Deep learning approaches to landmark detection in tsetse wing images.

Morphometric analysis of wings has been suggested for identifying and controlling isolated populations of tsetse (Glossina spp), vectors of human and animal trypanosomiasis in Africa. Single-wing images were captured from an extensive data set of field-collected tsetse wings of species Glossina pallidipes and G. m. morsitans. Morphometric analysis required locating 11 anatomical landmarks on each wing. The manual location of landmarks is time-consuming, prone to error, and infeasible for large data sets. We developed a two-tier method using deep learning architectures to classify images and make accurate landmark predictions. The first tier used a classification convolutional neural network to remove most wings that were missing landmarks. The second tier provided landmark coordinates for the remaining wings. We compared direct coordinate regression using a convolutional neural network and segmentation using a fully convolutional network for the second tier. For the resulting landmark predictions, we evaluate shape bias using Procrustes analysis. We pay particular attention to consistent labelling to improve model performance. For an image size of 1024 × 1280, data augmentation reduced the mean pixel distance error from 8.3 (95% confidence interval [4.4,10.3]) to 5.34 (95% confidence interval [3.0,7.0]) for the regression model. For the segmentation model, data augmentation did not alter the mean pixel distance error of 3.43 (95% confidence interval [1.9,4.4]). Segmentation had a higher computational complexity and some large outliers. Both models showed minimal shape bias. We deployed the regression model on the complete unannotated data consisting of 14,354 pairs of wing images since this model had a lower computational cost and more stable predictions than the segmentation model. The resulting landmark data set was provided for future morphometric analysis. The methods we have developed could provide a starting point to studying the wings of other insect species. All the code used in this study has been written in Python and open sourced.

Fisheries-induced disruptive selection.

Commercial harvesting is recognized to induce adaptive responses of life-history traits in fish populations, in particular by shifting the age and size at maturation through directional selection. In addition to such evolution of a target stock, the corresponding fishery itself may adapt, in terms of fishing policy, technological progress, fleet dynamics, and adaptive harvest. The aim of this study is to assess how the interplay between natural and artificial selection, in the simplest setting in which a fishery and a target stock coevolve, can lead to disruptive selection, which in turn may cause trait diversification. To this end, we build an eco-evolutionary model for a size-structured population, in which both the stock׳s maturation schedule and the fishery׳s harvest rate are adaptive, while fishing may be subject to a selective policy based on fish size and/or maturity stage. Using numerical bifurcation analysis, we study how the potential for disruptive selection changes with fishing policy, fishing mortality, harvest specialization, life-history tradeoffs associated with early maturation, and other demographic and environmental parameters. We report the following findings. First, fisheries-induced disruptive selection is readily caused by commonly used fishing policies, and occurs even for policies that are not specific for fish size or maturity, provided that the harvest is sufficiently adaptive and large individuals are targeted intensively. Second, disruptive selection is more likely in stocks in which the selective pressure for early maturation is naturally strong, provided life-history tradeoffs are sufficiently consequential. Third, when a fish stock is overexploited, fisheries targeting only large individuals might slightly increase sustainable yield by causing trait diversification (even though the resultant yield always remains lower than the maximum sustainable yield that could be obtained under low fishing mortality, without causing disruptive selection). We discuss the broader implications of our results and highlight how these can be taken into account for designing evolutionarily informed fisheries-management regimes.

A mathematical model of 'Pride and Prejudice'.

A mathematical model is proposed for interpreting the love story between Elizabeth and Darcy portrayed by Jane Austen in the popular novel Pride and Prejudice. The analysis shows that the story is characterized by a sudden explosion of sentimental involvements, revealed by the existence of a saddle-node bifurcation in the model. The paper is interesting not only because it deals for the first time with catastrophic bifurcations in romantic relation-ships, but also because it enriches the list of examples in which love stories are described through ordinary differential equations.

Current and future scenarios of suitability and expansion of cassava brown streak disease, Bemisia tabaci species complex, and cassava planting in Africa.

Cassava (Manihot esculenta) is among the most important staple crops globally, with an imperative role in supporting the Sustainable Development Goal of 'Zero hunger'. In sub-Saharan Africa, it is cultivated mainly by millions of subsistence farmers who depend directly on it for their socio-economic welfare. However, its yield in some regions has been threatened by several diseases, especially the cassava brown streak disease (CBSD). Changes in climatic conditions enhance the risk of the disease spreading to other planting regions. Here, we characterise the current and future distribution of cassava, CBSD and whitefly Bemisia tabaci species complex in Africa, using an ensemble of four species distribution models (SDMs): boosted regression trees, maximum entropy, generalised additive model, and multivariate adaptive regression splines, together with 28 environmental covariates. We collected 1,422 and 1,169 occurrence records for cassava and Bemisia tabaci species complex from the Global Biodiversity Information Facility and 750 CBSD occurrence records from published literature and systematic surveys in East Africa. Our results identified isothermality as having the highest contribution to the current distribution of cassava, while elevation was the top predictor of the current distribution of Bemisia tabaci species complex. Cassava harvested area and precipitation of the driest month contributed the most to explain the current distribution of CBSD outbreaks. The geographic distributions of these target species are also expected to shift under climate projection scenarios for two mid-century periods (2041-2060 and 2061-2080). Our results indicate that major cassava producers, like Cameron, Ivory Coast, Ghana, and Nigeria, are at greater risk of invasion of CBSD. These results highlight the need for firmer agricultural management and climate-change mitigation actions in Africa to combat new outbreaks and to contain the spread of CBSD.

Mechanistic reconciliation of community and invasion ecology.

Community and invasion ecology have mostly grown independently. There is substantial overlap in the processes captured by different models in the two fields, and various frameworks have been developed to reduce this redundancy and synthesize information content. Despite broad recognition that community and invasion ecology are interconnected, a process-based framework synthesizing models across these two fields is lacking. Here we review 65 representative community and invasion models and propose a common framework articulated around six processes (dispersal, drift, abiotic interactions, within-guild interactions, cross-guild interactions, and genetic changes). The framework is designed to synthesize the content of the two fields, provide a general perspective on their development, and enable their comparison. The application of this framework and of a novel method based on network theory reveals some lack of coherence between the two fields, despite some historical similarities. Community ecology models are characterized by combinations of multiple processes, likely reflecting the search for an overarching theory to explain community assembly and structure, drawing predominantly on interaction processes, but also accounting largely for the other processes. In contrast, most models in invasion ecology invoke fewer processes and focus more on interactions between introduced species and their novel biotic and abiotic environment. The historical dominance of interaction processes and their independent developments in the two fields is also reflected in the lower level of coherence for models involving interactions, compared to models involving dispersal, drift, and genetic changes. It appears that community ecology, with a longer history than invasion ecology, has transitioned from the search for single explanations for patterns observed in nature to investigate how processes may interact mechanistically, thereby generating and testing hypotheses. Our framework paves the way for a similar transition in invasion ecology, to better capture the dynamics of multiple alien species introduced in complex communities. Reciprocally, applying insights from invasion to community ecology will help us understand and predict the future of ecological communities in the Anthropocene, in which human activities are weakening species' natural boundaries. Ultimately, the successful integration of the two fields could advance a predictive ecology that is urgently required in a rapidly changing world.

Prejudice, privilege, and power: Conflicts and cooperation between recognizable groups.

The problem of cooperation remains one of the fundamental questions in the fields of biology, sociology, and economics. The emergence and maintenance of cooperation are naturally affected by group dynamics, since individuals are likely to behave differently based on shared group membership. We here formulate a model of socio-economic power between two prejudiced groups, and explore the conditions for their cooperative coexistence under two social scenarios in a well-mixed environment. Each scenario corresponds to an asymmetrical increase in the payoffs for mutual cooperation in either cross-group or within-group interactions. In the 'inter-dependence' scenario payoffs of cross-group cooperation are enhanced, while in the 'group-cohesion' scenario payoffs of within-group cooperation are enhanced. We find that stable cooperative coexistence is possible only in the inter-dependence scenario. The conditions for such coexistence are highly sensitive to prejudice, defined as the reduction in probability for cross-group cooperation, and less sensitive to privilege, defined as the enhancements to payoffs of cross-group cooperation.

Modeling Top-Down and Bottom-Up Drivers of a Regime Shift in Invasive Aquatic Plant Stable States.

The evidence for alternate stable states characterized by dominance of either floating or submerged plant dominance is well established. Inspired by an existing model and controlled experiments, we conceptually describe a dynamic that we have observed in the field using a simple model, the aim of which was to investigate key interactions of the shift between invasive floating and invasive submerged plant dominance, driven by the rapid decomposition of floating plants as a consequence of herbivory by biological control agents. This study showed that the rate of switch between floating and submerged invasive plant dominance, and the point in time at which the switch occurs, is dependent on the nutrient status of the water and the density of biological control agents on floating plant populations. Therefore, top-down invasive plant biological control efforts using natural enemies can affect systems on a wider scale than the intended agent - plant level, and can be significantly altered by bottom-up changes to the system, i.e., nutrient loading. The implications of this are essential for understanding the multiple roles invasive plants and their control have upon ecosystem dynamics. The results emphasize the importance of multi-trophic considerations for future invasive plant management and offer evidence for new pathways of invasion. The model outputs support the conclusion that, after the shift and in the absence of effective intervention, a submerged invasive stable state will persist.

Variability in life-history switch points across and within populations explained by Adaptive Dynamics.

Understanding the factors that shape the timing of life-history switch points (SPs; e.g. hatching, metamorphosis and maturation) is a fundamental question in evolutionary ecology. Previous studies examining this question from a fitness optimization perspective have advanced our understanding of why the timing of life-history transitions may vary across populations and environments. However, in nature we also often observe variability among individuals within populations. Optimization theory, which typically predicts a single optimal SP under physiological and environmental constraints for a given environment, cannot explain this variability. Here, we re-examine the evolution of a single life-history SP between juvenile and adult stages from an Adaptive Dynamics (AD) perspective, which explicitly considers the feedback between the dynamics of population and the evolution of life-history strategy. The AD model, although simple in structure, exhibits a diverse range of evolutionary scenarios depending upon demographic and environmental conditions, including the loss of the juvenile stage, a single optimal SP, alternative optimal SPs depending on the initial phenotype, and sympatric coexistence of two SP phenotypes under disruptive selection. Such predictions are consistent with previous optimization approaches in predicting life-history SP variability across environments and between populations, and in addition they also explain within-population variability by sympatric disruptive selection. Thus, our model can be used as a theoretical tool for understanding life-history variability across environments and, especially, within species in the same environment.

Heterogeneity in local density allows a positive evolutionary relationship between self-fertilisation and dispersal.

Despite empirical evidence for a positive relationship between dispersal and self-fertilization (selfing), theoretical work predicts that these traits should always be negatively correlated, and the Good Coloniser Syndrome of high dispersal and selfing (Cf. Baker's Law) should not evolve. Critically, previous work assumes that adult density is spatiotemporally homogeneous, so selfing results in identical offspring production for all patches, eliminating the benefit of dispersal for escaping from local resource competition. We investigate the joint evolution of dispersal and selfing in a demographically structured metapopulation model where local density is spatiotemporally heterogeneous due to extinction-recolonization dynamics. Selfing alleviates outcrossing failure due to low local density (an Allee effect) while dispersal alleviates competition through dispersal of propagules from high- to low-density patches. Because local density is spatiotemporally heterogeneous in our model, selfing does not eliminate heterogeneity in competition, so dispersal remains beneficial even under full selfing. Hence the Good Coloniser Syndrome is evolutionarily stable under a broad range of conditions, and both negative and positive relationships between dispersal and selfing are possible, depending on the environment. Our model thus accommodates positive empirical relationships between dispersal and selfing not predicted by previous theoretical work and provides additional explanations for negative relationships.

Evolutionary fields can explain patterns of high-dimensional complexity in ecology.

One of the properties that make ecological systems so unique is the range of complex behavioral patterns that can be exhibited by even the simplest communities with only a few species. Much of this complexity is commonly attributed to stochastic factors that have very high-degrees of freedom. Orthodox study of the evolution of these simple networks has generally been limited in its ability to explain complexity, since it restricts evolutionary adaptation to an inertia-free process with few degrees of freedom in which only gradual, moderately complex behaviors are possible. We propose a model inspired by particle-mediated field phenomena in classical physics in combination with fundamental concepts in adaptation, which suggests that small but high-dimensional chaotic dynamics near to the adaptive trait optimum could help explain complex properties shared by most ecological datasets, such as aperiodicity and pink, fractal noise spectra. By examining a simple predator-prey model and appealing to real ecological data, we show that this type of complexity could be easily confused for or confounded by stochasticity, especially when spurred on or amplified by stochastic factors that share variational and spectral properties with the underlying dynamics.

The transition from evolutionary stability to branching: A catastrophic evolutionary shift.

Evolutionary branching-resident-mutant coexistence under disruptive selection-is one of the main contributions of Adaptive Dynamics (AD), the mathematical framework introduced by S.A.H. Geritz, J.A.J. Metz, and coauthors to model the long-term evolution of coevolving multi-species communities. It has been shown to be the basic mechanism for sympatric and parapatric speciation, despite the essential asexual nature of AD. After 20 years from its introduction, we unfold the transition from evolutionary stability (ESS) to branching, along with gradual change in environmental, control, or exploitation parameters. The transition is a catastrophic evolutionary shift, the branching dynamics driving the system to a nonlocal evolutionary attractor that is viable before the transition, but unreachable from the ESS. Weak evolutionary stability hence qualifies as an early-warning signal for branching and a testable measure of the community's resilience against biodiversity. We clarify a controversial theoretical question about the smoothness of the mutant invasion fitness at incipient branching. While a supposed nonsmoothness at third order long prevented the analysis of the ESS-branching transition, we argue that smoothness is generally expected and derive a local canonical model in terms of the geometry of the invasion fitness before branching. Any generic AD model undergoing the transition qualitatively behaves like our canonical model.

Community analysis in directed networks: in-, out-, and pseudocommunities.

When analyzing important classes of complex interconnected systems, link directionality can hardly be neglected if a precise and effective picture of the structure and function of the system is needed. If community analysis is performed, the notion of "community" itself is called into question, since the property of having a comparatively looser external connectivity could refer to the inbound or outbound links only or to both categories. In this paper, we introduce the notions of in-, out-, and in-/out-community in order to correctly classify the directedness of the interaction of a subnetwork with the rest of the system. Furthermore, we extend the scope of community analysis by introducing the notions of in-, out-, and in-/out-pseudocommunity. They are subnetworks having strong internal connectivity but also important interactions with the rest of the system, the latter taking place by means of a minority of its nodes only. The various types of (pseudo-)communities are qualified and distinguished by a suitable set of indicators and, on a given network, they can be discovered by using a "local" searching algorithm. The application to a broad set of benchmark networks and real-world examples proves that the proposed approach is able to effectively disclose the different types of structures above defined and to usefully classify the directionality of their interactions with the rest of the system.