Programme 2017

Le MEE se tiendra les 30 et 31 mai 2017 à l’amphitéâtre Charles Flahaut de l’Institut de Botanique de Montpellier.

Mardi 30 mai :

8h30 – 9h00 : Accueil (Viennoiseries & Cafés)

9h00 – 9h45 : Chercheur invité : J.B. André (ENS, IEC, Paris) : Modeling the evolution of cooperation and fairness

I will present simple mathematical principles at the core of the evolution of cooperation and explain how they can help understand the nature of human sociality. I will then show how other approaches based on robotics can also help us understand the constraints that influence the evolution of social behaviors. 

9h50 – 10h10 : L.-P. Nguyen : On the difficult evolutionary transition from the free-living lifestyle to obligate symbiosis

The transition from free-living organisms to symbiotic complexes is one of the major transitions in evolution (MaynardSmith, 1997), which is thought to proceed in a stepwise manner: from free-living individuals, via facultative associations to obligate associations. The fact that obligate associations seem far more abundant than facultative associations suggests that facultative association is a mere transitory step in this major evolutionary transition. However, one may ask why this facultative phase would be so unstable as strains in facultative associations can benefit from both the hosts and the external environment. Moreover, the host environment may actually be quite hostile. Therefore, there may be conditions under which it is beneficial to stay only facultatively associated and thus use both mode of life. Employing the adaptive dynamic approach, we analyse a simple mathematical model and show that i) benefit from an association and host coexistence are prerequisite for both facultative and obligate association; ii) free-living lifestyle, facultative and obligate symbiosis form a continuum of the interdependency of two partners in an association just like the parasitism-mutualism continuum; iii) there is a link between the costliness of a transmission strategy and the degree of commitment to the associations, hence having the knowledge of one feature may help predict the outcome of the other.

10h15 – 10h35 : G. Doulcier : Modeling Nested Darwinian Populations

The hierarchical organisation of living systems, and concomitant opportunity for selection to operate at multiple levels, raises particular challenges for development of mathematical models. Major evolutionary transitions require additional care because of the need to capture the process by which higher levels of biological organisation becomes units of evolution in their own right. I will describe a general framework in which individuals as well as collectives of individuals follow Darwinian  processes. These models are inspired by laboratory experiments performed in state of the art millifluidic instruments in which egalitarian transitions, such as that leading to evolution of the eukaryotic cell, can be studied in real time. We hypothesize that this situation will lead to the
progressive selection of improved Darwinian properties at the collective level (namely increased reproductive output but also increased heredity) as individual level conflicts are solved. By closely examining the evolutionary consequences arising from the nesting of two Darwinian processes at two timescales we hope to reveal mechanisms underpinning the emergence of complex life.

10h40 – 11h00 : Cafés

11h00 – 11h20 : L. Rimbaud : Assessing the performance of landscape-based strategies to deploy major-gene resistance

Genetically-controlled plant resistance can reduce the damage caused by pathogens. However, pathogens have the ability to evolve and overcome such resistance. This often occurs very quickly after resistance is deployed, resulting in significant crop losses and a continuing need to breed new resistant cultivars. To tackle this issue, several strategies have been proposed to constrain the evolutionary potential of pathogen populations and thus increase the durability of resistance deployment. These strategies mainly rely on using different combinations of resistance sources in time, space, or both. In time, such combination consists of crop rotations. In space, resistance sources can be deployed in the same cultivar (pyramiding), in different cultivars within the same field (cultivar mixtures) or in different fields (mosaics). However, experimental assessment of the efficiency (i.e. ability to reduce disease impact) and the durability (i.e. ability to limit pathogen evolution and delay resistance breakdown) of different deployment strategies presents a major challenge.

Therefore, we developed a spatially-explicit stochastic model to assess the epidemiological and evolutionary outcomes of the major deployment options described above when one or two major genes for resistance are present. In addition, we analysed the impact of landscape organisation (as defined by the proportion of fields cultivated with a resistant cultivar, and their spatial aggregation) and epidemiological or evolutionary parameters (e.g. dispersal abilities, mutation rate, cost of infectivity) through sensitivity analyses and polynomial regression.

The model has been parameterised for wheat qualitative resistance to rusts, caused by fungi of the genus Puccinia, but can be applied to many other pathosystems. Early results suggest that strategies offering the best epidemiological control of the disease are not necessarily the most durable.

11h25 – 11h45 : M. Sofonea : Con Ebola Virus evolve to be less virulent in humans ?

Understanding Ebola Virus (EBOV) virulence evolution is not only timely but also raises specific questions because it causes on the most virulent human infections and it is capable of transmission after the death of its host. Using a compartmental epidemiological model that captures all of the virus’ transmission routes, we infer the evolutionary dynamics of case fatality ratio (CFR) on the scale of an outbreak and on the long term. Our major finding is that the virus’s specific life cycle imposes selection for high levels of virulence and that this pattern is robust to parameter variations in biological ranges. In addition to shedding a new light on the adaptive reasons of EBOV’s high virulence, these results generate testable predictions and contribute to informing public health policies. In particular, burial management stands out as the most appropriate intervention since it decreases the R0 of the epidemics, while imposing selection for less virulent strains.

12h00 – 13h30 : Repas

13h30 – 14h15 : Chercheur invité : S. Boitard (GenPhySE, Toulouse): Accounting for Linkage Disequilibrium in genome scans for selection without individual genotypes: the local score approach

Detecting genomic footprints of selection is an important step in the understanding of evolution. Accounting for linkage disequilibrium in genome scans increases detection power, but haplotype-based methods require individual genotypes and are not applicable on pool-sequenced samples. We propose to take advantage of the local score approach to account for linkage disequilibrium in genome scans for selection, cumulating (possibly small) signals from single markers over a genomic segment, to clearly pinpoint a selection signal. Using computer simulations, we demonstrate that this approach detects selection with higher power than several state-of-the-art single marker, windowing or haplotype-based approaches. We illustrate this on benchmark data sets including individual genotypes, for which we obtain similar results with the local score and one haplotype-based approach. Finally, we apply the local score approach to Pool-Seq data obtained from a divergent selection experiment on behavior in quail, and obtain precise and biologically coherent selection signals: while competing methods fail to highlight any clear selection signature, our method detects several regions involving genes known to act on social responsiveness or autistic traits. Although we focus here on the detection of positive selection from multiple population data, the local score approach is general and can be applied to other genome scans for selection or other genome-wide analyses such as GWAS

14h20 – 14h40 : M. Vérin : Host-parasite coevolution promotes the evolution of seed banking as a bet hedging strategy

Seed banking is a common bet-hedging strategy maximizing the fitness of organisms facing environmental stochasticity. Yet, this stochasticity is often viewed as fast and drastic shifts between good and bad years. We study here the evolution of the seed banking strategy in a gene-for-gene host-parasite system, characterized by indirect frequency-dependent selection at the coevolving loci. Cycles of coevolution promote a gradually changing environment for hosts and parasites, which drives and is driven in return by changes in the seed banking strategy. We study the evolution of host germination rate as a quantitative trait using both pairwise competition and multiple mutant competitions methods. The evolution of an optimal germination strategy is studied for fixed short and long term persistent seed bank, and for memoryless and non-memoryless germination functions. We demonstrate that coevolution promotes the evolution of several optimal seed bank strategy, depending on the speed and amplitude of coevolutionary cycles. Importantly, under a long term persistent bank and linkage disequilibrium between the coevolving locus and the germination loci, resistant hosts evolve a strong seed bank, while susceptible hosts do not. Our results provide a general understanding of the complex interplay between coevolutionary dynamics and evolution of seed banking strategies.

14h45 – 15h05 : A. Picot : The evolution of siderophore production as a competitive trait

Microbes have the potential to be highly cooperative organisms. The archetype of microbial cooperation is often considered to be the secretion of siderophores, molecules scavenging iron, where cooperation is threatened by “cheater” genotypes that use siderophores without making them. Here, we show that this view neglects a key piece of biology: siderophores are imported by specific receptors that constrain their use by competing strains. We study the effect of this specificity in an ecoevolutionary model, in which we vary siderophore sharing among strains, and compare fully shared siderophores with private siderophores. We show that privatizing siderophores fundamentally alters their evolution. Rather than a canonical cooperative good, siderophores become a competitive trait used to pillage iron from other strains. We also study the physiological regulation of siderophores using in silico long-term evolution. Although shared siderophores evolve to be downregulated in the presence of a competitor, as expected for a cooperative trait, privatized siderophores evolve to be upregulated. We evaluate these predictions using published experimental work, which suggests that some siderophores are upregulated in response to competition akin to competitive traits like antibiotics. Although siderophores can act as a cooperative good for single genotypes, we argue that their role in competition is fundamental to understanding their biology.

15h10 – 15h30 : Cafés

15h30 – 15h50 : J. Rode : Differential equation system of a theoretical ecosystem

This presentation will develop the main challenges and progress on the derivation of a generic differential equation system to improve our understanding of the behaviour of any ecosystem on the long term (scale of thousands of years, basically during the Holocene period). Indeed, the main aim of this equation system would be to model the whole ecosystem (flora, fauna, soil, atmosphere and mankind) in a parsimonious and generic enough way to be applied on different ecosystem types or biomes such as tropical forest, boreal forest, tundra, desert and ocean. Few different approaches will be developed, both theoretical and based on a data set.

15h55 – 16h15 : C. Carpentier & J. Centanni : Integrated modelling of ecosystems – theoretical approach and applications

Many theoretical approaches have been developed to study ecological systems and provide a basis for their understanding and management. One of them consists in representing ecological systems as networks: the components of the system are the nodes of a graph and these nodes are linked to each other by edges representing their interactions. Based on this approach, we are developing a qualitative discrete model integrating biotic, abiotic and anthropic components linked by some rules (processes). This model can then be used to study the trajectories followed by the ecosystem and, thereby, its stability properties. This model allows to determine the impact of different ecosystem management strategies (i.e. grazing impacts on biodiversity in Camargue) and to optimise and compromise between ecosystem services and biodiversity conservation (i.e. a mountain ecosystem, Chamrousse).

16h20 – 17h00 : Chercheur invité : C. Gaucherel (AMAP, Montpellier): Integrated modelling and development of an ecosystem

Complex ecosystems are difficult to model and are still poorly understood. In spite of the long lasting efforts to predict them, we still lack a real integrated framework to grasp their behaviours. In this paper, the concept of ecosystem development is proposed to understand the sharp regime shifts (Scheffer et al. 2001, Brook et al. 2013) they cumulate over the long term. To handle regime shifts, we develop an integrated model merging the physical, biological and social components and interactions (processes) of an ecosystem into a single graph representation. We then formalize sharp structural (topological) changes of the system with Boolean networks or Petri nets (Pommereau 2010, Gaucherel et al. 2017).

We illustrate this theoretical approach borrowed from discrete model formalisms on some ecosystems assumed to be representative of most ecosystems: the eusocial (termite and ant) insect colonies and their associated interactions (Hölldobler and Wilson 2009, Turner 2009). Boolean networks or Petri nets, combined with rigorous production/rewriting rules and syntax, simulate the ecosystem’s responses to strong perturbations. This deterministic and qualitative approach allows identifying fragile components, analysing ecosystem resilience to perturbations and to map out every lethal trajectory of the ecosystem. The rigorous and parsimonious abilities of such discrete models pave the way in reinterpreting and managing a wide range of ecosystems. I will also show preliminary results on more realistic ecosystems such as the Camargue area or Alp forests.

Brook, B. W., E. C. Ellis, M. P. Perring, A. W. Mackay, and L. Blomqvist. 2013. Does the terrestrial biosphere have planetary tipping points? Trends in Ecology & Evolution 28:396-401.
Gaucherel, C., H. Théro, A. Puiseux, and V. Bonhomme. 2017. Understand ecosystem regime shifts by modelling ecosystem development using Boolean networks. In review.
Hölldobler, B. and E. O. Wilson. 2009. The Superorganism. The Beauty, Elegance, and Strangeness of Insect Societies. W.W. Norton & Company, New-York, London.
Pommereau, F. 2010. Algebras of coloured Petri nets. Lambert Academic Publishing (LAP).
Scheffer, M., S. Carpenter, J. A. Foley, C. Folke, and B. Walker. 2001. Catastrophic shifts in ecosystems. Nature 413:591-596.
Turner, J. S. 2009. The extended organism: the physiology of animal-built structures. Harvard University Press.

Mercredi 31 mai :

8h30 – 9h00 : Accueil (Viennoiseries & Cafés)

9h00 – 9h00 : Chercheur invité : S. Kéfi (ISEM, Montpellier): Why and how does the diversity of interactions matter?

In natural communities, species interact with each other in different ways, including predation, competition, and facilitation. This complex spectrum of interactions constitutes a network of links that mediates ecological communities’ response to perturbations, such as exploitation and climate change. In the last decades, there have been great advances in the study of intricate ecological networks. We have, nonetheless, lacked both the data and the tools to more rigorously understand the patterning of multiple interaction types between species, as well as their consequences for community dynamics. Improving our understanding of the dynamics and resilience of complex ecological systems may rely on how the joint effects of different interactions types explain variations not explained by feeding interactions alone. In this talk, I’ll present recent efforts in analyzing and understanding ecological networks including different types of species interactions. I will argue that moving beyond unidimensional analyses of ecological networks may contribute to improving our understanding and predictive capacity of the way ecological systems respond to disturbances.

9h50 – 10h10 : P. Quévreux : Effects of relations between green and brown food webs on food web stability and ecosystem functioning

Theoretical food web models generally forget that food webs are split into the green and the brown food webs. The green food web produces biomass through photosynthesis and the brown food web recycles nutrients contained in the dead organic matter through decomposition. These two food webs are thus mutualistic but they can be competitive as well. In fact, decomposers also take up mineral nutrients depending on the stoichiometry of the dead organic matter and thus compete with primary producers. We propose here a model including a size structured food web with an allometric scaling of biological rates of species and a stoichiometric model of nutrient cycling through the living and non living compartments. We studied the impact of three factors on the stability and the functioning of the food web: the connexion between the green and the brown food web, the competitive advantage of decomposers relatively to primary producers for nutrient uptake and the stoichiometry of dead organic matter consumption by decomposers. We also tracked the time dynamics of energy and matter flows between the green and the brown food webs thanks to an adaptive foraging model, that enables consumers to adapt their diet to maximise their biomass absorption.

10h15 – 10h35 : T. Koffel : A theoretical study of facilitative succession and ecosystem development by nitrogen fixers 

Symbiotic nitrogen (N)-fixing organisms such as actinorhizal plants and legumes tend to thrive during primary succession, as typical bedrocks lack available N. In turn, fixed N accumulates in soils through biomass turnover and recycling, benefiting the whole community. Yet, it is unclear how this facilitation mechanism interacts with competition for other nutrients, e.g. phosphorus (P), and when this leads to succession mostly driven by facilitation. Here, we introduce a resource-explicit, community assembly model of N-fixing species competing for N and P and analyze successional trajectories along resource availability gradients using a recently developed extension of graphical resource competition theory.
We show that facilitative succession only occurs under very low N availability. It relies on the initial invasion of the bare substrate by the most efficient N-fixing strategies, sequentially displaced by colonizers that are more competitive but whose establishment relies on soil N accumulated by previous strategies. Facilitative succession comes with two characteristic signatures. First, the very strict order in which species can replace each other filters out the inherently random colonization process, leading to relatively ordered succession trajectories. Second, the late successional ecosystem presents alternate stable states, making late succession inherently prone to catastrophic shifts. Conversely, high environmental N availability inhibits N-fixation, leading to competition-dominated succession. Put together, these results contribute to an enriched version of Tilman’s resource-ratio theory of succession.

10h40 – 11h00 : Cafés

11h00 – 11h20 : P. Volte : Use of power tests before establishing a protocol to determine the quality of future occupancy results

Since march 2017, ECO-MED began a project to establish the detection rates of reptiles. The aim is to increase the quality of our surveys. Due to the inherent limitations in the operational capacities of the company, the protocol had to be the least intrusive possible for field expert. It has been decided to work with single season occupancy models. Two operating variables were used: the number of sites and the repetitions per sites. It is known that accuracy of studied parameters improves as these variables amount increase. We decided not to limit the number of samples but to fit with the maximum operational capacities of the company. However, we can question the level of precision that will be reached if we stay within these limits?

We decided to conduct power tests with different scenarios for each taxa. The tests required occupancy and detection input rates that were based on experts’ advice thus allowing us to generate simulations specific to each species. We developed an R script using the “unmarked” package. Generated graphs show the difference between the lower and upper limits of studied rates (= precision) as a function of the number of sites and repetitions (= operational capacities) for each species.

11h25 – 11h45 : N. Bertoldi : Stratégies évolutivement stables et équilibres de Nash : deux exemples d’un seul et même principe ?

Le concept de stratégie évolutivement stable constitue la clef de voûte de la théorie des jeux évolutionnaires. Telle que John Maynard Smith la définit, une stratégie évolutivement stable est un phénotype comportemental qui, une fois adopté par tous les membres d’une population, empêche à tout phénotype mutant d’envahir ladite population, sous l’action de la sélection naturelle. Elle rappelle ainsi l’un des concepts fondamentaux de la théorie générale des jeux, à savoir l’équilibre de Nash pour des jeux non-coopératifs. Un équilibre de Nash se définit, en effet, comme un ensemble de choix tel qu’aucun acteur ne peut espérer augmenter son gain tout en se démarquant de ce profil d’action. Il semblerait donc possible de définir une stratégie évolutivement stable comme étant un équilibre de Nash où les acteurs seraient des organismes qui pourraient choisir leur propre phénotype (ou bien celui de leurs enfants) et qui seraient en compétition pour l’accès aux ressources, ou encore à la reproduction, dans des conditions environnementales données. De plus, les gains correspondant à chaque stratégie devraient être compris comme des variations de la fitness darwinienne de tel ou tel phénotype. Mon intervention vise précisément à évaluer la pertinence d’un tel rapprochement, en mettant en lumière les analogies formelles qu’il est possible d’établir entre des modèles illustrant l’un et l’autre principe. Il s’agira en outre d’analyser les présupposés qui s’avèrent nécessaires pour mettre en œuvre une telle analogie, ainsi que le rapport de tels principes au cadre conceptuel plus général de la théorie de l’évolution.

12h00 – 13h30 : Repas

13h30 – 14h15 :  Chercheur invité : Y.-J.Shin (MARBEC, Montpellier) : Modelling the combined effects of fishing and environmental changes on marine ecosystems

In this presentation, we provide a few examples where scenarios and ecosystem models can provide useful support to ecosystem-based fisheries management. In particular, we explore the risks associated to the synergistic ecosystem impacts of fishing and environmental change across a suite of marine ecosystems using four state-of-the-art ecosystem models (EwE, OSMOSE, Atlantis, Multispecies Size Spectrum). Owing to the structural differences between the models, simulating changes in multiple drivers can potentially produce a wide array of ecosystem responses to changes. However, using standardized simulation experiments, similar responses obtained across this variety of ecosystem and models can help identify plausible scenarios of ecosystem change.

14h20 – 14h40 : T. Aubier : Speciation along ecological gradients and the costs of choosiness

Recently, research on speciation has shifted from a focus on the geographical context of differentiation to a focus on the role of natural and sexual selection in reproductive isolation. Yet, the spatial context may be of prime importance in the feasibility of speciation, and indeed, species often replace each other spatially along environmental gradients. In fact, we still lack good theoretical predictions to understand the role of the spatial dimension of ecological changes in parapatric speciation. In particular, the effects of the costs of choosiness have been ignored so far. Ecological speciation may fail if choosy females experience fitness costs. Notably, choosy females may remain unmated if they can assess only few mates in their lifetime. In parapatry, this effect may be especially strong because the relative proportion of the two taxa change across the ecological gradient, and females with uncommon phenotypes indeed risk remaining unmated if they are looking for a similar mate. Unfortunately, many theoretical models have used normalized mating probabilities so that all females eventually find a mate.

Here, using a spatially-explicit individual-based model, we relax this hypothesis and allow females to fail finding a mate if they are too choosy. We model parapatric populations where disruptive natural selection leads to divergence in an ecological trait, and where assortative mating can evolve, limiting gene flow between incipient species. We show that the size of the transition zone (or hybrid zone) and the number of males each female may encounter have a major influence on the probability of speciation. Intriguingly, we show that speciation along ecological gradients is hindered not only when females may encounter very few males, but also when they may encounter many males. Our predictions are explained by the direction of selection on the strength of mate choice of primarily affecting (maladapted) migrants dispersing through space.

14h45 – 15h05 : T. Brom : Maintenance of gametophytic self-incompatibility system in spatially structured population

Many flowering plants exhibit a mate choice based on pollen rejection. This self-incompatibility system enables hermaphroditic plants to recognize their own pollen and prevent its germination on their own pistil. In gametophytic self-incompatibility system, self-recognition depends on a multiallelic locus, the S locus. A mutation in the S locus can disable the expression of recognition specificities in pollen, pistil or both and allowed for self-reproduction. Self-compatible mutants benefit of a better propagation because they can self-pollinate and do not reject or be rejected by any potential mate. However, individuals born from self-fecundation can suffer inbreeding depression, decreasing their fitness. Previous theoretical works show that in panmictic populations the maintenance of self-incompatibility system is possible only in a reduce parameters range of self-pollination rate and inbreeding depression. Some authors made the hypothesis that spatial structuration of populations can be a key to understand the puzzling evolution and maintenance of self-incompatibility system.

We present an individual based model of hermaphroditic plants with gametophytic self-incompatibility system in a spatially structured metapopulation. We assessed the resistance of this system to the invasion by self-compatible mutants under various conditions of spatial structuration of the population. We showed that spatial structuration can have important effects on the inbreeding depression value required for self-incompatible system to resist to the invasion by a self-compatible mutant.

15h10 – 15h30 : Cafés

15h30 – 15h50 : O. Cotto : A dynamic eco-evolutionary model predicts slow response to alpine plants to climate warming

Withstanding extinction while facing rapid climate change depends on a species ability to track its ecological niche or evolve a new one. Current methods that predict climate-driven species range shifts use ecological modeling without eco-evolutionary dynamics. Here, we present an eco-evolutionary forecasting framework that combines niche modeling with individual-based demographic and genetic simulations. Applying our approach to four endemic perennial plant species of the Austrian Alps, we show that accounting for eco-evolutionary dynamics when predicting species responses to climate change is crucial. Perennial species persist in unsuitable habitats longer than predicted by niche modeling, causing delayed range losses; but their evolutionary responses are constrained because long-lived adults produce increasingly maladapted offspring. Decreasing population size due to maladaptation occurs faster than the contraction of the species range, especially for the most abundant species. Monitoring of species local abundance rather than their range may likely better inform on species extinction risks under climate change.

15h55 – 16h15 : N. Harmand : An adaptative fitness landscape model across environments

Fitness landscape models have a long history in evolutionary biology, as they provide a rich and easily
visualized topographical metaphor to describe adaptive processes. Phenotype-to-fitness landscapes in single environments have been widely used as purely heuristic models to provide mathematical formalizations of the mutational inputs along an adaptive trajectory. Coupled with standard population/quantitative genetics, they allow a complete theoretical description of the evolutionary process (mutation-selection). Such a framework, inspired by Fisher’s geometrical model, has proved very powerful to formulate general predictions that can be related to empirical patterns of selection or distributions of mutation fitness effects. With these data, it is thus possible to test and calibrate a fitness landscape, opening large opportunities to extrapolate predictions on long-term adaptation. However those landscapes remain widely described in a simple ecological context that stands as a limitation to model adaptation in natural heterogeneous environments. In this talk, I will present methods and applications to reveal the topography of a multi-environmental fitness landscape model from empirical patterns of short and long term adaptation across several environments along a gradient.

16h20 – 17h00 : Chercheur invité : G. Martin (ISEM, Montpellier) : Fisher’s geometrical model emerges as a property of complex integrated phenotypic networks

Models relating phenotype to fitness (phenotype-fitness landscapes) have seen important developments recently. They can roughly be divided into mechanistic models (e.g., metabolic networks) and more heuristic models like Fisher’s geometrical model. Each has its own drawbacks, but both yield testable predictions on how the context (genomic background or environment) affects the distribution of mutation effects on fitness and thus adaptation, and both have received some empirical support.

I will propose a derivation of Fisher’s model « from first principles », where the model emerges as a limit of a more general model, with features typical of phenotypic networks. Mutations within a given target (e.g. a gene set) are assumed to affect some set of unspecified phenotypic traits, which in turn affect fitness. Four main qualitative assumptions are further imposed : (i) high pleiotropy (any mutation in the target affects the same large set of traits) (ii) integration (these traits jointly determine a much smaller set of « adaptation traits ») (iii) optimization (« adaptation traits » determine fitness, via some optimizing function) and (iv) limited maladaptation (the background considered is not too far from a local fitness optimum, in phenotype space). Otherwise, the model remains fairly general regarding the phenotypic processes involved or the distribution of mutation effects. A statistical limit is obtained under these assumptions, by importing results from random matrix theory. This yields a fitness landscape that is effectively Fisher’s model: either in its isotropic form (all dimensions are equivalent) or a simple anisotropic version (a single dominant dimension). These results suggest that the premises of Fisher’s model might be more realistic than often claimed, which might explain its success in predicting several observed patterns of mutation effects on fitness, across species and environments. They also suggest testable insights on which features of mutation fitness effects may vary (or not) across contexts.