A roadmap towards predicting species interaction networks (across space and time).

Networks of species interactions underpin numerous ecosystem processes, but comprehensively sampling these interactions is difficult. Interactions intrinsically vary across space and time, and given the number of species that compose ecological communities, it can be tough to distinguish between a true negative (where two species never interact) from a false negative (where two species have not been observed interacting even though they actually do). Assessing the likelihood of interactions between species is an imperative for several fields of ecology. This means that to predict interactions between species-and to describe the structure, variation, and change of the ecological networks they form-we need to rely on modelling tools. Here, we provide a proof-of-concept, where we show how a simple neural network model makes accurate predictions about species interactions given limited data. We then assess the challenges and opportunities associated with improving interaction predictions, and provide a conceptual roadmap forward towards predictive models of ecological networks that is explicitly spatial and temporal. We conclude with a brief primer on the relevant methods and tools needed to start building these models, which we hope will guide this research programme forward. This article is part of the theme issue 'Infectious disease macroecology: parasite diversity and dynamics across the globe'.

Ecological interactions and the Netflix problem.

Species interactions are a key component of ecosystems but we generally have an incomplete picture of who-eats-who in a given community. Different techniques have been devised to predict species interactions using theoretical models or abundances. Here, we explore the K nearest neighbour approach, with a special emphasis on recommendation, along with a supervised machine learning technique. Recommenders are algorithms developed for companies like Netflix to predict whether a customer will like a product given the preferences of similar customers. These machine learning techniques are well-suited to study binary ecological interactions since they focus on positive-only data. By removing a prey from a predator, we find that recommenders can guess the missing prey around 50% of the times on the first try, with up to 881 possibilities. Traits do not improve significantly the results for the K nearest neighbour, although a simple test with a supervised learning approach (random forests) show we can predict interactions with high accuracy using only three traits per species. This result shows that binary interactions can be predicted without regard to the ecological community given only three variables: body mass and two variables for the species' phylogeny. These techniques are complementary, as recommenders can predict interactions in the absence of traits, using only information about other species' interactions, while supervised learning algorithms such as random forests base their predictions on traits only but do not exploit other species' interactions. Further work should focus on developing custom similarity measures specialized for ecology to improve the KNN algorithms and using richer data to capture indirect relationships between species.

Evolution of a transposon in Daphnia hybrid genomes.

BACKGROUND: Transposable elements play a major role in genome evolution. Their capacity to move and/or multiply in the genome of their host may have profound impacts on phenotypes, and may have dramatic consequences on genome structure. Hybrid and polyploid clones have arisen multiple times in the Daphnia pulex complex and are thought to reproduce by obligate parthenogenesis. Our study examines the evolution of a DNA transposable element named Pokey in the D. pulex complex. RESULTS: Portions of Pokey elements inserted in the 28S rRNA genes from various Daphnia hybrids (diploids and polyploids) were sequenced and compared to sequences from a previous study to understand the evolutionary history of the elements. Pokey sequences show a complex phylogenetic pattern. We found evidence of recombination events in numerous Pokey alleles from diploid and polyploid hybrids and also from non-hybrid diploids. The recombination rate in Pokey elements is comparable to recombination rates previously estimated for 28S rRNA genes in the congener, Daphnia obtusa. Some recombinant Pokey alleles were encountered in Daphnia isolates from multiple locations and habitats. CONCLUSIONS: Phylogenetic and recombination analyses showed that recombination is a major force that shapes Pokey evolution. Based on Pokey phylogenies, reticulation has played and still plays an important role in shaping the diversity of the D. pulex complex. Horizontal transfer of Pokey seems to be rare and hybrids often possess Pokey elements derived from recombination among alleles encountered in the putative parental species. The insertion of Pokey in hotspots of recombination may have important impacts on the diversity and fitness of this transposable element.

Immigration, local dispersal limitation, and the repeatability of community composition under neutral and niche dynamics.

Repeatability of community composition has been a critical aspect for community structure, which is closely associated with community stability, predictability, conservation biology and ecological restoration. It has been shown that both immigration and local dispersal limitation can affect the community composition in both neutral and niche model. Hence, we use a spatially explicit individual-based model to investigate the potential influence of immigration rate and strength of local dispersal limitation on repeatability in both neutral and niche models. Similarity measures are used to quantify repeatability. We examine the repeatability of community composition among replicate communities (which means the same community repeats many times), and between niche and neutral replicate communities. We find the correlation between repeatability and immigration rate is positive in the neutral model and an inverted unimodal in the niche model. The correlation between repeatability and local dispersal distance is positive in the niche model and negative in the neutral model. High repeatability between niche communities and neutral communities is observed with high immigration rates or when high local dispersal distance appears in the niche model or low local dispersal distance in the neutral model. Our results show that repeatability of community composition is not only dependent on the types of community models (niche vs. neutrality) but also strongly determined by immigration rates and local dispersal limitation.

A complex speciation-richness relationship in a simple neutral model.

Speciation is the "elephant in the room" of community ecology. As the ultimate source of biodiversity, its integration in ecology's theoretical corpus is necessary to understand community assembly. Yet, speciation is often completely ignored or stripped of its spatial dimension. Recent approaches based on network theory have allowed ecologists to effectively model complex landscapes. In this study, we use this framework to model allopatric and parapatric speciation in networks of communities. We focus on the relationship between speciation, richness, and the spatial structure of communities. We find a strong opposition between speciation and local richness, with speciation being more common in isolated communities and local richness being higher in more connected communities. Unlike previous models, we also find a transition to a positive relationship between speciation and local richness when dispersal is low and the number of communities is small. We use several measures of centrality to characterize the effect of network structure on diversity. The degree, the simplest measure of centrality, is the best predictor of local richness and speciation, although it loses some of its predictive power as connectivity grows. Our framework shows how a simple neutral model can be combined with network theory to reveal complex relationships between speciation, richness, and the spatial organization of populations.

How likely is speciation in neutral ecology?

Patterns of biodiversity predicted by the neutral theory rely on a simple phenomenological model of speciation. To further investigate the effect of speciation on neutral biodiversity, we analyze a spatially explicit neutral model based on population genetics. We define the metacommunity as a system of populations exchanging migrants, and we use this framework to introduce speciation with little or no gene flow (allopatric and parapatric speciation). We find that with realistic mutation rates, our metacommunity model driven by neutral processes cannot support more than a few species. Adding natural selection in the population genetics of speciation increases the number of species in the metacommunity, but the level of diversity found in the Barro Colorado Island is difficult to reach.