Antisense vivo-morpholino oligonucleotides (vivo-MOs) allow transient gene knockdown in adult organisms with high specificity and low toxicity. Vivo-MOs are used in cell culture and in many established model organisms, but a method for their use has not been described in threepsine stickleback ( Gasterosteus aculeatus (Linnaeus, 1758)). Stickleback are an emerging model system used in evolutionary and ecological genetic studies. While genomic techniques are commonly used in stickleback research, there are few studies and tools available to assess gene function in-vivo, especially for genes that may be difficult to knock out by CRISPR (e.g., lethal knock-outs). Here, we test the use of splice-blocking vivo-MOs for gene knockdown in stickleback using intraperitoneal injection of vivo-MOs targeting three candidate genes. Gene expression was assessed in the liver, spleen, and intestine. Successful knockdown of Spi1b was observed in the spleen, however, we observed no other significant knockdown at either timepoint tested. Injection of a fluorescently labeled control vivo-MO confirmed delivery to each target organ, validating this approach, but delivery was variable which may explain inconsistent effects. These results indicate that vivo-MOs have potential as a tool for in-vivo gene knockdown in stickleback. Optimizing delivery methods could improve reproducibility and knockdown efficiency in future studies.

Untargeted metabolomics offers a powerful lens for quantifying high-dimensional phenotypic variation within and among species in nature, but has yet to be widely adopted in evolutionary ecology. Some important initial questions are whether metabolome composition differs among populations, and to what extent such variation is genetic or plastic. Here, we use untargeted liquid chromatography tandem mass spectrometry to characterize the relative abundance of 5,939 molecular features of the threespine stickleback (Gasterosteus aculeatus) liver metabolome. Native lake populations differ in metabolome composition, reflecting effects of sex, size, geography, and population ecotype (benthic versus limnetic). Stickleback from these lakes were translocated to found new populations in nine recently fishless lakes, permuting fish ecotypes across benthic and limnetic lake habitats. Several generations later, metabolomes in these experimental populations reflect effects both of their genetic ancestry (e.g., taurocholic acid, a cholane steroid bile acid, was elevated in limnetic-ancestries), as well as their present habitat (e.g., acylcarnitines). Additionally, ecotypes transplanted into a habitat to which they were maladapted exhibited a distinctive metabolomic profile. We conclude that stickleback exhibit both heritable and plastic among-population differences in liver metabolome, which could represent an important phenotypic basis of rapid evolution, population divergence, and perhaps local adaptation.

Introduction Host-parasite interactions are ubiquitous and are important drivers of host diversification and evolution. In particular, host immune systems are frequent targets of parasite-driven selection. The resulting rapid evolution of immune genes is usually framed as an ongoing 'arms race' between a co-evolving pair of host and parasite species. However, immune evolution may often be driven by the acquisition of a new and unfamiliar parasite. For instance, when marine populations of threespine stickleback (Gasterosteus aculeatus) colonized freshwater lakes approximately 12,000 years ago, they encountered the freshwater-restricted cestode Schistocephalus solidus and evolved resistance.Methods We compared the transcriptomic responses of lab-reared sticklebacks from three populations of stickleback with varying cestode susceptibilities when exposed to several immune stimuli (alum, cestode protein, or a control injection).Results The resulting changes in expression reveal strong evidence of shared and population-specific responses during the evolution of defense against a new parasite. Our investigation highlights the roles of several key immunological processes in underlying a general physiological response to tissue damage (fibrosis) and the importance of regulating fibrosis as a necessary step for its co-option into defense against S. solidus tapeworms. Furthermore, we highlighted changes in the expression of fibrosis-associated genes, which facilitate faster and more targeted deployment of this defense mechanism against parasites. Fish from the most fibrosis-prone population exhibited constitutively higher expression of fibrosis-associated genes and stronger downregulation of these genes after an initial stimulus from injected cestode proteins.Conclusion Our results provide strong evidence that changes in gene regulation and increased negative feedback to mitigate immunopathology are essential steps in the evolutionary co-option of an existing pathway to defend against a new parasite infection.

Understanding how immune variation arises in natural populations requires disentangling the relative contributions of host genetic differences, environmental variation, and parasite effects, which is rarely possible in wild systems. Threespine stickleback populations vary in their use of intraperitoneal fibrosis as a defense against the helminth parasite Schistocephalus solidus, providing a natural system to study the genetic and ecological drivers of immune variation. We combined a 46-lake field survey with common garden experiments on 20 representative populations exposed to multiple parasite genotypes to test whether population differences in fibrosis persist under controlled conditions and whether they depend on parasite genotype or lake ecology. Fibrosis variation was strongly heritable, with both constitutive and inducible components persisting under common garden conditions. In contrast, parasite genotype had only a weak effect on fibrosis responses. Moreover, inducible fibrosis covaried with lake environmental conditions, with populations from more eutrophic-like lakes exhibiting stronger responses than those from more oligotrophic-like lakes. Together, these results reveal ecologically structured divergence in heritable immune responses among natural populations.

Our understanding of the vertebrate immune system is dominated by a few model organisms such as mice. This use of a few model systems is reasonable if major features of the immune systems evolve slowly and are conserved across most vertebrates, but may be problematic if there is substantial macroevolutionary change in immune responses. Here, we present a test of the macroevolutionary stability, across 14 species of ray-finned fishes, of the transcriptomic response to a standardized immune challenge. Intraperitoneal injection of an immune adjuvant (alum) induces a fibrosis response in nearly all jawed fishes, which in some species contributes to anti-helminth protection. Despite this conserved phenotypic response, the underlying transcriptomic response is highly inconsistent across species. Although many gene orthogroups exhibit differential expression between saline versus alum-injected fish in at least one species, few orthogroups exhibit consistent differential expression across species. This result suggests that although the phenotypic response to alum (fibrosis) is highly conserved, the underlying gene regulatory architecture is very flexible and cannot readily be extrapolated from any one species to fishes (or vertebrates) more broadly. The vertebrate immune response is remarkably changeable over macroevolutionary time, requiring a diversity of model organisms to describe effectively.

Conservation actions often assume implicitly that heritabilities are zero and that threatened populations cannot adapt to changing environments. To illustrate, we evaluated the last 10 y of recovery plans for US threatened and endangered species and found that only 4% assessed within-population adaptability. This omission reflects the common assumption that population adaptation is too slow or inconsequential to affect conservation practice in the short term. Yet, the median heritability (h2) and evolvability (IA) across many studies are not zero as assumed, but 0.3 and 0.4, respectively, based on a compilation of estimated values. This moderate heritability could rescue some populations. By compiling literature on conservation assessments, we detail how considering adaptability can shift conservation priorities, alter management recommendations, and provide additional ways to rescue declining populations and species. Based on these findings, we advocate for including population adaptability into conservation plans and adopting the prior expectation of moderate adaptability (h2 = 0.3, IA = 0.4) along with its uncertainty, as a starting point when better information is lacking. This moderate adaptability could allow some species to respond naturally to environmental change while directing limited resources toward the species that need it most.

Weighted Gene Co-expression Network Analysis (WGCNA) is routinely applied to pooled datasets from multiple biological populations, genotypes, or treatment groups, implicitly assuming a shared module structure across groups. While the distortion of pairwise correlations by pooling heterogeneous groups is well established statistically, three aspects of this problem have received little systematic attention in the context of co-expression network analysis: the extent to which pooling disrupts the discrete module-level community structure inferred by WGCNA; whether this disruption is detectable from the global topology metrics researchers routinely report; and how prevalent the pooling practice is in published multi-group WGCNA studies. Using analytical toy examples and a four-scenario simulation framework, we address all three questions. Module preservation Zsummary scores declined progressively with between-population divergence, from full preservation under identical populations (mean median Zsummary = 25.2 ± 3.3, 95% interval 19.0--30.7 across 20 simulation replicates) to substantial disruption when both network structure and mean expression differed (mean median Zsummary = 11.9 ± 1.0, 95% interval 10.2--13.5). This disruption was undetectable from global topology metrics: modularity and clustering coefficient remained stable across all scenarios, while edge density was sensitive but non-specific. These findings were corroborated in an empirical reanalysis of divergent lake and stream stickleback transcriptomes, where merged analysis collapsed 26 lake-specific and 59 stream-specific modules into only 19 merged modules. A survey of 100 publications found that 78.7% (95% CI 69.4--87.9%) of multi-group WGCNA studies with sufficient methodological reporting used a single merged analysis. Results were robust across network sizes of 250--1,000 genes and rewiring rates of 10--50%. We provide concrete recommendations including module preservation testing in both directions, population-specific baseline networks, and consensus WGCNA as a principled alternative.

Species introductions and transplants offer powerful contexts to understand evolutionary patterns and processes, and they are increasingly critical for conservation. However, introduction success varies widely, and predicting outcomes remains challenging. Introducing multiple source populations should increase the chance of success, while also providing an opportunity to explore the factors that predict success of individual source populations in the same environment. We used replicated, mixed-population introductions of >12,000 threespine stickleback (Gasterosteus aculeatus) to test whether source population success could be predicted by environmental matching between source and recipient environments and/or by intrinsic source population characteristics. We introduced four to eight source populations of stickleback into each of nine natural lakes and tracked their relative success over the following two years (up to two generations). Source population success was largely consistent across lakes, despite divergent environmental conditions. These results point to the importance of intrinsic source population characteristics rather than environmental matching in predicting introduction success in natural settings. Source populations that were consistently successful tended to have greater stress tolerance (mortality rate during translocation) and higher genetic diversity, though these relationships were not conclusive. Our study highlights the value of considering factors that generate fitness differences independent of environmental contexts in predicting ecological and evolutionary dynamics and planning conservation programs. ### Competing Interest Statement The authors have declared no competing interest. U.S. National Science Foundation, FAIN-2133740, DMS-1716803 Natural Sciences and Engineering Research Council, https://ror.org/01h531d29, RGPIN 249551-2013, RGPIN 2016-05143 Fonds de Recherche du Québec – Nature et Technologies, https://ror.org/00b9f9778, 2024-PR-324806 Swiss National Science Foundation, https://ror.org/00yjd3n13, TMAG-3_209309/1 Centre de la science de la biodiversité du Québec, https://ror.org/04zbfd360 Polar Knowledge Canada, https://ror.org/00rfash91

Abstract Antimicrobial resistance (AMR) genes are increasingly recognized as an emerging environmental contaminant. Yet, the ecological mechanisms shaping their distribution across natural landscapes remain poorly understood. Here, we quantified AMR gene abundances in microbial communities sampled from wild fish from eight freshwater lakes on Vancouver Island and paired these gene-level measurements with fine-scale limnological and land-use data. Using droplet digital PCR, field surveys, and an iterative spatial forecasting framework that integrates Random Forest models with regression kriging, we explored how watershed-scale processes relate to variation in AMR genes across lakes. Our analyses reveal potential associations between elevated AMR gene levels, changes in water quality, deforestation, and geographic proximity to salmon aquaculture. By integrating data across biological and spatial scales, from genes within microbial communities to lake-level conditions and landscape patterns, this study illustrates the value of combining quantitative molecular measurements with geospatial modeling to identify environmental factors that may promote antimicrobial resistance in natural systems. Our approach provides a proof-of-concept and a general predictive framework for generating hypotheses and informing future monitoring efforts aimed at understanding, managing, and forecasting environmental reservoirs of resistance. Significance Antimicrobial resistance (AMR) genes are ancient components of environmental microbiomes. Yet, the mechanisms that generate modern hotspots of resistance across natural landscapes remain unclear. Here, we reveal how watershed-scale environmental change, including water quality metrics linked with deforestation and proximity to salmon aquaculture, predicts elevated AMR gene levels in the microbiomes of wild fish populations. By combining quantitative droplet digital PCR with ecological data and geospatial modeling, we move beyond isolated surveillance data to identify ecological mechanisms that promote antimicrobial resistance in freshwater ecosystems. This integrative approach provides mechanistic insight into why certain habitats, and the organisms within them, become reservoirs of resistance while others do not. Our findings highlight the importance of ecological context in understanding resistance evolution and offer a predictive tool for informing proactive monitoring and management strategies.

AbstractHost-parasite interactions are likely to involve eco-evolutionary feedbacks. Parasites can impose strong selection on host immune traits. Evolution of these traits can suppress parasite abundance, which alters the strength of selection. Such eco-evo feedbacks can be hard to document in natural settings, where long-term interactions may have converged on relatively stable equilibria. However, when a species disperses into an unoccupied habitat, it can experience altered parasite abundance (e.g., enemy release) or encounter new parasite genotypes. These perturbations may be followed by observable eco-evolutionary dynamics as the newly assembled host-parasite communities interact and evolve. To test this hypothesis, we experimentally founded nine whole-lake populations of threespine stickleback and tracked changes in the prevalence of a parasite (Schistocephalus solidus) and a heritable host immune trait (fibrosis) for 7 years. In native source lakes, among-population differences in tapeworm prevalence and fibrosis were relatively stable across years, suggesting there are alternative eco-evolutionary equilibria. In contrast, the experimentally founded populations initially exhibited enemy release, followed by strong year-to-year fluctuations in infection and fibrosis. In lakes with high initial infection rates, fibrosis subsequently increased in severity, suppressing infections. Conversely, low fibrosis was followed by increased parasite prevalence. This temporal autocorrelation between parasite prevalence and immune traits weakened over time. These results are consistent with the existence of transient eco-evolutionary dynamics between host and parasite in newly founded populations.

Abstract Phenotypic differences among populations can arise through heritable genetic divergence, phenotypic plasticity, or both, making it difficult to determine whether trait-environment correlations observed in nature reflect adaptive evolution. Within threespine stickleback ( Gasterosteus aculeatus ) studies, numerous document morphological differences among allopatric-, parapatric-, and even sympatric populations. These phenotypic differences among populations are often correlated with diet and lake habitat (e.g., lake size), suggesting an adaptive value to the population differences. However, many studies of ecomorphological divergence in stickleback use wild-caught stickleback, which may differ due to evolution or plasticity. Although common garden experiments have confirmed that population differences can be heritable, such experiments typically entail small numbers of populations. Consequently, we still do not know to what extent well-known trait-environment correlations in stickleback are a result of evolution. To address this gap, we reared stickleback embryos from 27 lake populations on Vancouver Island, in a laboratory environment. Morphological differences among populations persist in common-garden fish, confirming a large role for divergent evolution. These heritable differences were associated with environmental variation among lakes, implying an adaptive value. However, some well-known trait-environment relationships in stickleback did not persist in common-garden fish and may be primarily plastic.

Many generalized species are composed of individual specialists that use a small subset of the population's resources. Individuals will specialize on different resources if they have different resource use efficiencies, that is, there are efficiency tradeoffs between different prey. Different resource-use efficiencies may reflect variable morphological, behavioral, or physiological capacities to handle alternative resources. Individual diet specialization can be related to sexual dimorphism, ontogenetic niches, or related to morphological or behavioral variation in the population. However, not all diet variation needs to be related to variation in efficiencies – diet variation may also be related to variation in social status, mating strategy, or territories. Studying diet variation is important both in ecology and evolution. Ecologically, diet variation is important not only when considering predator–prey interactions and competitive interactions but also in population dynamical studies. In evolutionary studies, diets of an individual determine its energy acquisition, which is strongly related to fitness. Thus, differential foraging rates on different prey types will have a big influence on both the ecological and evolutionary dynamics of populations.

The intestine plays a crucial role in physiology, nutrition, and immune function, but intestinal anatomy and cell types have yet to be fully characterized in many fish species, the most diverse group of vertebrates. To address this gap, we characterized the structure and composition of the intestine of threespine stickleback (Gasterosteus aculeatus), an emerging model teleost in biological research. Using histology, myeloperoxidase staining, single-cell RNA sequencing, and RNA in situ hybridization, we defined major intestinal epithelial, immune, stromal, and stem/progenitor populations. Goblet cells were abundant in proximal and hindgut, while myeloperoxidase-positive granulocytes were evenly distributed throughout the intestine. To facilitate future experimental studies of stickleback intestinal function, we also developed the first intestinal organoid culture from stickleback and show that these cultures recapitulate epithelial architecture and retain expression of canonical intestinal epithelial markers. This organoid platform enables future functional studies of mucosal immunity, host-microbe interactions, and intestinal physiology in stickleback and related teleosts. Together, our integrated approach provides a comprehensive cell atlas and a novel experimental model for studying digestive and immune functions in threespine stickleback.

Global changes in land use and nutrient cycling are transforming ecosystems at unprecedented rates, with significant consequences for infectious disease dynamics. Aquatic environments are particularly vulnerable because the interplay of habitat modification, nutrient enrichment, and biodiversity loss can drive pronounced changes in the community composition of food webs, including hosts and parasites. Yet, despite well-documented effects of habitat modification on aquatic communities and food webs, the mechanisms through which these changes influence infectious disease dynamics remain poorly resolved. This gap arises, in part, because it remains challenging to disentangle how multiple stressors interact to shape disease outcomes and quantify parasite levels and host densities from field-collected samples. Here, we illustrate two tools that might help address these challenges. First, highly sensitive droplet digital PCR can quantify infection loads even when the signal:noise ratio is low. Second, stepwise Bayesian path analyses can identify the direct and indirect pathways connecting land-use changes to infectious disease dynamics. As a case study, we examined cyclopoid copepods and their helminth parasite, Schistocephalus solidus, across 47 freshwater lakes on Vancouver Island, a region strongly shaped by commercial logging, including widespread clear-cutting of old-growth forests. Our results reveal a positive correlation between copepod density and deforestation, potentially mediated by associated changes in water quality and calanoid copepods, key competitors of the focal host. ddPCR enabled sensitive detection of extremely low parasite signals in field-collected copepods. We detected positive infections in only 19.5% of the lakes surveyed, highlighting the difficulty of assessing disease dynamics in natural populations. Nonetheless, this study highlights the challenges of linking land-use change to disease outcomes, while also demonstrating that sensitive molecular and statistical tools offer new ways to reveal these hidden connections.

Abstract Ecological speciation is now regarded as one of the primary processes by which new species are generated. Adaptive divergence in allopatry begins this process, but it is often unclear when and how mechanisms that promote reproductive isolation, such as assortative mating and selection against hybrids, evolve. Here, we test for evidence of these mechanisms across replicated secondary contact experiments in natural settings. We introduced four to eight allopatric populations of threespine stickleback ( Gasterosteus aculeatus ), in both single-ecotype and mixed-ecotype treatments, into nine natural lakes, after which we inferred mating patterns by genotyping the resulting F1 generation. Contrary to expectations from the literature, we found no evidence of assortative mating or partial reproductive isolation among the introduced source populations. Instead, we detected evidence of disassortative mating by source population in three lakes and some evidence of disassortative mating by ecotype in one lake. These mating patterns were both context-dependent and population-dependent, varying substantially across lakes receiving the same source populations, and with some source populations generally displaying greater tendencies for disassortment. The absence of positive assortative mating in any replicate demonstrates that adaptive divergence in allopatry alone might be insufficient to generateassortative mating in many cases, while the possibility of disassortative mating in these contexts poses an additional hurdle on the path toward speciation.

Recombination is central to genetics and to evolution of sexually reproducing organisms. However, obtaining accurate estimates of recombination rates, and of how they vary along chromosomes, continues to be challenging. To advance our ability to estimate recombination rates, we present Hi-reComb, a new method and software for estimation of recombination maps from bulk gamete chromosome conformation capture sequencing (Hi-C). Simulations show that Hi-reComb produces robust, accurate recombination landscapes. With empirical data from sperm of five fish species we show the advantages of this approach, including joint assessment of recombination maps and large structural variants, map comparisons using bootstrap, and workflows with trio phasing vs. Hi-C phasing. With off-the-shelf library construction and a straightforward rapid workflow, our approach will facilitate routine recombination landscape estimation for a broad range of studies and model organisms in genetics and evolutionary biology. Hi-reComb is open-source and freely available at https://github.com/millanek/Hi-reComb.

ABSTRACT Theory predicts that virulence evolves as a consequence of selection to optimize transmission, generating a trade off in which increased exploitation enhances transmission but shortens the infectious period. Despite its central role in evolutionary epidemiology, empirical support for this virulence–transmission trade off remains limited, has focused largely on microparasites, and often overlooks variation in host immunity which can fundamentally alter links between virulence and transmission. Here, we provide a rare empirical test of virulence–transmission dynamics in a macroparasite with a complex life cycle. Using a field survey of the helminth parasite, Schistocephalus solidus , and its second intermediate host, threespine stickleback ( Gasterosteus aculeatus ), we quantify how host immune variation shapes relationships among parasite burden (a proxy for virulence) and transmission potential to definitive hosts, piscivorous birds. Importantly, host populations in the focal lakes span a gradient of evolved immune strategies, from low to high fibrosis, a strong anti-growth resistance mechanism. We find that variation in immune timing across host populations constrains the window in which parasites can reach transmissible stages. Subsequent changes in parasite burden scale up to alter transmission potential and reveal a nonlinear relationship consistent with a virulence–transmission trade off. Transmission potential is highest at intermediate parasite burdens, which also corresponds to intermediate immune responses. Together, this work links within-host processes to population-level epidemiological outcomes and demonstrates how host immune variation can shape virulence-transmission relationships. Incorporating immune heterogeneity may therefore help reconcile the mixed empirical support for trade off theory.

Climate warming and chemical pollution shape aquatic ecosystems, yet the physiological mechanisms underlying their combined effects remain unclear. We investigated how projected increases in mean summer surface water temperature alter per- and polyfluoroalkyl substance (PFAS) toxicokinetics and their effects on the physiological performance of sheepshead minnows (Cyprinodon variegatus). Adult fish were chronically exposed to an environmentally relevant PFAS mixture (perfluorooctanesulfonate (PFOS) + perfluorooctanoate (PFOA)) under current and projected mean-temperature scenarios. Tissue PFAS concentrations, whole-organism metabolic rates, swimming performance, reproductive output, and somatic indices were assessed. Temperature modified PFAS tissue concentrations in a compound- and tissue-specific manner, notably promoting PFOA redistribution to eggs. Metabolic responses were temperature-dependent: at 26 °C, higher tissue PFAS concentrations were associated with elevated standard and maximum metabolic rates (SMR and MMR), maintaining aerobic scope (AS). At 28.5 °C, SMR remained stable while MMR and AS declined with rising PFAS, indicating less oxygen for energetically demanding activities. Despite unchanged performance outcomes for swimming and reproduction, increase in hepatosomatic index with increasing tissue PFAS concentrations and altered PFAS distribution suggests detoxification costs. These findings indicate that increases in mean water temperature are likely to exacerbate contaminant stress, with consequences for coastal fish population resilience and offspring development. PFAS risk assessment should consider costressors under projected warming.

MOTIVATION:Single-cell RNA sequencing (scRNA-seq) data analysis is often performed using network projections that produce co-expression networks. These network-based algorithms are attractive because regulatory interactions are fundamentally network-based and there are many tools available for downstream analysis. However, most network-based approaches have two major limitations. First, they are typically unipartite and therefore fail to capture higher-order information. Second, scRNA-seq data are often sparse, so most algorithms for constructing unipartite network projections are inefficient and may overestimate co-expression relationships, or may under-utilize the sparsity when clustering (e.g. with cosine distance). To address these limitations, we propose representing scRNA-seq expression data as hypergraphs, which are generalized graphs where a hyperedge can connect more than two nodes. In this context, hypergraph nodes represent cells, and hyperedges represent genes. Each hyperedge connects all cells in which its corresponding gene is actively expressed, indicating the expression of that gene across different cells. The resulting hypergraph can capture higher-order information and appropriately handle varying levels of data sparsity. This representation enables clustering algorithms to leverage higher-order relationships for improved cell-type differentiation. RESULTS:To distinguish cell types using hypergraph representations of scRNA-seq data, we introduce two novel clustering algorithms: (i) Dual-Importance Preference Hypergraph Walk (DIPHW) and (ii) Co-expression and Memory-Integrated Dual-Importance Preference Hypergraph Walk (CoMem-DIPHW). DIPHW is a new hypergraph-based random walk algorithm that computes cell embeddings by considering the relative importance of genes to cells and cells to genes, incorporating a preference exponent to facilitate clustering. CoMem-DIPHW integrates two unipartite projections, the gene co-expression and cell co-expression networks, along with the cell-gene expression hypergraph derived from single-cell abundance count data into the random walk model. The advantage of CoMem-DIPHW is that it accounts for both local information from single-cell gene expression and global information from pairwise similarity in the two co-expression networks. We benchmark the performance of our algorithms against established and state-of-the-art deep learning approaches using both real-world and simulated scRNA-seq data. Real-world datasets include cells from the human pancreas, mouse pancreas, human brain, and mouse brain tissues. We also use a ground-truth labeled cell-type annotation dataset based on human lung adenocarcinoma cell lines. Quantitative evaluation shows that CoMem-DIPHW consistently outperforms established algorithms and state-of-the-art deep learning algorithms for cell-type clustering. Our proposed algorithms show the greatest improvement on scRNA-seq data with weak modularity. Moreover, CoMem-DIPHW successfully annotates clusters with biologically relevant cell types. Our results highlight the utility of hypergraph representations in the analysis of scRNA-seq data. AVAILABILITY AND IMPLEMENTATION:Our methods are implemented in Python and available on GitHub (https://github.com/wanhe13/CoMem-DIPHW) and archived at Zenodo (https://doi.org/10.5281/zenodo.18927437).