with 0 ≲ sinv ≲ η, peaking at sinv ≈ η/2 This coexistence region is a direct consequence of frequency-dependent selection (Eq. (3)) The mean extinction time reaches maximum when sinv = sc ≈ η/2, i. p. 6e., Tmax ∼ exp(α 2/4), which is determined by α only
Generalizability and Implications
The framework is robust to the details of the exact ecological model used to mathematically represent the community dynamics, applying to various models including generalized Lotka-Volterra models and multiple variants of the consumer-resource model The results show that parent-mutant coexistence is a generic feature of population genetics in complex communities and is ubiquitous across all of the ecological models studied The theory makes testable predictions for highresolution strain tracking in natural microbiomes [65] and suggests that estimating both f∗ and sinv from observed trajectories could enable direct inference of the ecological feedback strength η from strain dynamics alone This framework has broader implications for inference using population genetics models, suggesting that current methods may yield biased estimates of effective population sizes, demographic histories, and selection coefficients
Methods and Model Extensions
The authors use dynamical mean-field theory (DMFT) to coarse-grain the effect of a highly-diverse community on a mutant’s frequency dynamics They derive an analytic formula for the fixation probability pfix in complex communities (Eq. (5)) The results are shown to generalize beyond the GLV model to many variants of consumer-resource models, leading to the same dynamics as in the MacArthur consumer-resource model The analysis involves solving stochastic differential equations using a generalized Euler method with demographic noise They also explore how different ecological properties of the underlying community control the strength of suppression by measuring α across a variety of parameter sets in the GLV model The study concludes that ecological interactions likely generically suppress fixation of neutral to moderately beneficial mutants, strongly deviating from the predictions of classical population genetics This is supported by simulations showing that our theory is indeed applicable to various ecological models The results are further tested in consumer-resource models where the packing fraction determines the strength of ecological suppression The fixation probabilities are more suppressed for a less packed community with lower packing fraction in external resource scenarios
Extensive Analysis of Fixation Probability
The fixation probability can be obtained by solving the stochastic differential equation, leading to pfix = Erfi(αs˜inv) − Erfi(α(˜sinv − f0)) - Erfi(αs˜inv) − Erfi(α(˜sinv − 1)) For neutral mutations (sinv = 0), the fixation probability is pfix ≈ exp(-2Neffsc) when ecological effects are large (α >> 1) This matches Kimura’s prediction when α≪1, but exponentially suppressed by α otherwise The behavior of pfix is divided into different regimes according to sinv, with the coexistence region being characterized by flowercoexist < fuppercoexist centered at f∗ ≈ s˜inv The mean extinction time reaches maximum when sinv = sc ≈ η/2, i.e., Tmax ∼ exp(α 2/4), which is determined by α only This suggests that the maximum time is exponentially longer as the fixation probability is exponentially suppressed, since both exponential behavior is caused by parent-mutant coexistence as explained in the main text
Inference and Future Directions
The results suggest that new empirical methods could be developed to quantify the impact of ecology during evolutionary dynamics by estimating η from strain dynamics alone Current methods largely neglect ecology and for this reason may yield biased estimates of effective population sizes, demographic histories, and selection coefficients Future work will extend the theory to a wider range of ecological and evolutionary settings, including ecosystems exhibiting limit cycles, multistability [68], and chaos [42] It will also be useful to understand how clonal interference can significantly modify evolutionary dynamics in complex ecological communities The study highlights that traditional metrics that ignore ecology may under- or over-estimate extinction risks in conservation genetics
Appendix Details
The methods include defining the generalized Lotka-Volterra model and various consumer-resource models, analyzing them using dynamical mean-field theory to derive Eq. (3) and the value of η The derivation involves approximating the community contribution by linear responses, which leads to an expression for η involving only the trace of (I - P)Q−1 in some resource models Appendix H provides analytical estimates for the mean extinction time, showing that it deviates from population genetics expectations when a coexistence region is present The simulation details include sampling random ecosystems at steady states and solving the equations using a generalized Euler method with demographic noise These simulations are repeated with multiple instances of demographic noise to obtain the statistical quantities such as fixation probabilities and mean absorption/extinction times The parameters used in the figures are specified in Appendix J, including initial mutant frequency f0 = 0. p. 809 The simulation parameters for various cases are detailed in Appendix I, including how to approximate species dynamics using a Gamma distribution The simulations confirm that our theory is applicable to various ecological models, with agreement between predictions and simulations for high ρ in generalized Lotka-Volterra models The fixation probabilities are more accurate than Kimura’s predictions in general, but there are some deviations for 0 < sinv < η where the parent and the mutant can transiently coexist The packing fraction is given by S∗/M∗, where S∗ is the number of surviving species and M∗ is the number of non-depleted resources The fixation probabilities are more suppressed for a less packed community with lower packing fraction in external resource scenarios The results for different consumer-resource models are qualitatively the same as in generalized Lotka-Volterra model
Summary of Key Findings
The central finding is that community-mediated feedbacks generically lead to emergent frequency-dependent selection across a variety of ecological models This effect can be captured by only a single new parameter η which quantifies the strength of community-mediated feedbacks on mutant dynamics When ecological feedback is strong, fixation probabilities for moderately beneficial mutants are suppressed compared to Kimura’s predictions This suppression depends on effective population size (Neff), weaker parent-mutant correlations (ρ), and decreases as ecosystems become more packed Parent-mutant coexistence alters fixation and extinction times, resulting in exponentially long absorption times for moderately beneficial mutants when ecological effects are strong The theory provides a principled method to apply existing population genetics models with frequency-dependent selection even in the presence of complex communities This framework predicts mutation outcomes even in complex ecological settings
Where the Gains Come From
The gains come from establishing a quantitative link between ecological interactions and evolutionary trajectories, showing that evolutionary outcomes can be predicted using simple models even in the presence of complex community feedbacks The analytic expression for fixation probability (Eq. (5)) provides a generalization of Kimura’s formula in the context of complex ecological communities The ability to quantify the strength of ecological effects through η allows for direct inference of this parameter from strain dynamics alone, suggesting an interesting new empirical method The analysis reveals that selection is now frequency-dependent since the selection coefficient sinv − ηf changes with mutant frequency This result shows that ecological interactions strongly suppress fixation of neutral to moderately beneficial mutants, strongly deviating from the predictions of classical population genetics
Final Conclusion on Suppression
The study concludes that ecological interactions likely generically suppress fixation of neutral to moderately beneficial mutants, strongly deviating from the predictions of classical population genetics The suppression is strongest when sinv = 0 and α >> 1, where pfix(sinv = 0) ∼ exp(-2Neffsc), which is the same as Kimura’s prediction but for s = sinv The presence of a large η now also exponentially suppresses pfix with positive sinv up to sinv ≈ sc:= η/2 This suggests that the coexistence region acts as an additional barrier to fixation, causing extinction even after crossing the drift threshold when f∗ < 1/2 and sinv < sc The maximum mean extinction time reaches maximum when sinv = sc ≈ η/2, i.e.
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
-
Improved AI can predict mutation outcomes in complex ecological communities by integrating population genetics models with dynamical mean-field theory to capture emergent frequency-dependent selection between parents and mutants.
-
The improved system can derive an analytic expression for fixation probability that extends Kimura’s formula, revealing that
ecological interactions strongly suppress the fixation of moderately beneficial mutations.
-
The system can predict the time it takes for a mutant to reach fixation or extinction, showing that "when ecological effects are strong (alpha >> 1), the mean times to absorption (Fig. 3a) and extinction times (Fig. 3b) become exponentially large for moderately beneficial mutants."
-
The improved AI can quantify the strength of community-mediated feedbacks using a single emergent parameter eta, which
can be computed from a community’s statistical properties such as its nichepacking.
-
The system can predict that fixation probability is
exponentially suppressed
for neutral mutations when ecological effects are large (alpha >> 1), matching Kimura's prediction only when alpha is small. -
The improved system can identify the critical invasion fitness sc, where
fixation probability switches from roughly 0 to almost 1,
allowing for the prediction of which mutants will fix or go extinct based on their fitness relative to this threshold. -
The system can infer ecological feedback strength eta directly from high-resolution strain tracking data by estimating both f star and sinv from observed trajectories, suggesting
an interesting new empirical method to quantify the impact of ecology during evolutionary dynamics.
-
The improved AI can design predictors for demographic inference frameworks that incorporate eta into coalescent models, potentially improving accuracy where
ecology plays an important role in the evolutionary dynamics.
Related papers
- From life-history model to theories of ecological diversity: the impact of Madhav Gadgil's work
- A multiscale discrete-to-continuum framework for structured population models
- Diversification under Global Competition: A Cutoff-Induced Localization Mechanism
- Extending indoor airborne infection risk modelling to 3D, with local surface deposition
- Drivers of periodicity in population dynamic models of long-lived, large mammals
- Mutant Fixation for a Stochastic Evolutionary Model in Fragmented Populations