Quantifying positional importance in food webs: A comparison of centrality indices

Abundant top predators increase species interaction network complexity in northeastern Chinese forests

Species interactions remain a cornerstone in shaping community dynamics and structure, alongside other factors, such as climate conditions and human activities. Although network structure is known to influence community stability and ecosystem functioning, the roles of top predators in shaping interaction network structure remain obscure. We examined a 5-7-year time series of species detections for mammal communities in multiple protected areas to investigate the association between top predators and interaction network structure. Our findings suggest that abundant species, day-active species and species with wide habitat breadth interact with more species, as did species that were more affected by vehicle disturbance. With increased densities of top predators, interaction networks exhibited greater complexity, with increased connectance, nestedness and average degree. An increased density of mesopredators, such as yellow-throated martens and badgers, was associated with sparser, less nested, but more centralized interaction networks. Top predators reduced the degree of highly interactive species, making them more specialized, and increased the degree of less abundant species, making them more general. In particular, this redistribution of interactions was not driven by direct changes in species density of top predators but seemingly by non-consumptive or indirect effects. Our findings emphasize the pivotal role of the main predators in structuring interactions within northeastern China's mammal communities, with large implications for conservation and management.

The network architecture and phylogeographic drivers of interactions between rodents and seed plants at continental scales

Rodents are known to interact with seed plants in three different ways, including predation in situ, scatter hoarding and larder hoarding of seeds. These behaviours span a spectrum from mutualistic seed dispersal to predation, and they are related to species' and environmental characteristics. We used interaction networks to evaluate the structure and drivers of rodent-seed plant interactions, including geography, phylogeny and traits at continental scales. We constructed five aggregated networks, each representing a continent and containing three subnetworks defined by foraging behaviours, tested questions about their network structures and analysed the driving signals shaping rodent-seed plant interactions at network and species levels. Rodent-seed plant networks varied across continents. We found most rodents exhibited a significant propensity for one foraging behaviour and detected significant modular structures in both aggregated networks and subnetworks. We detected significant co-phylogenetic signals between rodents and seed plants. Distance matrix-based regressions on interaction and module dissimilarity of rodents suggest geographical and phylogenetic forces are important in the assembly of rodent-seed plant networks. In addition, multiple species traits correlated with the roles of rodents within aggregated networks; however, the specific traits associated with these roles varied among interaction types. Our results highlight that geography and phylogenetics are dominant in structuring the architecture of rodent-seed plant networks at continental scales and reveal challenges regarding spatial and taxa coverage in rodent-seed plant interactions.

Connectome embedding in multidimensional graph spaces

Connectomes' topological organization can be quantified using graph theory. Here, we investigated brain networks in higher dimensional spaces defined by up to 10 graph theoretic nodal properties. These properties assign a score to nodes, reflecting their meaning in the network. Using 100 healthy unrelated subjects from the Human Connectome Project, we generated various connectomes (structural/functional, binary/weighted). We observed that nodal properties are correlated (i.e., they carry similar information) at whole-brain and subnetwork level. We conducted an exploratory machine learning analysis to test whether high-dimensional network information differs between sensory and association areas. Brain regions of sensory and association networks were classified with an 80-86% accuracy in a 10-dimensional (10D) space. We observed the largest gain in machine learning accuracy going from a 2D to 3D space, with a plateauing accuracy toward 10D space, and nonlinear Gaussian kernels outperformed linear kernels. Finally, we quantified the Euclidean distance between nodes in a 10D graph space. The multidimensional Euclidean distance was highest across subjects in the default mode network (in structural networks) and frontoparietal and temporal lobe areas (in functional networks). To conclude, we propose a new framework for quantifying network features in high-dimensional spaces that may reveal new network properties of the brain.

Mycobiome analysis of leaf, root, and soil of symptomatic oil palm trees (Elaeis guineensis Jacq.) affected by leaf spot disease

Recently, attention has been shifting toward the perspective of the existence of plants and microbes as a functioning ecological unit. However, studies highlighting the impacts of the microbial community on plant health are still limited. In this study, fungal community (mycobiome) of leaf, root, and soil of symptomatic leaf-spot diseased (SS) oil palm were compared against asymptomatic (AS) trees using ITS2 rRNA gene metabarcoding. A total of 3,435,417 high-quality sequences were obtained from 29 samples investigated. Out of the 14 phyla identified, Ascomycota and Basidiomycota were the most dominant accounting for 94.2 and 4.7% of the total counts in AS, and 75 and 21.2% in SS, respectively. Neopestalotiopsis is the most abundant genus for AS representing 8.0% of the identified amplicons compared to 2.0% in SS while Peniophora is the most abundant with 8.6% of the identified amplicons for SS compared to 0.1% in AS. The biomarker discovery algorithm LEfSe revealed different taxa signatures for the sample categories, particularly soil samples from asymptomatic trees, which were the most enriched. Network analysis revealed high modularity across all groups, except in root samples. Additionally, a large proportion of the identified keystone species consisted of rare taxa, suggesting potential role in ecosystem functions. Surprisingly both AS and SS leaf samples shared taxa previously associated with oil palm leaf spot disease. The significant abundance of Trichoderma asperellum in the asymptomatic root samples could be further explored as a potential biocontrol agent against oil palm disease.

Approaching the Ecological Role of the Squat Lobster (Munida gregaria) and the Fuegian Sprat (Sprattus fuegensis) in the Francisco Coloane Marine Area (Magellan Strait, Chile) Using a Pelagic Food Web Model

The structure and functioning of the food web of the Francisco Coloane Marine Area in the Magellan Strait, Chile, was quantified, with an emphasis on identifying the ecological role of the squat lobster (Munida gregaria) and the Fuegian sprat (Sprattus fuegensis). Food web indicators, the trophic level, and centrality indices were estimated using Ecopath with Ecosim. Dynamic simulations were carried out to evaluate the ecosystem impacts of biomass changes in squat lobster and Fuegian sprat. The model calculated a total ecosystem biomass of 71.7 t km^-2 and a total primary production of 2450.9 t km^-2 year^-1. Squat lobster and Fuegian sprat were located in specific trophic levels of 2.3 and 2.7, respectively. Squat lobster reduction produced a decrease in the biomass of red cod (42-56%) and humpback whales (25-28%) and Fuegian sprat reduction a decrease in penguins (15-37%) and seabirds (11-34%). The Francisco Coloane Area is an immature ecosystem with productivity and energy flows values within those reported for productive ecosystems; the role of the squat lobster seems to be related to the structure of the food web, and the role of the Fuegian sprat seems to be related to the functioning of the ecosystem and to the energy transfer to top predators.

A network simplification approach to ease topological studies about the food-web architecture

Food webs studies are intrinsically complex and time-consuming. Network data about trophic interaction across different large locations and ecosystems are scarce in comparison with general ecological data, especially if we consider terrestrial habitats. Here we present a complex network strategy to ease the gathering of the information by simplifying the collection of data with a taxonomic key. We test how well the topology of three different food webs retain their structure at the resolution of the nodes across distinct levels of simplification, and we estimate how community detection could be impacted by this strategy. The first level of simplification retains most of the general topological indices; betweenness and trophic levels seem to be consistent and robust even at the higher levels of simplification. This result suggests that generalisation and standardisation, as a good practice in food webs science, could benefit the community, both increasing the amount of open data available and the comparison among them, thus providing support especially for scientists that are new in this field and for exploratory analysis.

Seasonal dynamics of plant pollinator networks in agricultural landscapes: how important is connector species identity in the network?

Farm habitat enrichment is crucial for sustainable production of pollinator-dependent crops. Correct choice of crop and non-crop plant species in habitat management support resilient pollinator fauna and effective pollination service delivery. We identify key network metrices to recognize suitable crop and non-crop plant species for farm habitat enrichment. We also highlight the importance of seasonal variation of the key plant and pollinator species that will crucially inform farm management. Crop species played a key role in maintaining plant-pollinator network integrity. In contrast to the conventional practice of focussing on non-crop plants for pollination service restoration, we find crop plants across seasons hold a key role in maintaining healthy plant-pollinator networks. Our study highlights the importance of non-bee pollinators especially, flies and butterflies in sustaining healthy plant-pollinator network. Bees were important as connector species and controlled other species in the network. Only 16.67% bees and 33.33% of plant species acted as connector species. Our study also shows that the identity of connector species in a plant-pollinator network can change drastically across seasons.

The Structure of Ecological Networks Across Levels of Organization

Interactions connect the units of ecological systems, forming networks. Individual-based networks characterize variation in niches among individuals within populations. These individual-based networks merge with each other, forming species-based networks and food webs that describe the architecture of ecological communities. Networks at broader spatiotemporal scales portray the structure of ecological interactions across landscapes and over macroevolutionary time. Here, I review the patterns observed in ecological networks across multiple levels of biological organization. A fundamental challenge is to understand the amount of interdependence as we move from individual-based networks to species-based networks and beyond. Despite the uneven distribution of studies, regularities in network structure emerge across scales due to the fundamental architectural patterns shared by complex networks and the interplay between traits and numerical effects. I illustrate the integration of these organizational scales by exploring the consequences of the emergence of highly connected species for network structures across scales.

Deciphering the expression dynamics of ANGPTL8 associated regulatory network in insulin resistance using formal modelling approaches

ANGPTL8 is a recently identified novel hormone which regulates both glucose and lipid metabolism. The increase in ANGPTL8 during compensatory insulin resistance has been recently reported to improve glucose tolerance and a part of cytoprotective metabolic circuit. However, the exact signalling entities and dynamics involved in this process have remained elusive. Therefore, the current study was conducted with a specific aim to model the regulation of ANGPTL8 with emphasis on its role in improving glucose tolerance during insulin resistance. The main contribution of this study is the construction of a discrete model (based on kinetic logic of René Thomas) and its equivalent Stochastic Petri Net model of ANGPTL8 associated Biological Regulatory Network (BRN) which can predict its dynamic behaviours. The predicted results of these models are in‐line with the previous experimental observations and provide comprehensive insights into the signalling dynamics of ANGPTL8 associated BRN. The authors’ results support the hypothesis that ANGPTL8 plays an important role in supplementing the insulin signalling pathway during insulin resistance and its loss can aggravate the pathogenic process by quickly leading towards Diabetes Mellitus. The results of this study have potential therapeutic implications for treatment of Diabetes Mellitus and are suggestive of its potential as a glucose‐lowering agent.

Host taxonomy constrains the properties of trophic transmission routes for parasites in lake food webs

Some parasites move from one host to another via trophic transmission, the consumption of the parasite (inside its current host) by its future host. Feeding links among free-living species can thus be understood as potential transmission routes for parasites. As these links have different dynamic and structural properties, they may also vary in their effectiveness as trophic transmission routes. That is, some links may be better than others in allowing parasites to complete their complex life cycles. However, not all links are accessible to parasites as most are restricted to a small number of host taxa. This restriction means that differences between links involving host and non-host taxa must be considered when assessing whether transmission routes for parasites have different food web properties than other links. Here we use four New Zealand lake food webs to test whether link properties (contribution of a link to the predator's diet, prey abundance, prey biomass, amount of biomass transferred, centrality, and asymmetry) affect trophic transmission of parasites. Critically, we do this using both models that neglect the taxonomy of free-living species and models that explicitly include information about which free-living species are members of suitable host taxa. Although the best-fit model excluding taxonomic information suggested that transmission routes have different properties than other feeding links, when including taxonomy, the best-fit model included only an intercept. This means that the taxonomy of free-living species is a key determinant of parasite transmission routes and that food-web properties of transmission routes are constrained by the properties of host taxa. In particular, many intermediate hosts (prey) attain high biomasses and are involved in highly central links while links connecting intermediate to definitive (predator) hosts tend to be dynamically weak.

The structural importance of less abundant species in Prince William Sound food web

Rarity of species is often considered to set priorities for biodiversity conservation. Less abundant species are expected to be at higher risk of extinction and make significant contribution to food web functioning. However, the relationship between species abundance and position in food webs is still unclear. Here we tested possible correlations between species abundance and structural position in Prince William Sound food web. Species abundance was inferred from biomass data and structural position was characterized by 13 centrality indices.

We found that less abundant species have higher trophic positions and display more generalist feeding strategies. However, positive correlations link most of the centrality indices to population size. Thus, being locally rare translates into more peripheral food web positions and implies marginal roles in the spread of indirect effects. Species characterized by largest population size are responsible for the transfer of largest amounts of biomass and regulate the transmission of indirect effects. Less abundant species are of marginal structural importance and are exposed to impacts mediated by larger populations. In Prince William Sound ecosystem, rarity is associated with critical food web positions and does not simply reflect a marginal contribution to biodiversity.

We suggest that knowing the food web position of rare species might help to formulate more effective, system-level solutions for their conservation, rather than simply focusing on the direct treatment of symptoms.

Ranking species in mutualistic networks

Understanding the architectural subtleties of ecological networks, believed to confer them enhanced stability and robustness, is a subject of outmost relevance. Mutualistic interactions have been profusely studied and their corresponding bipartite networks, such as plant-pollinator networks, have been reported to exhibit a characteristic "nested" structure. Assessing the importance of any given species in mutualistic networks is a key task when evaluating extinction risks and possible cascade effects. Inspired in a recently introduced algorithm--similar in spirit to Google's PageRank but with a built-in non-linearity--here we propose a method which--by exploiting their nested architecture--allows us to derive a sound ranking of species importance in mutualistic networks. This method clearly outperforms other existing ranking schemes and can become very useful for ecosystem management and biodiversity preservation, where decisions on what aspects of ecosystems to explicitly protect need to be made.

Supporting crop pollinators with floral resources: network-based phenological matching

The production of diverse and affordable agricultural crop species depends on pollination services provided by bees. Indeed, the proportion of pollinator-dependent crops is increasing globally. Agriculture relies heavily on the domesticated honeybee; the services provided by this single species are under threat and becoming increasingly costly. Importantly, the free pollination services provided by diverse wild bee communities have been shown to be sufficient for high agricultural yields in some systems. However, stable, functional wild bee communities require floral resources, such as pollen and nectar, throughout their active season, not just when crop species are in flower. To target floral provisioning efforts to conserve and support native and managed bee species, we apply network theoretical methods incorporating plant and pollinator phenologies. Using a two-year dataset comprising interactions between bees (superfamily Apoidea, Anthophila) and 25 native perennial plant species in floral provisioning habitat, we identify plant and bee species that provide a key and central role to the stability of the structure of this community. We also examine three specific case studies: how provisioning habitat can provide temporally continuous support for honeybees (Apis mellifera) and bumblebees (Bombus impatiens), and how resource supplementation strategies might be designed for a single genus of important orchard pollinators (Osmia). This framework could be used to provide native bee communities with additional, well-targeted floral resources to ensure that they not only survive, but also thrive.

A bio-inspired methodology of identifying influential nodes in complex networks

How to identify influential nodes is a key issue in complex networks. The degree centrality is simple, but is incapable to reflect the global characteristics of networks. Betweenness centrality and closeness centrality do not consider the location of nodes in the networks, and semi-local centrality, leaderRank and pageRank approaches can be only applied in unweighted networks. In this paper, a bio-inspired centrality measure model is proposed, which combines the Physarum centrality with the K-shell index obtained by K-shell decomposition analysis, to identify influential nodes in weighted networks. Then, we use the Susceptible-Infected (SI) model to evaluate the performance. Examples and applications are given to demonstrate the adaptivity and efficiency of the proposed method. In addition, the results are compared with existing methods.

Robustness of empirical food webs with varying consumer's sensitivities to loss of resources

Food web responses to species loss have been mostly studied in binary food webs, thus without accounting for the amount of energy transferred in consumer-resource interactions. We introduce an energetic criterion, called extinction threshold, for which a species goes secondarily extinct when a certain fraction of its incoming energy is lost. We study the robustness to random node loss of 10 food webs based on empirically-derived weightings. We use different extinction scenarios (random removal and from most- to least-connected species), and we simulate 10(5) replicates for each extinction threshold to account for stochasticity of extinction dynamics. We quantified robustness on the basis of how many additional species (i.e. secondary extinctions) were lost after the direct removal of species (i.e. primary extinctions). For all food webs, the expected robustness linearly decreases with extinction threshold, although a large variance in robustness is observed. The sensitivity of robustness to variations in extinction threshold increases with food web species richness and quantitative unweighted link density, while we observed a nonlinear relationship when the predictor is food web connectance and no relationship with the proportion of autotrophs.

Environ centrality reveals the tendency of indirect effects to homogenize the functional importance of species in ecosystems

Ecologists and conservation biologists need to identify the relative importance of species to make sound management decisions and effectively allocate scarce resources. We introduce a new method, termed environ centrality, to determine the relative importance of a species in an ecosystem network with respect to ecosystem energy-matter exchange. We demonstrate the uniqueness of environ centrality by comparing it to other common centrality metrics and then show its ecological significance. Specifically, we tested two hypotheses on a set of 50 empirically based ecosystem network models. The first concerned the distribution of centrality in the community. We hypothesized that the functional importance of species would tend to be concentrated into a few dominant species followed by a group of species with lower, more even importance as is often seen in dominance-diversity curves. Second, we tested the systems ecology hypothesis that indirect relationships homogenize the functional importance of species in ecosystems. Our results support both hypotheses and highlight the importance of detritus and nutrient recyclers such as fungi and bacteria in generating the energy-matter flow in ecosystems. Our homogenization results suggest that indirect effects are important in part because they tend to even the importance of species in ecosystems. A core contribution of this work is that it creates a formal, mathematical method to quantify the importance species play in generating ecosystem activity by integrating direct, indirect, and boundary effects in ecological systems.

Combining a dispersal model with network theory to assess habitat connectivity

Assessing the potential for threatened species to persist and spread within fragmented landscapes requires the identification of core areas that can sustain resident populations and dispersal corridors that can link these core areas with isolated patches of remnant habitat. We developed a set of GIS tools, simulation methods, and network analysis procedures to assess potential landscape connectivity for the Delmarva fox squirrel (DFS; Sciurus niger cinereus), an endangered species inhabiting forested areas on the Delmarva Peninsula, USA. Information on the DFS's life history and dispersal characteristics, together with data on the composition and configuration of land cover on the peninsula, were used as input data for an individual-based model to simulate dispersal patterns of millions of squirrels. Simulation results were then assessed using methods from graph theory, which quantifies habitat attributes associated with local and global connectivity. Several bottlenecks to dispersal were identified that were not apparent from simple distance-based metrics, highlighting specific locations for landscape conservation, restoration, and/or squirrel translocations. Our approach links simulation models, network analysis, and available field data in an efficient and general manner, making these methods useful and appropriate for assessing the movement dynamics of threatened species within landscapes being altered by human and natural disturbances.

Generalized walks-based centrality measures for complex biological networks

A strategy for zooming in and out the topological environment of a node in a complex network is developed. This approach is applied here to generalize the subgraph centrality of nodes in complex networks. In this case the zooming in strategy is based on the use of some known matrix functions which allow focusing locally on the environment of a node. When a zooming out strategy is applied new matrix functions are introduced, which give a more global picture of the topological surrounds of a node. These indices permit a modulation of the scales at which the environment of a node influences its centrality. We apply them to the study of 10 protein-protein interaction (PPI) networks. We illustrate the similarities and differences between the generalized subgraph centrality indices as well as among them and some classical centrality measures. We show here that the use of centrality indices based on the zooming in strategy identifies a larger number of essential proteins in the yeast PPI network than any of the other centrality measures studied.

Googling food webs: can an eigenvector measure species' importance for coextinctions?

A major challenge in ecology is forecasting the effects of species' extinctions, a pressing problem given current human impacts on the planet. Consequences of species losses such as secondary extinctions are difficult to forecast because species are not isolated, but interact instead in a complex network of ecological relationships. Because of their mutual dependence, the loss of a single species can cascade in multiple coextinctions. Here we show that an algorithm adapted from the one Google uses to rank web-pages can order species according to their importance for coextinctions, providing the sequence of losses that results in the fastest collapse of the network. Moreover, we use the algorithm to bridge the gap between qualitative (who eats whom) and quantitative (at what rate) descriptions of food webs. We show that our simple algorithm finds the best possible solution for the problem of assigning importance from the perspective of secondary extinctions in all analyzed networks. Our approach relies on network structure, but applies regardless of the specific dynamical model of species' interactions, because it identifies the subset of coextinctions common to all possible models, those that will happen with certainty given the complete loss of prey of a given predator. Results show that previous measures of importance based on the concept of "hubs" or number of connections, as well as centrality measures, do not identify the most effective extinction sequence. The proposed algorithm provides a basis for further developments in the analysis of extinction risk in ecosystems.

Keystone species and food webs

Different species are of different importance in maintaining ecosystem functions in natural communities. Quantitative approaches are needed to identify unusually important or influential, 'keystone' species particularly for conservation purposes. Since the importance of some species may largely be the consequence of their rich interaction structure, one possible quantitative approach to identify the most influential species is to study their position in the network of interspecific interactions. In this paper, I discuss the role of network analysis (and centrality indices in particular) in this process and present a new and simple approach to characterizing the interaction structures of each species in a complex network. Understanding the linkage between structure and dynamics is a condition to test the results of topological studies, I briefly overview our current knowledge on this issue. The study of key nodes in networks has become an increasingly general interest in several disciplines: I will discuss some parallels. Finally, I will argue that conservation biology needs to devote more attention to identify and conserve keystone species and relatively less attention to rarity.

Using network centrality measures to manage landscape connectivity

We use a graph-theoretical landscape modeling approach to investigate how to identify central patches in the landscape as well as how these central patches influence (1) organism movement within the local neighborhood and (2) the dispersal of organisms beyond the local neighborhood. Organism movements were theoretically estimated based on the spatial configuration of the habitat patches in the studied landscape. We find that centrality depends on the way the graph-theoretical model of habitat patches is constructed, although even the simplest network representation, not taking strength and directionality of potential organisms flows into account, still provides a coarse-grained assessment of the most important patches according to their contribution to landscape connectivity. Moreover, we identify (at least) two general classes of centrality. One accounts for the local flow of organisms in the neighborhood of a patch, and the other accounts for the ability to maintain connectivity beyond the scale of the local neighborhood. Finally, we study how habitat patches with high scores on different network centrality measures are distributed in a fragmented agricultural landscape in Madagascar. Results show that patches with high degree and betweenness centrality are widely spread, while patches with high subgraph and closeness centrality are clumped together in dense clusters. This finding may enable multispecies analyses of single-species network models.