Biology subjects

Etienne, R.

Publications and source records attributed to Etienne, R..

17 recordsLinked to original sources

bioRxiv2026

Identifying Evolutionary Relatedness Effects on Diversification from Phylogenies using Neural Networks

Reconstructing the forces that shaped macroevolutionary histories from extant phylogenies is fundamentally challenging: richly parameterized diversification models are often only weakly identifiable; different evolutionary mechanisms can yield nearly indistinguishable tree shapes. Here we use a model with evolutionary relatedness dependence to evaluate how much information about such forces can be recovered from simulated trees. We train graph neural networks and long short-term memory classifiers to distinguish three scenarios of feedback of diversity on diversification: effect of phylogenetic diversity (total branch length), evolutionary distinctiveness (average phylogenetic distance of a species to all other species in a clade), and nearest-neighbor distance (phylogenetic distance to the mostly closely related species). We also train a suite of regression networks to estimate the underlying diversification parameters. We then analyze classification performance, calibration of predicted class probabilities, regression errors, and their dependence on tree size and on the strength and sign of richness and relatedness effects. Across network architectures and complexity levels, scenario classification is only moderately accurate and strongly asymmetric as revealed by the confusion matrix. Trees generated under an effect of nearest-neighbor distance on diversification tend to be correctly classified, whereas those with an effect of evolutionary distinctiveness are frequently misclassified. Regression networks systematically shrink predictions toward the empirical mean, especially for complex models, suggesting broad regions of parameter space with low identifiability. Strong global dependence of diversification rates on diversity further erodes recoverability by driving large variations in tree size that mask the subtler signatures of related-ness effects. In contrast, sufficiently strong speciation-relatedness effects can carve out narrow regions of parameter space in which scenarios and parameters become practically recoverable. Together, our results provide a map of when neural networks can and cannot infer diversification mechanisms from extant trees under our evolutionary relatedness dependence model, and they underscore the need for additional data or constraints when using flexible diversification models for macroevolutionary inference.

evolutionary biology↗

bioRxiv2025

The phylogenetic Janzen-Connell effect can explain multiple macroecological and macroevolutionary patterns

The classical Janzen-Connell (J-C) hypothesis states that species-specific natural enemies induce local-density dependence which explains high diversity observed in tropical tree communities. Recent studies have extended the hypothesis to include the fact that these natural enemies often attack phylogenetically related species as well, which reduces diversity and hence the explanatory power of the hypothesis. However, these studies consider ecological time scales at which the phylogenetic relatedness between species is constant, and thus ignore that at evolutionary time scales the population dynamics change the phylogeny. We use a spatially explicit eco-evolutionary model of the interplay between the dynamics of the phylogeny and the species abundances to study the predictions for common macroevolutionary and macroecological patterns. We confirm that hyperdiversity can be maintained by the J-C effect but the phylogenetic relatedness effect weakens the rarity advantage, reducing diversity. Incorporating protracted speciation further improves the result. Our model predicts a triphasic species-area relationship at much shorter time scales than in neutral scenarios. The species-abundance distribution can have one or two modes depending on the dispersal distance. Phylogenetic trees show diversification slowdowns and imbalance, consistent with empirical patterns. We study a new pattern of phylogenetic relatedness through space and find that small dispersal distance causes clusters of species with large phylogenetic distance to the community while large dispersal distance makes species distribute uniformly. As an illustration of how our models predictions can be compared to empirical data we study macroecological and phylogenetic patterns of the Barro Colorado Island (BCI) tree community. We conclude that the spatially explicit eco-evolutionary phylogenetic J-C effect can explain commonly observed macroevolutionary and macroecological patterns, providing an alternative way, while the traditional methods are prone, to demonstrate the role of J-C effect on species community assembly.

evolutionary biology↗

bioRxiv2025

Additive effects of environmental and demographic variation shape the repeatability of evolution across replicated experiments

The repeatability of evolution is fundamentally important for understanding the origin and diversification of life as well as for developing evolutionary forecasting tools. Repeatability is limited by stochasticity, here defined as changes that are independent of genotypic fitness effects. Over short timescales, the two main sources of stochasticity of evolutionary change are environmental stochasticity and demographic (life-history) stochasticity. Quantifying the relative importance of these two sources of stochasticity in driving fitness outcomes is crucially important for predicting evolutionary responses. To gain insights in the effects of stochasticity, five institutes replicated an evolutionary experiment exposing Caenorhabditis elegans to novel rearing conditions. Replication across the institutes led to variation in selective environments, e.g. through divergent microbiomes among institutes. Replication within institutes was done across demographic treatments that influence the potential for population-size dependent fluctuations in allele frequencies (drift) and genetic hitchhiking (draft). We found high among-institute variation in fitness outcomes, which was partially explained by variation in microbiota. Whereas lab-specific effects explained most of the variance in mean fitness, the repeatability of fitness outcomes depended more on demographic heterogeneity. Specifically, population bottlenecks resulted in high among-replicate variation in fitness. When combined, environmental and demographic stochasticity additively reduced repeatability, underlining their additive importance in developing evolutionary forecasting tools. These results further highlight the importance of statistically integrating heterogeneity in experimental evolution to identify factors constraining outcome repeatability and study replicability.

evolutionary biology↗

bioRxiv2025

Genomic plasticity of the Azospirillum genus in a biotechnological context

Extensive agriculture and the use of chemical fertilizers cause notable environmental impacts on multiple levels, from reducing soil microbiota diversity to groundwater contamination. In this context, the usage of plant growth-promoting bacteria (PGPB) presents a sustainable alternative to enhance crop production while mitigating these adverse effects. Azospirillum, a bacterial genus renowned for its beneficial capabilities, particularly phytohormone production, is a key component of many commercial inoculants. In this work, we performed a comparative genomic analysis of all publicly available Azospirillum genomes and four novel isolates belonging to our microbial collection. Our analysis identified a species complex within the genus, which we designate the A. brasilense species complex, comprising species already used in commercial bioconsortia. This complex is characterized by a core set of exclusive genes linked to chemotaxis and host-recognition capability. Furthermore, we also validated the biosafety of the A. brasilense species complex and confirmed the plant growth-promoting potential of our novel isolates, highlighting their suitability for developing new biofertilizers.

microbiology↗

bioRxiv2025

An Alternative Approach to Compute the Likelihood for the DAISIE (Dynamic Assembly of Island biota through Speciation, Immigration, and Extinction) Model

Understanding island biodiversity requires studying the processes driving its assembly and examining how these processes vary over time, across lineages, and across different islands. The DAISIE (Dynamic Assembly of Island biota through Speciation, Immigration, and Extinction) framework has been developed for this purpose, and has advanced our understanding of island community assembly by estimating, from phylogenetic data, the contribution of the processes of colonization, speciation and extinction to island community assembly. However, the model assumes uniform colonization and diversification rates across lineages, ignoring potential variation in these rates caused by, for example, lineage-specific traits. This assumption thus restricts the models capacity to capture complex colonization and diversification dynamics, and may consequently bias inference of colonization and diversification dynamics on islands. An extension of the framework behind these more complex dynamics is therefore desired. However, for this state-dependent colonization and diversification extension, the current computation of the likelihood of the model given phylogenetic data is computationally prohibitive. In this study, we therefore propose an alternative approach to computing the DAISIE likelihood, under the assumption of diversity-independent colonization and diversification. Our novel approach is based on the pruning algorithm (which calculates the likelihoods of comparative data given a tree and a model) that has been used for computing the likelihood of SSE (State-dependent Speciation and Extinction) models, which traces lineage history backward in time from the present day to the root of the tree. We demonstrate that our alternative approach reproduces DAISIEs predictions under diversity-independent colonization and diversification. This is important because no independent evaluation of the likelihood computation in DAISIE has been available. Our alternative approach not only confirms DAISIE results for diversity-independent colonization and diversification, it also gives more confidence in the results under diversity-dependence. Furthermore, our alternative approach offers greater computational efficiency. Finally, it provides a flexible framework for incorporating state-dependent dynamics in colonization, speciation, and extinction processes.

evolutionary biology↗

bioRxiv2024

Performance and Robustness of Parameter Estimation from Phylogenetic Trees Using Neural Networks

Species diversification is characterized by speciation and extinction, the rates of which can, under some assumptions, be estimated from time-calibrated phylogenies. However, maximum likelihood estimation methods (MLE) for inferring rates are limited to simpler models and can show bias, particularly in small phylogenies. Likelihood-free methods to estimate parameters of diversification models using deep learning have started to emerge, but how robust neural network methods are at handling the intricate nature of phylogenetic data remains an open question. Here we present a new ensemble neural network approach to estimate diversification parameters from phylogenetic trees that leverages different classes of neural networks (dense neural network, graph neural network, and long short-term memory recurrent network) and simultaneously learns from graph representations of phylogenies, their branching times and their summary statistics. Our best-performing ensemble neural network (which corrects graph neural network result using a recurrent neural network) can compute estimates faster than MLE and is less affected by tree size. Our analysis suggests that the primary limitation to accurate parameter estimation is the amount of information contained within a phylogeny, as indicated by its size and the strength of effects shaping it. In cases where MLE is unavailable, our neural network method provides a promising alternative for estimating phylogenetic tree parameters. If there are detectable phylogenetic signals present, our approach delivers results that are comparable to MLE but without inherent biases.

evolutionary biology↗

bioRxiv2023

Impact of Evolutionary Relatedness on Species Diversification: A New Birth-Death Model

Slowdowns in lineage accumulation are often observed in phylogenies of extant species. One explanation is the presence of ecological limits to diversity and hence to diversification. Previous research has examined whether and how species richness (SR) impacts diversification rates, but rarely considered the evolutionary relatedness (ER) between species, although ER can affect the degree of interaction between species, which likely sets these limits. To understand the influences of ER on species diversification and the interplay between SR and ER, we present a simple birth-death model in which the speciation rate depends on the ER. We use different metrics of ER that operate at different scales, ranging from branch/lineage-specific to clade-wide scales. We find that the scales at which an effect of ER operates yield distinct patterns in various tree statistics. When ER operates across the whole tree, we observe smaller and more balanced trees, with speciation rates distributed more evenly across the tips than in scenarios with lineage-specific ER effects. Importantly, we find that negative SR dependence of speciation masks the impact of ER on some of the tree statistics. Our model allows diverse evolutionary trajectories for producing imbalanced trees, which are commonly observed in empirical phylogenies but have been challenging to replicate with earlier models.

evolutionary biology↗

bioRxiv2023

Identifying summary statistics for approximate Bayesian computation in a phylogenetic island biogeography model

Estimation of parameters of evolutionary island biogeography models, such as colonization and diversification rates, is important for a better understanding of island systems. A popular statistical inference framework is likelihood-based estimation of parameters using island species richness and phylogenetic data. Likelihood approaches require that the likelihood can be computed analytically or numerically, but with the increasing complexity of island biogeography models, this is often unfeasible. Simulation-based estimation methods may then be a promising alternative. One such method is approximate Bayesian computation (ABC), which compares summary statistics of the empirical data with the output of model simulations. However, ABC demands the definition of summary statistics that sufficiently describe the data, which is yet to be explored in island biogeography. Here, we propose a set of summary statistics and use it in an ABC framework for the estimation of parameters of an island biogeography model, DAISIE (Dynamic Assembly of Island biota through Speciation, Immigration and Extinction). For this model, likelihood-based inference is possible, which gives us the opportunity to assess the performance of the summary statistics. DAISIE currently only allows maximum likelihood estimation (MLE), so we additionally develop a likelihood-based Bayesian inference framework using Markov Chain Monte Carlo (MCMC) to enable comparison with the ABC results (i.e., making the same assumptions on prior distributions). We simulated phylogenies of island communities subject to colonization, speciation, and extinction using the DAISIE simulation model and compared the estimated parameters using the three inference approaches (MLE, MCMC and ABC). Our results show that the ABC algorithm performs well in estimating colonization and diversification rates, except when the species richness or amount of phylogenetic information from an island are low. We find that compared to island species diversity statistics, summary statistics that make use of phylogenetic and temporal patterns (e.g., the number of species through time) significantly improve ABC inference accuracy, especially in estimating colonization and anagenesis rates, as well as making inference converge considerably faster and perform better under the same number of iterations. Island biogeography is rapidly developing new simulation models that can explain the complexity of island biodiversity, and our study provides a set of informative summary statistics that can be used in island biogeography studies for which likelihood-based inference methods are not an option.

evolutionary biology↗

bioRxiv2023

Inferring state-dependent diversification rates using approximate Bayesian computation (ABC)

State-dependent speciation and extinction (SSE) models provide a framework for quantifying whether species traits have an impact on evolutionary rates and how this shapes the variation in species richness among clades in a phylogeny. However, SSE models are becoming increasingly complex, limiting the application of likelihood-based inference methods. Approximate Bayesian computation (ABC), a likelihood-free approach, is a potentially powerful alternative for estimating parameters. One of the key challenges in using ABC is the selection of efficient summary statistics, which can greatly affect the accuracy and precision of the parameter estimates. In state-dependent diversification models, summary statistics need to capture the complex relationships between rates of diversification and species traits. Here, we develop an ABC framework to estimate state-dependent speciation, extinction and transition rates in the BiSSE (binary state dependent speciation and extinction) model. Using different sets of candidate summary statistics, we then compare the inference ability of ABC with that of using likelihood-based maximum likelihood (ML) and Markov chain Monte Carlo (MCMC) methods. Our results show the ABC algorithm can accurately estimate state-dependent diversification rates for most of the model parameter sets we explored. The inference error of the parameters associated with the species-poor state is larger with ABC than in the likelihood estimations only when the speciation rate is highly asymmetric between the two states ({lambda}1 /{lambda} 0 = 5). Furthermore, we find that the combination of normalized lineage-through-time (nLTT) statistics and phylogenetic signal in binary traits (Fitz and Purviss D) constitute efficient summary statistics for the ABC method. By providing insights into the selection of suitable summary statistics, our work aims to contribute to the use of the ABC approach in the development of complex state-dependent diversification models, for which a likelihood is not available.

evolutionary biology↗

bioRxiv2023

From Fossils to Living Canids: Two Contrasting Perspectives on Biogeographic Diversification

The Canidae are an ecologically important group of dog-like carnivores that arose in North America and spread across the planet around 10 million years ago. The current distribution patterns of species, coupled with their phylogenetic structure, suggest that Canidae diversification may have occurred at varying rates across different biogeographic areas. However, such extant-only analyses undervalued the groups rich fossil history because of a limitation in methods development. Current State-dependent Speciation and Extinction (SSE) models are (i) often parameter-rich which hinders reliable application to relatively small clades such as the Caninae (the only extant subclade of the Canidae consisting of 36 extant species); and (ii) often assume as possible states only the states that extant species present. Here we extend the SSE method SecSSE to apply to phylogenies with extinct species as well (111 Caninae species) and compare the results to those of analyses with the extant-species-only phylogeny. The results on the extant-species tree suggest that distinct diversification patterns are related to geographic areas, but the results on the complete tree do not support this conclusion. Furthermore, our extant-species analysis yielded an unrealistically low estimate of the extinction rate. These contrasting findings suggest that information from extinct species is different from information from extant species. A possible explanation for our results is that extinct species may have characteristics (causing their extinction), which may be different from the characteristics of extant species that caused them to be extant. Hence, we conclude that differences in biogeographic areas probably did not contribute much to the variation in diversification rates in Caninae.

evolutionary biology↗

bioRxiv2022

A simple island biodiversity model is robust to trait dependence in diversification and colonization rates

The application of state-dependent speciation and extinction (SSE) models to phylogenetic trees has revealed an important role for traits in diversification. However, this role remains comparatively unexplored on islands, which can include multiple independent clades resulting from different colonization events. Here, we perform a robustness study to identify how trait-dependence in rates of island colonization, extinction and speciation (CES rates) affects the estimation accuracy of a phylogenetic model that assumes no rate variation between trait states. We extend the DAISIE (Dynamic Assembly of Islands through Speciation, Immigration and Extinction) simulation model to include state-dependent rates, and evaluate the robustness of the DAISIE inference model using simulated data. Our results show that when the CES rate differences between trait states are moderate, DAISIE shows negligible error for a variety of island diversity metrics. However, for large differences in speciation rates, we find large errors when reconstructing clade size variation and non-endemic species diversity through time. We conclude that for many biologically realistic scenarios with trait-dependent speciation and colonization, island diversity dynamics can be accurately estimated without the need to explicitly model trait dynamics. Nonetheless, our new simulation model may provide a useful tool for studying patterns of trait variation.

evolutionary biology↗

bioRxiv2021

Social modulation of oogenesis and egg-laying in Drosophila melanogaster

Being part of a group facilitates cooperation between group members, but also creates competition for limited resources. This conundrum is problematic for gravid females who benefit from being in a group, but whose future offspring may struggle for access to nutrition in larger groups. Females should thus modulate their reproductive output depending on their social context. Although social-context dependent modulation of reproduction is documented in a broad range of species, its underlying mechanisms and functions are poorly understood. In the fruit fly Drosophila melanogaster, females actively attract conspecifics to lay eggs on the same resources, generating groups in which individuals may cooperate or compete. The tractability of the genetics of this species allows dissecting the mechanisms underlying physiological adaptation to their social context. Here, we show that females produce eggs increasingly faster as group size increases. By laying eggs faster in group than alone, females appear to reduce competition between offspring and increase their likelihood of survival. In addition, females in a group lay their eggs during the light phase of the day, while isolated females lay them during the night. We show that responses to the presence of others are determined by vision through the motion detection pathway and that flies from any sex, mating status or species can trigger these responses. The mechanisms of this modulation of egg-laying by group is connected to a lifting of the inhibition of light on oogenesis and egg-laying by stimulating hormonal pathways involving juvenile hormone. Because modulation of reproduction by social context is a hallmark of animals with higher levels of sociality, our findings represent a protosocial mechanism in a species considered solitary that may have been the target of selection for the evolution of more complex social systems.

animal behavior and cognition↗

bioRxiv2021

The effect of climate change on the distribution of Canidae

Land use by humans and climate change have been seriously affecting the distribution of species resulting in a quarter of all known mammals currently threatened with extinction. Here, we modeled the present and future potential distributions of all 36 extant Canidae species to evaluate their response to future climate scenarios. In addition, we tested if canids were likely to experience evolutionary rescue, which could allow some species to adapt to climate change. Our results suggest that global warming will cause most species to lose or maintain their ranges, while a few will have the potential to benefit from future conditions and considerably expand their geographic distributions. Some canids have the potential to experience evolutionary rescue, but Atelocynus microtis and Chrysocyon brachyurus are two concerning cases that do not show this capacity to adapt given the current pace of climate change. We also reveal that most Canidae hotspot regions are outside protected areas, which may be useful for the identification of key areas for conservation.

ecology↗

bioRxiv2020

A unified model of species abundance, genetic diversity, and functional diversity revealsthe mechanisms structuring ecological communities

Biodiversity accumulates hierarchically by means of ecological and evolutionary processes and feedbacks. Reconciling the relative importance of these processes is hindered by current theory, which tends to focus on a single spatial, temporal or taxonomic scale. We introduce a mechanistic model of community assembly, rooted in classic island biogeography theory, which makes temporally explicit joint predictions across three biodiversity data axes: i) species richness and abundances; ii) population genetic diversities; and iii) trait variation in a phylogenetic context. We demonstrate that each data axis captures information at different timescales, and that integrating these axes enables discriminating among previously unidentifiable community assembly models. We combine our massive eco-evolutionary synthesis simulations (MESS) with supervised machine learning to fit the parameters of the model to real data and infer processes underlying how biodiversity accumulates, using communities of tropical trees, arthropods, and gastropods as case studies that span a range of spatial scales.

evolutionary biology↗

bioRxiv2019

Performance in a novel environment subject to ghost competition.

BackgroundA central tenet of the evolutionary theory of communities is that competition impacts evolutionary processes such as local adaptation. Species in a community exert a selection pressure on other species and may drive them to extinction. We know, however, very little about the influence of unsuccessful or ghost species on the evolutionary dynamics within the community.\n\nMethodsHere, we studied the long-term influence of a ghost competitor on the performance of a more successful species using experimental evolution. We transferred the spider mite Tetranychus urticae onto a novel host plant under initial presence or absence of a competing species, the congeneric mite T. ludeni.\n\nResultsThe latter species unintentionally went extinct soon after the start of the experiment, but we nevertheless completed the experiment and found that the initial density of this ghost competitor positively affected the performance (i.e. fecundity) of the more successful species.\n\nThis effect on T. urticae even lasted for at least 25 generations.\n\nDiscussionOur study supports the hypothesis that early experienced selection pressures can exert a persistent evolutionary signal on species performance in novel environments.

evolutionary biology↗

bioRxiv2017

A simple spatially explicit neutral model explains range size distribution of reef fishes

AimThe great variation in range sizes among species has fascinated ecologists for decades. In reef-associated fish species, which live in fragmented habitats and adopt a wide range of dispersal strategies, we may expect species with greater dispersal ability to spread over larger ranges. However, empirical evidence for such a positive relationship between dispersal and range size in reef fishes remains scarce. Here, we unveil the more nuanced role of dispersal on the range size distribution of reef associated fishes using empirical data and a novel spatially explicit model.\n\nLocationTropical Eastern Pacific\n\nMajor taxa studiedReef-associated fishes\n\nMethodsWe estimated range size distributions for six different guilds of all reef-associated fishes with different dispersal abilities. We used a one-dimensional spatially explicit neutral model, which simulates the distribution of species along a linear coastline to explored the effect of dispersal, speciation and sampling on the distribution of range sizes. Our model adopts a more realistic gradual speciation process (protracted speciation) and incorporates important long distance dispersal events with a fat-tail dispersal kernel. We simulated our model using a highly efficient coalescence approach, which guarantees the metacommunity, is sampled at dynamic equilibrium. We fitted the model to the empirical data using an approximate Bayesian computation approach, with a sequential Monte Carlo algorithm.\n\nResultsStochastic birth, death, speciation and dispersal events alone can accurately explain empirical range size distributions for six different guilds of tropical, reef-associated fishes. Variation in range size distributions among guilds are explained purely by differences in dispersal ability with the best dispersers covering larger ranges.\n\nMain conclusionsA simple combination of neutral processes with guild-specific dispersal ability provides a general explanation for both within- and across-guild range size variation. Our results support the theoretically expected, but empirically much debated, hypothesis that dispersal promotes range size.

ecology↗

bioRxiv2016

Inferring the role of habitat dynamics in driving diversification: evidence for a species pump in Lake Tanganyika cichlids

Geographic isolation that drives speciation is often assumed to slowly increase over time, for instance through the formation of rivers, the formation of mountains or the movement of tectonic plates. Cyclic changes in connectivity between areas may occur with the advancement and retraction of glaciers, with water level fluctuations in seas between islands or in lakes that have an uneven bathymetry. These habitat dynamics may act as a driver of allopatric speciation and propel local diversity. Here we present a parsimonious model of the interaction between cyclical (but not necessarily periodic) changes in the environment and speciation, and provide an ABC-SMC method to infer the rates of allopatric and sympatric speciation from a phylogenetic tree. We apply our approach to the posterior sample of an updated phylogeny of the Lamprologini, a tribe of cichlid fish from Lake Tanganyika where such cyclic changes in water level have occurred. We find that water level changes play a crucial role in driving diversity in Lake Tanganyika. We note that if we apply our analysis to the Most Credible Consensus (MCC) tree, we do not find evidence for water level changes influencing diversity in the Lamprologini, suggesting that the MCC tree is a misleading representation of the true species tree. Furthermore, we note that the signature of habitat dynamics is found in the posterior sample despite the fact that this sample was constructed using a species tree prior that ignores habitat dynamics. However, in other cases this species tree prior might erase this signature. Hence we argue that in order to improve inference of the effect of habitat dynamics on biodiversity, phylogenetic reconstruction methods should include tree priors that explicitly take into account such dynamics.

evolutionary biology↗