Shai Pilosof
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Shai Pilosof has a total of 170 co-authors. Here are the top co-authors by number of shared publications:
- Mercedes Pascual
Department of Biology and Department of Environmental Studies, New York University
8 shared publications Boris R. Krasnov
Mitrani Department of Desert Ecology, Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus, 84990, Israel
6 shared publications- Boris R. Krasnov
Mitrani Department of Desert Ecology, Swiss Institute for Dryland Environmental and Energy Research, Jacob Blaustein Institutes for Desert Research Ben‐Gurion University of the Negev Midreshet Ben‐Gurion Israel
6 shared publications - Matan Markfeld
Department of Life Sciences Ben‐Gurion University of the Negev Beer‐Sheva Israel
6 shared publications - Qixin He
Department of Biological Sciences, Purdue University, West Lafayette, IN, USA
6 shared publications - Rami Puzis
Ben-Gurion University of the Negev
6 shared publications
Latest Publications
- Open Access
- 25 September 2026
- Published by California Digital Library (CDL) in EcoEvoRxiv
Disentangling the drivers of parasite occurrence in a native and a non-native small mammal across a land-use gradient
Land-use change can reorganize host--parasite systems by altering environmental conditions, host communities, and transmission pathways. Yet parasite occurrence remains difficult to predict as it is shaped simultaneously by the environment, host identity, individual host attributes, the within-host microbiome, and parasite biology. We integrated these components within a single framework for two coexisting small mammals along a land-use gradient in northeastern Madagascar: the native, Malagasy-endemic shrew tenrec, Microgale brevicaudata (Tenrecidae), and the non-native, invasive black rat, Rattus rattus (Muridae). We characterized gut protozoan and nematode parasites in 505 hosts and used Hierarchical Modeling of Species Communities (HMSC) to quantify the relative contributions of environmental, host, microbiome, and parasite-level factors to parasite occurrence. Host species was the dominant driver overall, accounting for approximately half of the weighted explained variance, followed by environmental conditions, but their relative importance varied substantially among parasites. Parasite taxonomic group explained part of this variation, indicating that parasite biology also shaped responses to the same environmental and host-level drivers. Land-use responses differed sharply between hosts: parasite richness in M. brevicaudata was highest in the primary forest, whereas richness in R. rattus was highest in disturbed habitats. Thus, land-use change did not generate a common increase or decline in parasitism, but redistributed parasite richness between coexisting native and non-native hosts. Together, these results show that responses of host--parasite systems to land-use change emerge from the interaction of host identity, environmental conditions, and parasite biology, and cannot be understood from any single level alone. This multi-level perspective can improve predictions of infection risk as global change expands the interface between native and non-native hosts.
- 25 September 2026
- Published by California Digital Library (CDL) in EcoEvoRxiv
Open AccessDisentangling the drivers of parasite occurrence in a native and a non-native small mammal across a land-use gradient
Land-use change can reorganize host--parasite systems by altering environmental conditions, host communities, and transmission pathways. Yet parasite occurrence remains difficult to predict as it is shaped simultaneously by the environment, host identity, individual host attributes, the within-host microbiome, and parasite biology. We integrated these components within a single framework for two coexisting small mammals along a land-use gradient in northeastern Madagascar: the native, Malagasy-endemic shrew tenrec, Microgale brevicaudata (Tenrecidae), and the non-native, invasive black rat, Rattus rattus (Muridae). We characterized gut protozoan and nematode parasites in 505 hosts and used Hierarchical Modeling of Species Communities (HMSC) to quantify the relative contributions of environmental, host, microbiome, and parasite-level factors to parasite occurrence. Host species was the dominant driver overall, accounting for approximately half of the weighted explained variance, followed by environmental conditions, but their relative importance varied substantially among parasites. Parasite taxonomic group explained part of this variation, indicating that parasite biology also shaped responses to the same environmental and host-level drivers. Land-use responses differed sharply between hosts: parasite richness in M. brevicaudata was highest in the primary forest, whereas richness in R. rattus was highest in disturbed habitats. Thus, land-use change did not generate a common increase or decline in parasitism, but redistributed parasite richness between coexisting native and non-native hosts. Together, these results show that responses of host--parasite systems to land-use change emerge from the interaction of host identity, environmental conditions, and parasite biology, and cannot be understood from any single level alone. This multi-level perspective can improve predictions of infection risk as global change expands the interface between native and non-native hosts.
- Open Access
- 16 September 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Metabolite correlation networks reveal complex phenotypes of adaptation-driving mutations
Metabolic adaptation is often driven by mutations in pleiotropic genes that affect metabolism globally. However, identifying which phenotypes produced by such mutations drive selection remains difficult. Here, we consider metabolite correlation networks as a systems-level phenotype whose emergent properties reflect the combined effects of biochemical reactions and regulation. We hypothesize that mutation-induced changes in the covariance structure of these networks constitute complex phenotypes linking genetic change to bacterial fitness, and that node-level shifts can identify metabolites and pathways mechanistically associated with adaptation. We test this hypothesis by applying adaptive laboratory evolution to a metabolically suboptimal Escherichia coli strain in which metK, encoding methionine adenosyltransferase (MAT), is replaced by an ortholog from Ureaplasma urealyticum. Untargeted metabolite correlation networks reveal that evolution shifts S-adenosylmethionine (SAM), the MAT product, from a peripheral node to a key connector between network clusters. None of the accumulated pleiotropic mutations causing the shift directly affect MAT activity or SAM metabolism, indicating that this transition emerges from system-level metabolic reorganization. Targeted metabolomics of nodes showing similar shifts identifies additional metabolites involved in SAM-related pathways. These findings establish metabolite correlation networks as a framework for mapping genotype–phenotype–fitness relationships and prioritizing metabolic nodes linked to adaptive mechanisms.
- 16 September 2026
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Open AccessMetabolite correlation networks reveal complex phenotypes of adaptation-driving mutations
Metabolic adaptation is often driven by mutations in pleiotropic genes that affect metabolism globally. However, identifying which phenotypes produced by such mutations drive selection remains difficult. Here, we consider metabolite correlation networks as a systems-level phenotype whose emergent properties reflect the combined effects of biochemical reactions and regulation. We hypothesize that mutation-induced changes in the covariance structure of these networks constitute complex phenotypes linking genetic change to bacterial fitness, and that node-level shifts can identify metabolites and pathways mechanistically associated with adaptation. We test this hypothesis by applying adaptive laboratory evolution to a metabolically suboptimal Escherichia coli strain in which metK, encoding methionine adenosyltransferase (MAT), is replaced by an ortholog from Ureaplasma urealyticum. Untargeted metabolite correlation networks reveal that evolution shifts S-adenosylmethionine (SAM), the MAT product, from a peripheral node to a key connector between network clusters. None of the accumulated pleiotropic mutations causing the shift directly affect MAT activity or SAM metabolism, indicating that this transition emerges from system-level metabolic reorganization. Targeted metabolomics of nodes showing similar shifts identifies additional metabolites involved in SAM-related pathways. These findings establish metabolite correlation networks as a framework for mapping genotype–phenotype–fitness relationships and prioritizing metabolic nodes linked to adaptive mechanisms.
- Open Access
- 20 August 2026
- Published by California Digital Library (CDL) in EcoEvoRxiv
Contextual evidence for categorising interactions and guiding their discovery
A surge of recent methods tackles incomplete ecological interaction data through link prediction. However, validating predictions means field-sampling an immense number of candidate interactions, possibly with a poorly suited method. This costly endeavour lacks operative guides grounded in strong ecological and statistical theory. We present a guided-sampling framework combining link prediction with within-system variability across replicated networks as contextual evidence, generating a fine-resolution, ecologically-aware link taxonomy. It resolves categories ecologists have long struggled to separate, such as forbidden interactions versus those merely missing from the local sample, pinpoints the most probable unknowns, and directs targeted, cost-efficient discovery. We formalise how confidence in each assignment is quantified and grows as evidence accumulates. Treating validation as evidence accumulation makes it a tractable sampling problem rather than a permanent limit on knowledge. We provide a worked analysis of an empirical data set and an online guide at http://lpguide.ecomplab.com.
- 20 August 2026
- Published by California Digital Library (CDL) in EcoEvoRxiv
Open AccessContextual evidence for categorising interactions and guiding their discovery
A surge of recent methods tackles incomplete ecological interaction data through link prediction. However, validating predictions means field-sampling an immense number of candidate interactions, possibly with a poorly suited method. This costly endeavour lacks operative guides grounded in strong ecological and statistical theory. We present a guided-sampling framework combining link prediction with within-system variability across replicated networks as contextual evidence, generating a fine-resolution, ecologically-aware link taxonomy. It resolves categories ecologists have long struggled to separate, such as forbidden interactions versus those merely missing from the local sample, pinpoints the most probable unknowns, and directs targeted, cost-efficient discovery. We formalise how confidence in each assignment is quantified and grows as evidence accumulates. Treating validation as evidence accumulation makes it a tractable sampling problem rather than a permanent limit on knowledge. We provide a worked analysis of an empirical data set and an online guide at http://lpguide.ecomplab.com.
- Open Access
- 4 July 2026
- Published by Wiley in Methods in Ecology and Evolution
- Vol. 17 (8) , 2494-2509
Predicting ecological interactions across space through pairwise integration of latent network patterns
Ecological communities are complex and exhibit considerable spatial variability, presenting challenges in accurately understanding these systems. A primary obstacle in ecological research is the existence of ‘missing links’ between species: inevitable unobserved interactions that limit our comprehension of ecological networks and their response to change. While link prediction methods have been developed to address this challenge, most approaches overlook the intrinsic spatial variability of ecological systems. We introduce a flexible, spatially explicit framework based on matrix decomposition that leverages latent structural patterns to predict missing interactions and their strength, without requiring species traits or environmental data. The framework integrates information from paired auxiliary and target networks (locations) using thresholded SVD for link prediction. We applied it to plant–pollinator networks across the Canary Islands, performing pairwise predictions between locations, comparing them to within‐location predictions (as a control), and quantifying how spatial variability influences predictive performance. Predictions revealed that latent network structure contains substantial predictive information, with scores consistently exceeding a random baseline (mean F 0.5 = 0.67 ± 0.02 SD), while being less sensitive to interaction strength. The method enabled identifying plausible gaps in the data and producing ecologically coherent predictions. Incorporating information from auxiliary locations enhanced predictive accuracy in certain cases, but success depended on spatial context: predictions were most reliable when derived from nearby, ecologically similar locations, and declined with increasing geographic and ecological distance, consistent with a distance‐decay effect. We conclude that the predictability of missing links is spatially variable, reflecting both network and species‐level heterogeneity. These patterns provide insights into network structure and the ecological processes shaping it, complementing trait‐based approaches. While network structure offers rich predictive information, spatial context is essential for applying it effectively: ignoring spatial variability can obscure ecological signals and inflate predictive error. Our framework is computationally efficient, transferable, and readily applicable to any system with spatial or temporal replication. It can be used for a variety of ecological contexts, including island systems, fragmented landscapes, and environmental gradients, making it a practical and scalable tool for advancing link prediction in ecology.
- 4 July 2026
- Published by Wiley in Methods in Ecology and Evolution
- Vol. 17 (8) , 2494-2509
Open AccessPredicting ecological interactions across space through pairwise integration of latent network patterns
Ecological communities are complex and exhibit considerable spatial variability, presenting challenges in accurately understanding these systems. A primary obstacle in ecological research is the existence of ‘missing links’ between species: inevitable unobserved interactions that limit our comprehension of ecological networks and their response to change. While link prediction methods have been developed to address this challenge, most approaches overlook the intrinsic spatial variability of ecological systems. We introduce a flexible, spatially explicit framework based on matrix decomposition that leverages latent structural patterns to predict missing interactions and their strength, without requiring species traits or environmental data. The framework integrates information from paired auxiliary and target networks (locations) using thresholded SVD for link prediction. We applied it to plant–pollinator networks across the Canary Islands, performing pairwise predictions between locations, comparing them to within‐location predictions (as a control), and quantifying how spatial variability influences predictive performance. Predictions revealed that latent network structure contains substantial predictive information, with scores consistently exceeding a random baseline (mean F 0.5 = 0.67 ± 0.02 SD), while being less sensitive to interaction strength. The method enabled identifying plausible gaps in the data and producing ecologically coherent predictions. Incorporating information from auxiliary locations enhanced predictive accuracy in certain cases, but success depended on spatial context: predictions were most reliable when derived from nearby, ecologically similar locations, and declined with increasing geographic and ecological distance, consistent with a distance‐decay effect. We conclude that the predictability of missing links is spatially variable, reflecting both network and species‐level heterogeneity. These patterns provide insights into network structure and the ecological processes shaping it, complementing trait‐based approaches. While network structure offers rich predictive information, spatial context is essential for applying it effectively: ignoring spatial variability can obscure ecological signals and inflate predictive error. Our framework is computationally efficient, transferable, and readily applicable to any system with spatial or temporal replication. It can be used for a variety of ecological contexts, including island systems, fragmented landscapes, and environmental gradients, making it a practical and scalable tool for advancing link prediction in ecology.
- Open Access
- 10 June 2026
- Published by The Royal Society in Journal of The Royal Society Interface
- Vol. 23 (239)
The structure of pairwise competition drives sudden regime shifts in a microbe–plasmid model
Plasmids play a crucial role in bacterial communities by facilitating the horizontal transfer of genetic material, particularly genes that confer resistance to different stressor agents. In this work, we introduce a microbe–plasmid model to study the conditions under which a plasmid-carrying bacterial subpopulation can invade its corresponding plasmid-free counterpart, as well as those under which it can be eradicated from the population through potential intervention strategies. We reveal the variety of dynamic regimes of the model depending on the parameter values. In particular, we show that a heterogeneous structure of the competition between the two subpopulations drives the possibility of coexistence of stationary states (i.e. bistability), which creates conditions for sudden regime shifts and hysteresis. We found that the success of efforts to prevent or eliminate plasmid-encoded antibiotic resistance (ABR) depends not only on the type or strength of the intervention but also critically on the underlying ecological interactions that govern competition.
- 10 June 2026
- Published by The Royal Society in Journal of The Royal Society Interface
- Vol. 23 (239)
Open AccessThe structure of pairwise competition drives sudden regime shifts in a microbe–plasmid model
Plasmids play a crucial role in bacterial communities by facilitating the horizontal transfer of genetic material, particularly genes that confer resistance to different stressor agents. In this work, we introduce a microbe–plasmid model to study the conditions under which a plasmid-carrying bacterial subpopulation can invade its corresponding plasmid-free counterpart, as well as those under which it can be eradicated from the population through potential intervention strategies. We reveal the variety of dynamic regimes of the model depending on the parameter values. In particular, we show that a heterogeneous structure of the competition between the two subpopulations drives the possibility of coexistence of stationary states (i.e. bistability), which creates conditions for sudden regime shifts and hysteresis. We found that the success of efforts to prevent or eliminate plasmid-encoded antibiotic resistance (ABR) depends not only on the type or strength of the intervention but also critically on the underlying ecological interactions that govern competition.
- Open Access
- 26 May 2026
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 22 (5) , e1014339
The interplay between ecological networks drives host-plasmid community dynamics
Plasmids drive evolution by transferring traits across microbial hosts. Transmission depends on both host–plasmid (infection) and plasmid–plasmid (compatibility) interactions, yet how the structure of these networks shapes transmission remains poorly understood. We hypothesized that these two ecological networks interact in non-additive ways to influence community outcomes. To test this, we developed a stochastic agent-based model that embeds both network structures and simulates coupled host–plasmid dynamics. We systematically varied the structure of each network, both individually and in combination, to isolate the effect of structure on host-plasmid dynamics. A modular (interactions organized into clusters) and hub (interactions concentrated on the highly connected) plasmid-plasmid compatibility network promoted transient host coexistence, while a modular host-plasmid infection network promoted plasmid diversity and stable host coexistence. Importantly, structured networks interacted non-additively, and their impact was most apparent when plasmid carriage imposed a moderate fitness cost on hosts. For example, combining a modular infection network with a hub compatibility network reversed the expected plasmid prevalence patterns, demonstrating that the structure of one network can counteract the effects of the other. We further re-parameterized our model to recapitulate empirical host-plasmid community dynamics, showing that infection network structure can strongly shape plasmid prevalence even in the presence of substantial biological heterogeneity. Our results highlight the necessity of jointly considering host–plasmid infection and plasmid–plasmid compatibility networks to understand host–plasmid community dynamics and their eco-evolutionary potential. More broadly, this work provides an initial mechanistic framework for generating testable hypotheses and underscores that systems involving multiple hosts and infectious agents require explicit consideration of how different ecological networks interact to shape community dynamics.
- 26 May 2026
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 22 (5) , e1014339
Open AccessThe interplay between ecological networks drives host-plasmid community dynamics
Plasmids drive evolution by transferring traits across microbial hosts. Transmission depends on both host–plasmid (infection) and plasmid–plasmid (compatibility) interactions, yet how the structure of these networks shapes transmission remains poorly understood. We hypothesized that these two ecological networks interact in non-additive ways to influence community outcomes. To test this, we developed a stochastic agent-based model that embeds both network structures and simulates coupled host–plasmid dynamics. We systematically varied the structure of each network, both individually and in combination, to isolate the effect of structure on host-plasmid dynamics. A modular (interactions organized into clusters) and hub (interactions concentrated on the highly connected) plasmid-plasmid compatibility network promoted transient host coexistence, while a modular host-plasmid infection network promoted plasmid diversity and stable host coexistence. Importantly, structured networks interacted non-additively, and their impact was most apparent when plasmid carriage imposed a moderate fitness cost on hosts. For example, combining a modular infection network with a hub compatibility network reversed the expected plasmid prevalence patterns, demonstrating that the structure of one network can counteract the effects of the other. We further re-parameterized our model to recapitulate empirical host-plasmid community dynamics, showing that infection network structure can strongly shape plasmid prevalence even in the presence of substantial biological heterogeneity. Our results highlight the necessity of jointly considering host–plasmid infection and plasmid–plasmid compatibility networks to understand host–plasmid community dynamics and their eco-evolutionary potential. More broadly, this work provides an initial mechanistic framework for generating testable hypotheses and underscores that systems involving multiple hosts and infectious agents require explicit consideration of how different ecological networks interact to shape community dynamics.
- Open Access
- 14 January 2026
- Published by Oxford University Press (OUP) in ISME Communications
- Vol. 6 (1)
Host traits and environmental factors shape infection heterogeneity in wild rat–protozoa networks
The occurrence of microbes in animal hosts is highly heterogeneous, shaped by interactions among host traits, environmental context, and microbial diversity. Understanding this heterogeneity is particularly critical for endoparasite infections, where some hosts harbor diverse, high-burden assemblages that elevate disease spread and spillover risk. Yet the mechanisms underlying such heterogeneity remain poorly understood in wild systems, especially at the individual-host level. We addressed this challenge by studying protozoan infections in introduced black rats (Rattus rattus) across environmental gradients in Madagascar. Using network-based stochastic block modeling, we identified three infection profiles capturing meaningful variation in protozoan richness and composition, providing a structured framework for understanding heterogeneity. To uncover the predictors of these profiles, we trained machine-learning models incorporating host traits with environmental variables. Our models consistently outperformed no-skill baselines, with host traits contributing |$\sim$|40% more to predictions than environmental factors. Body mass and gut microbiome composition emerged as the strongest host predictors, while rat and other non-native species densities were the most influential environmental predictors. These results show that infection heterogeneity arises from the interplay of intrinsic host traits and extrinsic environmental conditions. Our approach illustrates how combining network analysis with predictive modeling can (i) uncover latent heterogeneity in host–microbe associations, (ii) identify the relative contribution of the factors driving this heterogeneity, and (iii) predict host infection profiles. Our framework advances microbial ecology by linking host traits, microbial communities, and environmental context, while also informing disease ecology at human–animal interfaces where zoonotic pathogens circulate.
- 14 January 2026
- Published by Oxford University Press (OUP) in ISME Communications
- Vol. 6 (1)
Open AccessHost traits and environmental factors shape infection heterogeneity in wild rat–protozoa networks
The occurrence of microbes in animal hosts is highly heterogeneous, shaped by interactions among host traits, environmental context, and microbial diversity. Understanding this heterogeneity is particularly critical for endoparasite infections, where some hosts harbor diverse, high-burden assemblages that elevate disease spread and spillover risk. Yet the mechanisms underlying such heterogeneity remain poorly understood in wild systems, especially at the individual-host level. We addressed this challenge by studying protozoan infections in introduced black rats (Rattus rattus) across environmental gradients in Madagascar. Using network-based stochastic block modeling, we identified three infection profiles capturing meaningful variation in protozoan richness and composition, providing a structured framework for understanding heterogeneity. To uncover the predictors of these profiles, we trained machine-learning models incorporating host traits with environmental variables. Our models consistently outperformed no-skill baselines, with host traits contributing |$\sim$|40% more to predictions than environmental factors. Body mass and gut microbiome composition emerged as the strongest host predictors, while rat and other non-native species densities were the most influential environmental predictors. These results show that infection heterogeneity arises from the interplay of intrinsic host traits and extrinsic environmental conditions. Our approach illustrates how combining network analysis with predictive modeling can (i) uncover latent heterogeneity in host–microbe associations, (ii) identify the relative contribution of the factors driving this heterogeneity, and (iii) predict host infection profiles. Our framework advances microbial ecology by linking host traits, microbial communities, and environmental context, while also informing disease ecology at human–animal interfaces where zoonotic pathogens circulate.
- Open Access
- 1 December 2025
- Published by Wiley in Journal of Applied Ecology
- Vol. 63 (1)
Local knowledge enhances the sustainability of interconnected fisheries
Local knowledge (LK) refers to the ancestral understanding that Indigenous Peoples and local communities have developed over centuries through trial‐and‐error and hands‐on management of natural resources. LK may provide valuable insights for biodiversity conservation and human well‐being. However, its effectiveness remains under‐explored at large scales, especially where multiple communities manage ecosystems. One example is fisheries, which form complex, interconnected networks where fish move across spatial boundaries between managed areas. Fisheries are critical for food security and income, yet face threats from overharvesting. Fisheries Co‐Management (FCM)—a partnership between local communities and governments—leverages LK. However, the value of LK in designing protection strategies remains unclear. Using a process‐based dynamical model parameterized with empirical data, we evaluated FCM strategies for pirarucu ( Arapaima gigas ) fisheries, which form a metapopulation network of protected and unprotected lakes in the Brazilian Amazon. We combined our metapopulation model with LK, fish biology and network theory to assess how lake protection, fishing quotas and illegal fishing impact pirarucu population abundance at the riverscape scale. By analysing 13 FCM‐protected lakes and 18 unprotected lakes, we contrasted six hypothesis‐driven management strategies against the current one, which is based on LK. In all strategies, protected lakes support higher pirarucu populations and buffer against increased fishing pressure, while unprotected lakes face population collapse due to the lack of fishing regulations. While a strategy that provides the best outcomes in terms of metapopulation persistence was based on pirarucu carrying capacity, the currently applied FCM strategy closely matched its efficacy. Synthesis and applications . Our modelling approach allows managers to compare alternative conservation strategies under different socio‐ecological scenarios, highlighting trade‐offs and guiding investment of effort and resources. While immediately valuable for pirarucu management in the Middle Juruá, the framework scales across tiers of applicability, each requiring progressively greater model adaptation: from supporting FCM in other Amazonian regions (with minimal adjustment), to adaptation for other riverine fisheries and ultimately to broader socio‐ecological systems. In this way, we provide both system‐specific insights and a flexible tool for advancing sustainable management of natural resources across contexts.
- 1 December 2025
- Published by Wiley in Journal of Applied Ecology
- Vol. 63 (1)
Open AccessLocal knowledge enhances the sustainability of interconnected fisheries
Local knowledge (LK) refers to the ancestral understanding that Indigenous Peoples and local communities have developed over centuries through trial‐and‐error and hands‐on management of natural resources. LK may provide valuable insights for biodiversity conservation and human well‐being. However, its effectiveness remains under‐explored at large scales, especially where multiple communities manage ecosystems. One example is fisheries, which form complex, interconnected networks where fish move across spatial boundaries between managed areas. Fisheries are critical for food security and income, yet face threats from overharvesting. Fisheries Co‐Management (FCM)—a partnership between local communities and governments—leverages LK. However, the value of LK in designing protection strategies remains unclear. Using a process‐based dynamical model parameterized with empirical data, we evaluated FCM strategies for pirarucu ( Arapaima gigas ) fisheries, which form a metapopulation network of protected and unprotected lakes in the Brazilian Amazon. We combined our metapopulation model with LK, fish biology and network theory to assess how lake protection, fishing quotas and illegal fishing impact pirarucu population abundance at the riverscape scale. By analysing 13 FCM‐protected lakes and 18 unprotected lakes, we contrasted six hypothesis‐driven management strategies against the current one, which is based on LK. In all strategies, protected lakes support higher pirarucu populations and buffer against increased fishing pressure, while unprotected lakes face population collapse due to the lack of fishing regulations. While a strategy that provides the best outcomes in terms of metapopulation persistence was based on pirarucu carrying capacity, the currently applied FCM strategy closely matched its efficacy. Synthesis and applications . Our modelling approach allows managers to compare alternative conservation strategies under different socio‐ecological scenarios, highlighting trade‐offs and guiding investment of effort and resources. While immediately valuable for pirarucu management in the Middle Juruá, the framework scales across tiers of applicability, each requiring progressively greater model adaptation: from supporting FCM in other Amazonian regions (with minimal adjustment), to adaptation for other riverine fisheries and ultimately to broader socio‐ecological systems. In this way, we provide both system‐specific insights and a flexible tool for advancing sustainable management of natural resources across contexts.
- Open Access
Predicting ecological interactions across space through pairwise integration of latent network patterns
Ecological communities are complex and exhibit considerable spatial variability, presenting challenges in accurately understanding these systems. A primary obstacle in ecological research is the existence of ‘missing links’ between species: inevitable unobserved interactions that limit our comprehension of ecological networks and their response to change. While link prediction methods have been developed to address this challenge, most approaches overlook the intrinsic spatial variability of ecological systems. We introduce a flexible, spatially explicit framework based on matrix decomposition that leverages latent structural patterns to predict missing interactions and their strength, without requiring species traits or environmental data. The framework integrates information from paired auxiliary and target networks (locations) using thresholded SVD for link prediction. We applied it to plant–pollinator networks across the Canary Islands, performing pairwise predictions between locations, comparing them to within-location predictions (as a control), and quantifying how spatial variability influences predictive performance. Predictions revealed that latent network structure contains substantial predictive information, with F 0.5 scores consistently exceeding a random baseline (mean F 0.5 = 0.67 ± 0.02 SD), while being less sensitive to interaction strength. The method enabled identifying plausible gaps in the data and producing ecologically coherent predictions. Incorporating information from auxiliary locations enhanced predictive accuracy in certain cases, but success depended on spatial context: predictions were most reliable when derived from nearby, ecologically similar locations, and declined with increasing geographic and ecological distance, consistent with a distance-decay effect. We conclude that the predictability of missing links is spatially variable, reflecting both network and species-level heterogeneity. These patterns provide insights into network structure and the ecological processes shaping it, complementing trait-based approaches. While network structure offers rich predictive information, spatial context is essential for applying it effectively: ignoring spatial variability can obscure ecological signals and inflate predictive error. Our framework is computationally efficient, transferable, and readily applicable to any system with spatial or temporal replication. It can be used for a variety of ecological contexts, including island systems, fragmented landscapes, and environmental gradients, making it a practical and scalable tool for advancing link prediction in ecology.
Open AccessPredicting ecological interactions across space through pairwise integration of latent network patterns
Ecological communities are complex and exhibit considerable spatial variability, presenting challenges in accurately understanding these systems. A primary obstacle in ecological research is the existence of ‘missing links’ between species: inevitable unobserved interactions that limit our comprehension of ecological networks and their response to change. While link prediction methods have been developed to address this challenge, most approaches overlook the intrinsic spatial variability of ecological systems. We introduce a flexible, spatially explicit framework based on matrix decomposition that leverages latent structural patterns to predict missing interactions and their strength, without requiring species traits or environmental data. The framework integrates information from paired auxiliary and target networks (locations) using thresholded SVD for link prediction. We applied it to plant–pollinator networks across the Canary Islands, performing pairwise predictions between locations, comparing them to within-location predictions (as a control), and quantifying how spatial variability influences predictive performance. Predictions revealed that latent network structure contains substantial predictive information, with F 0.5 scores consistently exceeding a random baseline (mean F 0.5 = 0.67 ± 0.02 SD), while being less sensitive to interaction strength. The method enabled identifying plausible gaps in the data and producing ecologically coherent predictions. Incorporating information from auxiliary locations enhanced predictive accuracy in certain cases, but success depended on spatial context: predictions were most reliable when derived from nearby, ecologically similar locations, and declined with increasing geographic and ecological distance, consistent with a distance-decay effect. We conclude that the predictability of missing links is spatially variable, reflecting both network and species-level heterogeneity. These patterns provide insights into network structure and the ecological processes shaping it, complementing trait-based approaches. While network structure offers rich predictive information, spatial context is essential for applying it effectively: ignoring spatial variability can obscure ecological signals and inflate predictive error. Our framework is computationally efficient, transferable, and readily applicable to any system with spatial or temporal replication. It can be used for a variety of ecological contexts, including island systems, fragmented landscapes, and environmental gradients, making it a practical and scalable tool for advancing link prediction in ecology.
- Open Access
- 14 November 2025
- Published by Wiley in Ecology Letters
- Vol. 28 (11) , e70255
Differential Assembly of Core and Non‐Core Host‐Microbe Network Structures Along a Land‐Use Change Gradient
Microbial communities are fundamental to host health, yet their assembly dynamics under environmental change remain poorly understood. We analysed individual‐level host‐microbe networks in the non‐native wild black rats ( Rattus rattus ) across a land‐use gradient in Madagascar. By applying a moving prevalence threshold, we distinguished between core and non‐core microbes and compared the assembly drivers shaping their network structures. Non‐core microbes formed fragmented, modular networks shaped mainly by heterogeneous selection, reflecting environmental filtering. In contrast, core microbes exhibited stable, less modular networks driven primarily by stochastic ecological drift. These distinct assembly processes persisted across thresholds, highlighting fundamental differences in microbial structuring. Land‐use change significantly influenced the modular structure of non‐core microbes but had minimal effects on core microbes, demonstrating the differential sensitivity of microbial groups to environmental variation. This study advances our understanding of host–microbe interactions and provides a framework for assessing microbiome assembly under anthropogenic change.
- 14 November 2025
- Published by Wiley in Ecology Letters
- Vol. 28 (11) , e70255
Open AccessDifferential Assembly of Core and Non‐Core Host‐Microbe Network Structures Along a Land‐Use Change Gradient
Microbial communities are fundamental to host health, yet their assembly dynamics under environmental change remain poorly understood. We analysed individual‐level host‐microbe networks in the non‐native wild black rats ( Rattus rattus ) across a land‐use gradient in Madagascar. By applying a moving prevalence threshold, we distinguished between core and non‐core microbes and compared the assembly drivers shaping their network structures. Non‐core microbes formed fragmented, modular networks shaped mainly by heterogeneous selection, reflecting environmental filtering. In contrast, core microbes exhibited stable, less modular networks driven primarily by stochastic ecological drift. These distinct assembly processes persisted across thresholds, highlighting fundamental differences in microbial structuring. Land‐use change significantly influenced the modular structure of non‐core microbes but had minimal effects on core microbes, demonstrating the differential sensitivity of microbial groups to environmental variation. This study advances our understanding of host–microbe interactions and provides a framework for assessing microbiome assembly under anthropogenic change.