Skip to main content
NCBI home page
Nature Communications logoLink to Nature Communications
. 2026 Sep 7;17:10582. doi: 10.1038/s41467-026-77595-w

Divergent microbial functional pathways and declining redundancy in lake sediments across eutrophication gradients

John K Pearman 1,✉, Jack Sissons 1, Joseph Kanyi Kihika 1, Sean Waters 2, Andrew B H Rees 3, Jamie D Howarth 3, Marcus J Vandergoes 4, Susanna A Wood 5
PMCID: PMC13646285  PMID: 42844285

Abstract

Microbial communities underpin nutrient cycling across aquatic ecosystems and are strongly influenced by human disturbance, yet the implication for functional redundancy remains poorly understood. Lake eutrophication is a global threat to freshwater systems, primarily driven by human activities within lakes and surrounding catchments. We hypothesize that eutrophication alters functional redundancy of key metabolic pathways in lake sediments. To test this, we conducted a national scale metagenomic study of surface sediments from 144 New Zealand lakes spanning a broad nutrient gradient. Increasing eutrophication is associated with reduced taxon-based functional redundancy, the potential of multiple species to perform the same function, indicating a narrower taxa pool supporting core functions. In contrast, abundance-based functional redundancy, characterized by the abundance of organisms that can perform the function, showed pathway-specific responses across trophic gradients. For example, nitrification and denitrification have greater abundance-based functional redundancy under higher nutrient conditions, whereas phosphorus transport shows reduced redundancy. By eroding microbial functional redundancy, eutrophication may increase susceptibility of microbial communities to environmental perturbations, potentially compromising their ability to sustain key metabolic processes. As lakes play key roles in biogeochemical cycles, the lower stability of vital biogeochemical pathways in lake sediments is likely to have an impact across ecosystems globally.

Subject terms: Molecular ecology, Environmental microbiology, Freshwater ecology, Metagenomics

In this study, the authors conduct a metagenomic study of 144 Aotearoa-New Zealand lakes to show that eutrophication impacts the functional redundancy of metabolic functions in surface sediment microbes, by narrowing of the species pool, although abundance shifts could offer some resilience.

Introduction

Human activity is leaving an indelible mark on Earth’s ecosystems, rapidly restructuring the composition and functioning of ecological communities1,2. Freshwater lakes are among the most sensitive sentinels of this transformation3. They respond not only to stressors that directly impact them, but also integrate disturbances from their catchments, making them early warning systems for anthropogenically induced change at local to regional scales4,5.

Microbial communities drive nutrient cycles that sustain lake ecosystems, yet they are profoundly altered by human disturbance6,7. Land-use intensification and corresponding shifts in the trophic states of lakes consistently reshape microbial community composition across spatial and temporal scales6,8–10. These taxonomic shifts can have effects beyond just compositional changes with implications on the functional potential (the metabolic capability) of microbial communities and subsequent consequences for ecosystem processes and services11.

The capacity of ecosystems to withstand such reconfiguration depends on functional redundancy, the coexistence of species that have similar functional roles but different ecological niches12–14. Higher redundancy can stabilize ecosystem functioning by maintaining key processes even when individual taxa/organisms decline or are lost12,15. Recent work by Fässler et al.16 used information-theoretic operationalizations to investigate functional redundancy at the single-trait level. They defined functional redundancy as the capacity of a community to maintain a given function despite the depletion or loss of individual species, enabling the investigation of trends in different functions across environmental pressure gradients. This definition of functional redundancy can be characterized in different ways. Firstly, taxon-based functional redundancy assesses whether the community contains multiple species capable of performing a certain function16. Secondly, abundance-based functional redundancy takes into account the abundance of organisms that can perform a certain function in the community16. While higher functional redundancy generally increases ecosystem resilience and stability13, its effect depends on the type of environmental change. Changes that fundamentally restructure metabolic niches, for example, shifts in redox conditions or electron donor/acceptor availability, may alter the abundance or composition of entire functional groups, even when redundancy within those groups is high. Therefore, functional redundancy provides robustness to many disturbances, but extreme or transformative environmental changes may still impact key ecosystem processes.

Investigations into the prevalence of microbial functional redundancy have produced contrasting results. Some studies report strong correlations between taxonomic changes and function, suggesting low functional redundancy in surface water samples in marine17 and freshwater environments18, whereas others have found high levels of functional redundancy at global scales19, including marine environments20. Despite the central role of sediment microbes in lake ecosystem functioning, the extent of their functional redundancy remains poorly understood. Studies from Poyang Lake, China, have reported limited functional redundancy21, while a paleolimnological investigation in Aotearoa, New Zealand, indicated low microbial resilience to natural perturbations22. The extent to which these communities can maintain functional stability under increasing anthropogenic pressure is a critical but unresolved question.

Lakes across Aotearoa New Zealand provide a natural experiment for testing how anthropogenic pressures shape microbial function across multiple environmental gradients. A national-scale lake study8,23 sampled lake sediments across Aotearoa New Zealand, of which 144 lakes were selected for investigation in the current study (Fig. 1A). Spanning 12° of latitude and encompassing pronounced gradients (Fig. 1B) in trophic state (sediment bacterial trophic index (SBTI), sediment nutrient concentrations (phosphorus (TP), nitrogen (TN), TN:TP ratio and sulfur (S)), elevation (Elev) and catchment land-use (high productivity grassland (HPG)), these systems provide a unique opportunity to examine how lake and catchment degradation influence microbial community function. The microbial communities in these lake sediments vary taxonomically across a trophic gradient8, providing an opportunity to focus on this as a measure of lake degradation and investigate its relationship with the functional potential of microbial communities in lake sediments. To assess this, we applied metagenomics to our nationwide lake surface sediment sample set to test two hypotheses: (1) The functional potential of sediment microbiomes varies across trophic states, with key biogeochemical pathways (e.g., carbon cycling, nutrient transformations) responding differently to the specific pressures associated with each trophic state; and (2) Functional redundancy buffers key biogeochemical processes across trophic gradients; however, redundancy declines under increasing anthropogenic pressure.

Fig. 1. Spatial distribution and principal component analysis of environmental variables across Aotearoa-New Zealand lakes.

Fig. 1

Open in a new tab

Map of the lakes investigated in the study (A), and principal component analysis of the environmental and land use data, with vectors showing trends of environmental variables (B). Points colored by trophic state. TN sediment total nitrogen, TP sediment total phosphorus, S sediment Sulfur, TOC sediment total organic carbon, HPG proportion of high-producing grassland in catchment, SBTI surface bacterial trophic index7, Depth maximum lake depth, Elev Lake Elevation, Lat Latitude and Area Lake Area. Lake characteristics are in Supplementary Data 1 and data for B can be found in Source Data file.

In this work, we show that eutrophication leads to a reduction in taxon-based functional redundancy, with metabolic pathways becoming increasingly reliant on fewer taxa to function. The decline in taxa-based functional redundancy in eutrophic lake sediments suggests heightened vulnerability to both natural and anthropogenic disturbances. Abundance-based functional redundancy showed pathway-specific responses to eutrophication, with increases in redundancy evident for key metabolic pathways (especially nitrogen-related). The increase in abundance-based functional redundancy in key pathways alongside shifts in community composition could counteract the reduction in taxon-based redundancy. Given the key role of lakes in global nutrient and carbon cycling, shifts in functional redundancy of metabolic pathways in lakes could have influences across terrestrial–aquatic–atmospheric interfaces.

Results

The non-redundant database constructed for bacteria in lake surface sediments resulted in a total of 156,537,568 open reading frames (ORF), of which 60,876,927 were annotated against the KEGG (Kyoto Encyclopedia of Genes and Genomes) database with 52,238,998 having a class level taxonomic classification. A total of 304,889 ORFS belonging to KEGG Orthologs (KO) terms which were indicative of key metabolic pathways had a class-level taxonomic classification.

To assess if there was a relationship between the taxonomic composition of the lake sediments and the functional potential, a Mantel correlation test was undertaken between the current functional data and previously published 16S rRNA bacterial metabarcoding8. The correlation revealed a weak but significant (Mantel r = 0.146, p = 0.004) relationship between the microbial taxonomic composition and their functional potential.

To assess if there were statistical differences in the functional potential of microbial communities in lake sediments, multivariate analysis was undertaken. The microbial functional potential (KEGG KO’s) showed clear differences across trophic state (adonis: F = 4.035, p < 0.001; Fig. 2A). At a national scale, functional potential of specific pathways also varied significantly with environmental factors, most strongly with trophic state (as determined using a surface bacterial trophic index [SBTI7]) and the extent of high-productivity grassland in the catchment (Fig. 2B and Supplementary Data 2).

Fig. 2. Multivariate distribution of samples and correlation of functional pathways with environmental variables.

Fig. 2

Open in a new tab

t-distributed Stochastic Neighbor Embedding (t-SNE) multivariate plot showing the relationship of functional potential (counts per million mapped reads per KO) to trophic state (A). Spearman correlations between indicative counts per million mapped reads of indicative genes for specific functional pathways and various environmental parameters (B). Points in A are colored by trophic state. Statistically significant results are denoted by either a + (positive) or – (negative sign). SOX = Sulfur oxidation; DSR Dissimilatory sulfate reduction; ASR Assimilatory sulfate reduction; P transport Phosphorus transport; N fixation Nitrogen fixation; DNRA Dissimilatory nitrate reduction to ammonia; ANR Assimilatory nitrate reduction; WL Wood-Lungdahl pathway; rTCA reductive tricarboxylic acid. SBTI Sediment Bacteria Trophic Index; HPC High productivity grassland; TOC Total Organic Carbon; OM Organic Matter; TN Total Nitrogen; TP Total Phosphorus; TN:TP Nitrogen: Phosphorus ratio; S Sulfur. Source data are provided as a Source Data file.

To give an assessment of how the taxonomic composition of the metagenomic data shifted with trophic state, the metagenomic reads were taxonomically classified with sylph, and the composition was assessed by trophic state (Supplementary Fig. 1). Gammaproteobacteria were observed to increase in abundance with trophic state from 19% in oligotrophic lakes to 31% in supertrophic lakes. Alphaproteobacteria and Nitrospira declined in abundance with trophic state from 13.5% and 8.2% in oligotrophic lakes to 2.7% and 0.2% in supertrophic lakes, respectively.

Pathway-environment relationships revealed clear trends (Fig. 2B and Supplementary Data 2). Calvin cycle carbon fixation was positively correlated with multiple environmental variables, including trophic state (SBTI; R = 0.42; p < 0.001), Total Nitrogen: Total Phosphorous (TN:TP; R = 0.32; p < 0.001) and S (R = 0.31; p < 0.001). Methanogenesis also showed positive correlations with most variables, including trophic state (R = 0.23; p = 0.005), organic matter (R = 0.29; p < 0.001), and TN:TP (R = 0.34; p < 0.001). In contrast, assimilatory nitrate reduction (ANR) and denitrification showed significant negative correlations with trophic state (ANR: R = −0.63; p < 0.001; denitrification: R = −0.34; p < 0.001), TN (ANR: R = −0.24; p = 0.003; denitrification: R = −0.32; p < 0.001), and TN:TP (ANR: R = −0.20; p = 0.015; denitrification: R = −0.30; p < 0.001). Nitrogen fixation was positively associated with nearly all variables except TP and high-productivity grassland. Phosphate transport (Pst) decreased with trophic state (R = −0.24; p = 0.003) and high-productivity grassland cover (R = −0.29; p < 0.001) but increased significantly with organic matter (R = 0.20; p = 0.013), and TN:TP (R = 0.24; p = 0.003). Among sulfur pathways, assimilatory sulfate reduction showed significant negative relationships with environmental variables, excluding TP.

To assess the effect of trophic state on functional redundancy of specific pathways, we applied the approach of Fässler et al.16 to examine changes in both taxon- and abundance-based functional redundancy. The Fässler et al.16 equations indicate that maximum functional redundancy is obtained when the value equals 0, and more negative numbers suggest lower levels of functional redundancy.

The highest levels of taxa-based functional redundancy were observed for the methanogenesis pathway (median score = −1.15; Fig. 3A and Supplementary Table 1) with 13 classes contributing to the pathway. Nitrification (median score = −1.50) and the Wood-Lungdahl pathway (median score = −1.51) exhibited the next highest levels of functional redundancy with 82 and 132 classes contributing to these pathways, respectively (Fig. 3A and Supplementary Fig. 2 and Supplementary Table 1). P transport had the highest abundance-based functional redundancy (Fig. 3B and Supplementary Table 1; median score = −0.14) followed by ANR (median score = −0.44) and Respiration (median score = −0.47), with high contributions from Gammaproteobacteria in all pathways (Supplementary Fig. 2).

Fig. 3. Functional redundancy across Aotearoa-New Zealand lakes.

Fig. 3

Open in a new tab

Taxon-based (A), and Abundance-based (B) functional redundancy across the Aotearoa-New Zealand lakes metagenomic dataset for specific functional pathways. Maximum functional redundancy is obtained at values of 0 and the greater the negative number the lower levels of functional redundancy. SOX Sulfur oxidation; DSR Dissimilatory sulfate reduction; ASR Assimilatory sulfate reduction; P transport Phosphorus transport; N fixation Nitrogen fixation; DNRA Dissimilatory nitrate reduction to ammonia; ANR Assimilatory nitrate reduction; WL Wood-Lungdahl pathway; rTCA reductive tricarboxylic acid. Source data are provided as a Source Data file. Boxplots were created from 144 lakes for each metabolic pathway for both the taxon-based and abundance-based redundancy analysis. For lakes with replication a lake mean was calculated and used for the boxplot. The boxplot shows the median (center) and the first quartile (bottom) and third quartile (top) of the data with the whiskers representing 1.5 x the interquartile range.

The lowest levels of taxa-based functional redundancy were observed for the sulfur cycling pathway, sulfur oxidation (median score = −3.20). The rTCA cycle had the lowest abundance-based (median score = −1.56).

Spearman’s correlations were used to assess the effect of trophic state on the functional redundancy of functional pathways. For most pathways, there was a negative correlation between taxon-based functional redundancy and trophic status (Table 1). This reduction in taxon-based functional redundancy agreed with a general decrease in the number of classes contributing to the functional pathways with increasing trophic state (Fig. 4; Supplementary Table 2).

Table 1.

Correlation analysis of functional redundancy for various pathways against trophic state

Pathway Taxon-based Abundance-based
r p r p
Calvin cycle −0.49 5.806e−10 −0.15 0.064
rTCA cycle 0.27 0.0008 −0.02 0.787
WL −0.06 0.462 −0.05 0.538
Respiration −0.19 0.020 0.05 0.574
Methanogenesis −0.08 0.353 0.11 0.184
ANR −0.23 0.005 −0.38 3.115e−6
DNRA −0.12 0.131 0.08 0.357
Denitrification −0.63 <2.2e−16 0.27 0.001
Nitrification −0.05 0.552 0.43 7.648e−8
N Fixation −0.43 9.132e−8 0.18 0.032
P transport −0.27 0.001 −0.33 7.281e−5
ASR −0.29 0.0004 −0.36 1.052e−5
DSR −0.16 0.05 −0.06 0.443
SOX −0.19 0.021 0.12 0.155
Open in a new tab

Two sided Spearman’s Rank correlation analysis between trophic state (low oligotrophic, high supertrophic) and functional redundancy for specific functional pathways. SOX Sulfur oxidation, DSR Dissimilatory sulfate reduction, ASR Assimilatory sulfate reduction, N fixation Nitrogen fixation, DNRA Dissimilatory nitrate reduction to ammonia, ANR Assimilatory nitrate reduction, WL Wood-Lungdahl pathway. Significant results are depicted in bold font.

Fig. 4. Diversity of taxonomic classes per functional pathway across trophic states.

Fig. 4

Open in a new tab

Boxplot of the number of classes contributing to each pathway across trophic states. SOX Sulfur oxidation; DSR Dissimilatory sulfate reduction; ASR Assimilatory sulfate reduction; P transport Phosphorus transport; N fixation Nitrogen fixation; DNRA Dissimilatory nitrate reduction to ammonia; ANR Assimilatory nitrate reduction; WL Wood-Lungdahl pathway; rTCA reductive tricarboxylic acid. Boxplots were created from the number of taxonomic classes detected in 144 lakes. Boxes are colored by trophic state (47 Oligotrophic lakes; 30 Mesotrophic lakes; 37 Eutrophic lakes and 30 Supertrophic lakes). Source data are provided as a Source Data file. The boxplot shows the median (center) and the first quartile (bottom) and third quartile (top) of the data with the whiskers representing 1.5 x the interquartile range.

Abundance-based functional redundancy analysis indicated that there was a positive relationship between trophic state and denitrification and nitrification, while there was a significant negative correlation with assimilatory nitrate reduction, assimilatory sulfate reduction and P transport (Table 1). For the nitrification and denitrification pathways, Gammaproteobacteria were the dominant class, increasing in contribution to both the community abundance and functional abundance in higher trophic states (Fig. 5). Gammaproteobacteria contributed less to the functional abundance than to community abundance in both pathways. For nitrification, Verrucomicrobia, Myxococcia, and Nitrospiria in oligotrophic lakes contributed more to functional abundance than to their relative community abundance. The contribution of Nitrospiria to functional abundance compared to relative community abundance declined in eutrophic lakes. Vicinamibacteria showed a higher contribution to functional abundance in the denitrification pathway slightly increasing in lakes of higher trophic state. In assimilatory nitrate reduction, both Gammaproteobacteria and Alphaproteobacteria contributed more to functional abundance than to community abundance, with Gammaproteobacteria increasing and Alphaproteobacteria decreasing at higher trophic states (Supplementary Fig. 3). For ASR, Gammaproteobacteria contributed the most to both the community and the functioning of the pathway across the trophic states with Acidiobacteria, Myxococcia and Planctomycetia increasing their contribution to the function relative to the community in higher trophic states (Supplementary Fig. 3).

Fig. 5. Community abundance vs functional abundance of functional pathways across trophic states.

Fig. 5

Open in a new tab

The contribution of specific classes (top 5 for each trophic state are shown) to each functional pathway in terms of community abundance and functional abundance on log transformed axes. The data is derived from the mean values obtained from 144 lakes split by trophic state (47 Oligotrophic lakes; 30 Mesotrophic lakes; 37 Eutrophic lakes; and 30 Supertrophic lakes). The dotted line indicates the x:y ratio. Classes below the line contributed more to functional abundance than to community abundance. Points are colored according to taxonomic class. Pathways that showed significant correlations across trophic state for abundance-based functional redundancy are shown with non-significant pathways depicted in Supplementary Fig. 4. P transport Phosphorus transport, ASR Assimilatory sulfate reduction, ANR Assimilatory nitrate reduction, N fixation Nitrogen fixation. Source data are provided as a Source Data file.

For the methanogenesis pathway, Methanomicrobia was the dominant contributor across all trophic states for both functional and community abundance (Supplementary Fig. 2). In the DNRA pathway, Ignavibacteria exhibited the highest functional contribution despite low community abundance (Supplementary Fig. 4). Gammaproteobacteria contributed more to the functional abundance of the SOX pathway than to their relative community abundance (Supplementary Fig. 4).

Discussion

We conducted a national-scale investigation of the functional potential of lake sediment bacterial communities, examining trends in functional redundancy. Lake sediments contain both organisms that have settled from the water column and actively living sediment-dwelling microbes, providing an integrated record that is less sensitive to short-term temporal variation. Previous work by Pearman et al.7 demonstrated that a sediment bacterial trophic index could be developed, which correlates strongly with the Trophic Level Index, a measure of eutrophication in the water column. By sampling the surface sediment, we were able to investigate recent patterns in the lake's history and understand how functional potential varies across a trophic gradient.

Our results revealed a statistically significant, albeit weak, positive relationship between 16S rRNA taxonomic community differentiation at the ASV level and functional potential, as determined by annotated ORFs from metagenomics. The positive correlation suggests that changes in the taxonomy of the community contributes to variations in functional potential at a national scale, aligning with observations from other ecosystems such as prairie soils24 and in the surface seawaters of the Bay of Banyuls sur Mer (France) in the Mediterranean17. However, the weak correlation observed in the present study indicates that functional potential can be shaped by environmental drivers distinct from those influencing taxonomic composition, a finding which is consistent with ocean microbiomes20. Although our study did not directly measure ecosystem processes, variation in functional redundancy may influence the stability and resilience of key ecosystem functions under environmental stress, highlighting the ecological relevance of the patterns we observe.

Previous bacterial metabarcoding research on lake sediments of Aotearoa-New Zealand7,8 and lake planktonic and sediment microbes globally6,25 highlighted that agricultural land use intensification and subsequent eutrophication were a substantial driver of the taxonomic composition of lake microbial communities. A similar trend was noted in the metagenomics in the present study, with significant shifts in functional potential amongst lake sediments of different trophic states. Genes indicative of functional pathways showed differential responses to trophic state. For instance, a positive relationship was noted between the abundance (copies per million mapped reads) of indicative genes for methanogenesis and organic matter (Fig. 2B). Methane production in lakes is primarily undertaken in the sediments or oxgen deplete sections of the water column26,27 where anoxic conditions and high organic matter prevail. In eutrophic lakes, increased primary productivity often leads to greater deposition of organic matter in the sediments, creating conditions that favor methanogenesis28. This is exemplified in the current dataset with the increasing abundance of Methanomicrobia in more eutrophic lakes compared with its near absence in oligotrophic lake surface sediments. In oligotrophic lakes, where surface sediment is likely still aerobic, it is probable that methanogenesis is occurring in deeper sediment horizons where anaerobic conditions prevail. In contrast, indicative genes for phosphorus transport were negatively correlated with trophic state and high productivity grassland in the catchment but positively correlated with TN:TP. This suggests that the potential for active transport of phosphorus is lower in the eutrophic lakes, where phosphorus is more readily available, as indicated by the TN:TP correlation. The intensification of agriculture in catchments surrounding lakes has led to increased eutrophication and is a significant issue globally6,29–31. Historically, intensive dairy farming in Aotearoa-New Zealand has used relatively high levels of phosphorus fertilizer31,32 and leaching of fertilizers has potentially provided additional levels of phosphorus into lake sediments as indicated by the opposite trends between high productivity grassland and TN: TP for phosphorus transport.

Across the functional pathways investigated, there was, in general, a significant decrease in taxon-based functional redundancy as trophic state increased, except for the rTCA, which had a significant positive relationship (Table 1). This reduction in taxon-based functional redundancy agrees with the observation that lakes of higher trophic states, in general, have few taxonomic classes involved in the functional pathways compared to more oligotrophic lakes. Given that functional redundancy has been shown to enhance stability and resilience to perturbations13, this suggests that various metabolic pathways in eutrophic lake sediments may be more prone to disruption as they are reliant on fewer taxa, which could be lost during substantial perturbations. However, alternate stable state theory suggests that there is a switch from macrophyte dominated clear water states and de-vegetated turbid states driven by changes in the intensification of the catchment (and subsequent eutrophication)33,34, which is in contrast to the reduction in taxon-based redundancy observed in this study. However, while taxon-based functional redundancy is declining along the trophic gradient, the significant increase in abundance-based functional redundancy for key metabolic pathways such as nitrification, denitrification and nitrogen fixation was observed. Shifts in the composition of the sediment community, such as increases in Gammaproteobacteria, Desulfobacteria and Syntrophia and decreases in Alphaproteobacteria and Nitrospiria with increasing trophic state were also noted. The combination of increased functional redundancy in important metabolic pathways and shifts in the community composition could provide resilience in the community and reinforce an alternative stable state, thus increasing the challenges of lake restoration management to switch the system back to a clear water state. While this study focused on the functional redundancy of individual pathways with trophic state it may provide insights into the resilience of regime shifts in lake systems. This works shows oligotrophic lakes have a more diverse community contributing to the functional pathways (high taxon-based functional redundancy) creating resilience while more eutrophic lakes have a reduced taxonomic-based functional redundancy (lower diversity) but have a more even distribution between the community abundance and functional abundance for key metabolic pathways creating a higher abundance-based functional redundancy which along with shifts in the community composition enable the potential stabilization of an alternative stable state.

Abundance-based functional redundancy had a positive correlation with trophic state for nitrification, denitrification and nitrogen fixation, and a negative correlation for P transport, assimilatory nitrate reduction and assimilatory sulfate reduction. These differences are likely driven by variations in environmental conditions that selectively favor abundance changes of specific organisms, resulting in changes to the abundance-based functional abundancy. Both nitrification and denitrification pathways showed an increase in abundance-based functional redundancy with higher trophic state, while there was a reduction in taxon-based functional abundance (only significant for denitrification). This indicates that in these higher trophic states fewer classes are contributing to the functional pathway (reduction in taxon-based functional redundancy) while the abundance of the taxa undertaking the pathway is more evenly distributed (increase in abundance-based functional redundancy). This increased evenness in the contribution of taxa in the community to the functional pathway can be highlighted by the reduction in the deviation away from the 1:1 line between community abundance and functional abundance of Gammaproteobacteria for both pathways and Nitrospiria in the nitrification pathway. Gammaproteobacteria were the dominant component in the community for both pathways, and this contribution increases in lakes with a higher trophic state, which is in agreement with previous work indicating that Gammaproteobacteria are copiotrophs that are often prevalent in freshwater sediment environments7,35. Gammaproteobacteria often have preferences for organic-rich sediments, typical of eutrophic lakes affected by agricultural run-off, as shown in lakes of the Eastern Plain of China36. The contribution of Gammaproteobacteria to the functional abundance also increases, which is likely in part due to the reduction of Nitrospiria in more eutrophic lakes for the nitrification pathway. The contribution of Nitrospiria to the community declines markedly in the sediment communities with increasing trophic state. The presence of Nitrospiria in oligotrophic lakes is in agreement with other work on freshwater sediments where the phylum Nitrospirae was found in the surface sediments of Lake Sarnen, Switzerland, which has low ammonia and high oxygen concentrations37. Baumann et al.37 found potential evidence for Nitrospira to undertake comammox (the oxidation of ammonia to nitrate in a single cell) in an oligotrophic lake, which was absent in more eutrophic lakes. Further work, including the creation of metagenomic assembled genomes, could be undertaken to understand the different strategies of nitrification across trophic states in freshwater lakes of Aotearoa, New Zealand. Eutrophication can also impact the level of anoxia in the water column and sediments and these changes in the redox zonation may have a profound effect on the presence of nitrification and denitrification, which require the presence/absence of oxygen in the surface sediments38. For example, the reduction in the contribution of Nitrospiria to both the community and function in highly eutrophic lakes suggests increasingly anaerobic sediment conditions, with nitrification likely shifting into the oxygenated water column. Shifts of Nitrospiria from sediments to the water column have previously been reported in storm sewer systems39. The shifting of the redox zonation has the potential to affect the functional redundancy of the microbial community in the surface sediments as processes move to different environments (e.g., into the water column or deeper into the sediment). Future studies could investigate trends in functional redundancy across multiple environments within the lake (e.g., surface waters, bottom water, surface sediment) to get a deeper understanding on how functional redundancy is changing within lakes.

For phosphorus transport, there was a negative relationship between abundance-based functional redundancy and trophic state. This indicates that the functional abundance is deviating from the community abundance. This is evidenced by the fact that, despite Gammaproteobacteria increasing as a proportion of the total community, the contribution to phosphorus transport is declining. Conversely, the contribution of Anaerolineae to phosphorus transport increased compared to their community abundance. In freshwater sediments Chloroflexota, including Anaerolineae, have been shown to have preferences for eutrophic lakes and high concentrations of total phosphorus40 and have been shown to play roles in phosphorus metabolism in lake sediments41. This shows that in lakes experiencing eutrophication the functional potential of phosphorus transport is shifting amongst taxa. While Gammaproteobacteria are still the main contributors to phosphorus transport in nutrient-rich lake sediments, there is an increasing proportion of the functional potential being accounted for by taxa (Anaerolineae) that represent a small component of the community. As a higher proportion of the functional potential of a pathway is being accounted for by taxa that are at low abundance in the community, the higher the chances are that perturbations to the community could cause the disappearance of those taxa and have substantial impacts on the ability of a community to maintain that function.

This study demonstrates distinct trends in functional redundancy across a trophic gradient for multiple pathways, focusing on functional potential in surface sediments using metagenomics. Future research could employ metatranscriptomics to target living sediment microbes, revealing how the active members of the community respond across the trophic gradient. It should also be noted that the investigation was undertaken at the class level due to limitations in the taxonomic classification of ORFs. Therefore, there is a potential that at finer taxonomic levels, such as the species level, different patterns in functional redundancy might be observed across the trophic gradient. As the sequencing of microbial genomes increases, the ability to assign taxonomy to ORFs will increase, and future studies will have the potential to investigate functional redundancy at the species level. Similarly, this study was undertaken at the KEGG Orthologs level and investigations at different levels (e.g., ORFs or metabolic pathways) could give a different result with the finer the scale studied the less likely for redundancy to be present12.

Overall, our findings indicate that eutrophication drives a reduction of functional redundancy, especially taxon-based redundancy, with metabolic pathways increasingly reliant on fewer taxa to function. Because functional redundancy is a key determinant of ecosystem stability, its decline in eutrophic lake sediments suggests heightened vulnerability to both natural and anthropogenic disturbances. However, increases in abundance-based functional redundancy for key metabolic pathways (especially nitrogen-related) and shifts in the community composition could enable the creation and stabilization of alternative states as the increased abundance-based functional redundancy counteracts the reduction in taxon-based redundancy. Given the central role of lakes in global nutrient and carbon cycling, which can buffer against climate change impacts42 these changes are likely to have consequences that extend beyond individual lake systems. Shifts in the functional redundancy of metabolic pathways in lakes could cascade and influence connectivity and feedback across terrestrial–aquatic–atmospheric interfaces.

Methods

Sampling and extraction

Sediment samples used in this study were collected from 144 lakes (183 samples; some lakes had triplicate samples taken from the same point in the lake) across Aotearoa New Zealand between October 2018 and April 2021 (Fig. 1 and Supplementary Data 1). Lakes were accessed with the permission of regional councils, local iwi and landowners. The lakes were selected to be representative of the various natural and anthropogenic conditions experienced by lakes across Aotearoa New Zealand. The sampled lakes ranged from sea level (e.g., Lakes Moeraki, Paringa) to 1799 m (Alta Tarn). Lake depths varied from ~ 1 m (e.g., Runanga Lake, Tomahawk Lagoon) to greater than 400 m (e.g., Lakes Manapouri and Te Anau). In general, sampling was undertaken at the deepest parts of the lakes except for the lakes deeper than 100 m, where sampling was undertaken in shallower bays.

Samples of sediment were taken from the top 2 cm of the ponar grab and collected in 500 mL containers and shipped chilled to the laboratory. For metagenomics subsamples (~20 g) were taken and frozen (−20 °C) until processing. DNA was extracted using an Alkaline Buffer Extraction Protocol43 in dedicated UV-sterilized laboratories under sterile conditions. The VAHTS Universal Plus DNA Library Prep Kit was used for library preparation, and samples were sequenced on an Illumina (San Diego, CA, USA) NovaSeq 6000 machine by Azenta Life Science (Suzhou, China).

Lake characteristics are provided in Supplementary Data 1. Surface sediment samples were taken for sediment biogeochemistry analysis, with nutrient and elemental analysis detailed in Pearman et al.8. Briefly, total nitrogen and total organic carbon were analyzed using catalytic combustion and separated using a thermal conductivity detector. Inductively coupled plasma mass spectrometry was used to assess the elemental concentrations (see Supplementary Data 1). Catchment characteristics were determined using the Land Cover Database version 5 (LCDBv5, Landcare Research New Zealand Ltd., Lincoln, New Zealand). Trophic state was determined for each lake using the Sediment Bacterial Trophic Index (SBTI) as detailed by Pearman et al.7, The SBTI consists of numeric values, which were subsequently grouped into four trophic states with Oligotrophic: SBTI < 3; Mesotrophic: SBTI 3 – 4; Eutrophic: SBTI 4 – 5; Supertrophic: SBTI > 5.

Bioinformatics

Raw reads from 220 samples were used for the creation of the non-redundant gene database (only 183 samples, which had a trophic state, were used for the analysis of the data). The raw data is stored in the Aotearoa Genomic Data Repository under the accession https://doi.org/10.57748/3jdy-rc84. These were processed using Trimmomatic v 0.3944 using ILLUMINACLIP and the following options: LEADING:5 TRAILING:5 SLIDINGWINDOW:10:15 MINLEN:30. Cleaned reads were then assembled using Megahit v1.2.945 for each individual sample, and contigs shorter than 500 bp were removed with seqmagick v0.8.646. Using the software Prodigal v2.6.347 ORFs were predicted for each sample. The ORFs from each sample were concatenated together, and a mmseqs v14.7e28448 database was created. A non-redundant gene database was constructed, adapting the approach of Feng et al.49. This database was then clustered at 95% identity with 90% of the shorter gene being aligned (--cluster-mode 2 --min-seq-id 0.95 -c 0.9 --cov-mode 1) using linclust from mmseqs, creating a non-redundant gene database of New Zealand lake sediments and filtered to remove sequences shorter than 100 bp. Reference sequences for each cluster were retrieved using mmseqs and annotated using EggNOG-mapper v2.1.1250,51 using DIAMOND v2.1.8.16252.

For read mapping, the non-redundant gene database was filtered to those sequences that were annotated against the KEGG database and reads for each sample were annotated using bwa mem53 and then counts per million (CPM) were calculated. The CPM was calculated by first finding the reads per kilobase (RPK) and then dividing by the total RPK and scaling to a million reads.

Taxonomic classification of the raw reads to give an assessment of the community composition was undertaken by sylph v0.9.054 against the GTDB55 v232 database. Taxonomic classification of the ORFs was undertaken using mmseqs against the GDTB55 ORF database for bacteria using the default parameters. Taxonomical classification of the coding sequences against the NCBI non-redundant protein database, based on Xiong et al.56, although for this study we were interested in eukaryotic sequences, was undertaken using blastp within DIAMOND (max target seqs 10). The lowest common ancestor was found for each coding sequence based on the hits that had scores with greater than 50% similarity and were within 10% of the highest bitscore for each ORF. Where the blastp results were attributed to eukaryotes, they were added to the bacterial taxonomy from GDTB.

Statistical analysis

The script for analysis can be found at https://github.com/jkpearmanbioinf/Lake-Metagenomic-Redundancy with files to run the script available at https://doi.org/10.6084/m9.figshare.30881612.

Principal Component Analysis (PCA) was undertaken on the environmental dataset within R v4.5.157 to give an indication of the environmental gradients present. The environmental data (nutrients, catchment characteristics, depth and elevation) were scaled and centered using the R package vegan v2.7-558 with the PCA assessed and visualized using ggfortify v0.4.1459 and ggplot2 v3.3.660.

Multivariate analysis was undertaken at the KEGG Orthology level (KO) on a Hellinger-transformed data table with a distance matrix constructed using Bray-Curtis. Visualization was undertaken in ggplot after dimensionality reduction using t-distributed Stochastic Neighbor Embedding (t-SNE) in the Rtsne package v0.1761. Mantel correlations were undertaken using the functional data in the current paper and compared against a subset of the 16S rRNA metabarcoding at the ASV level data published in Pearman et al.8.

Indicative genes (KEGG Orthologs; KO numbers) were determined for specific pathways based on the literature (Supplementary Table 3) and ORFs that were annotated with these KO numbers were subset for further investigation. For lakes with triplicate samples, the mean copies per million mapped reads for the lake was calculated.

Correlation analysis was undertaken in R using cor.test with the Spearman rank algorithm, with p-value adjustments undertaken using the Benjamini-Hochberg equation. Heatmap plots were constructed based on ORFs that were classified at the class level. For the heatmap analysis, ORFs with only higher-level classifications (e.g., phylum) were removed from the dataset.

Taxon-based and abundance-based functional redundancy were determined using the equations of Fässler et al.16. For lakes with triplicate samples, mean values of functional redundancy were calculated before further analysis. For the abundance vector, the relative contribution of each taxonomic class (only genes classified at the class level were included) was calculated across the whole dataset, then subset for each functional pathway to the classes that could perform the function. The contribution of these taxa was then normalized, and the n_reference was determined as the total number of classes that could undertake the function across the whole dataset for comparison between samples, as suggested by Fässler et al.16. For the function vector, the counts per million of the class to the specific function were determined and normalized. Differences in functional redundancy amongst different functional pathways and within functional pathways across trophic state were tested using the Kruskall-Wallis test and pairwise comparisons undertaken using the Wilcox test in R. Spearman’s correlations were undertaken between the functional redundancy results and the SBTI7 to assess the effect of trophic state on functional redundancy.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information (170.1KB, pdf)
Supplementary Data 1 (114.5KB, xls)
Supplementary Data 2 (60KB, xls)
Reporting Summary (1.6MB, pdf)
Transparent Peer Review file (594.2KB, pdf)

Source data

Source Data (269.7KB, xlsx)

Acknowledgements

We thank all members of the Lakes380 team for field assistance and laboratory analysis (ourlakesourfuture.co.nz/lakes/). In particular, we acknowledge Lucy Thompson, Georgia Thomson-Laing (Cawthron) and Lizette Reyes (Earth Sciences New Zealand) for sample processing. We acknowledge the support of New Zealand regional authorities that provided data and permission to use it and assisted with access to the sampling sites: Northland Regional Council, Auckland Council, Waikato Regional Council, Bay of Plenty Regional Council, Hawkes Bay Regional Council, Taranaki District Council, Horizon Regional Council, Greater Wellington Regional Council, Marlborough District Council, Tasman District Council, West Coast Regional Council, Environment Canterbury, Otago Regional Council and Environment Southland. The authors thank iwi and landowners nationwide for their assistance with sampling, accessing sites and guidance throughout this work. The Department of Conservation is acknowledged for assistance with permitting.

Author contributions

J.K.P. and S.A.W. conceptualized the study. J.K.P., S.W., A.B.H.R., J.D.H., M.J.V., and S.A.W. undertook the field work while J.S. and J.K.K. performed the laboratory work. J.K.P. analyzed the data, and J.K.P. and S.A.W. wrote the original draft of the manuscript. All authors contributed to the review and editing of drafts.

Peer review

Peer review information

Nature Communications thanks Ramiro Logares, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This research was funded by the New Zealand Ministry of Business, Innovation and Employment Smart Idea project “Applying a functional evidence approach to prioritize lake restoration initiatives (CAWX2303)” and research programs “Our Lakes, Our Future (CAWX2305)” and “Our lakes’ health; past, present, future (C05X1707)”.

Data availability

Raw sequence data, is available upon application to the Aotearoa Genomic Data Repository under dataset ID AGDR00072 and can be found at https://data.agdr.org.nz/discovery/NZ-00072. Access to these data is fully available upon application via the Aotearoa Genomic Data Repository application link and is reviewed by the authors with response to the application expected within days. Raw data access is managed through established cultural and ethical protocols because the data may contain information considered taonga (treasured) and potentially sensitive to Māori (indigenous people of New Zealand) communities associated with these lakes and organisms. Processed data from this study used to run the below scripts can be found at https://doi.org/10.6084/m9.figshare.30881612. Source data are provided with this paper.

Code availability

The script for analysis can be found at https://github.com/jkpearmanbioinf/Lake-Metagenomic-Redundancy.

Competing interests

The authors declare no competing interests

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77595-w.

References

  • 1.Dornelas, M. et al. Assemblage time series reveal biodiversity change but not systematic loss. Science344, 296–299 (2014). [DOI] [PubMed] [Google Scholar]
  • 2.Philippot, L., Griffiths, B. S. & Langenheder, S. Microbial Community Resilience across Ecosystems and Multiple Disturbances. Microbiol. Mol. Biol. Rev. 85 10.1128/mmbr.00026-20 (2021). [DOI] [PMC free article] [PubMed]
  • 3.Adrian, R. et al. Lakes as sentinels of climate change. Limnol. Oceanogr.54, 2283–2297 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Dudgeon, D. Multiple threats imperil freshwater biodiversity in the Anthropocene. Curr. Biol.29, R960–R967 (2019). [DOI] [PubMed] [Google Scholar]
  • 5.Li, M. et al. The diversity and biogeography of bacterial communities in lake sediments across different climate zones. Environ. Res.263, 120028 (2024). [DOI] [PubMed] [Google Scholar]
  • 6.Kraemer, S. A. et al. A large-scale assessment of lakes reveals a pervasive signal of land use on bacterial communities. ISME J.14, 3011–3023 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Pearman, J. K. et al. A bacterial index to estimate lake trophic level: national scale validation. Sci. Total Environ.812, 152385 (2022). [DOI] [PubMed] [Google Scholar]
  • 8.Pearman, J. K. et al. Deterministic processes drive national-scale patterns in lake surface sediment bacteria and eukaryotic assemblage composition. Limnol. Oceanogr. 68, 1–295 (2022).
  • 9.Capo, E. et al. Tracking a century of changes in microbial eukaryotic diversity in lakes driven by nutrient enrichment and climate warming. Environ. Microbiol.19, 2873–2892 (2017). [DOI] [PubMed] [Google Scholar]
  • 10.Shen, M. et al. Trophic status is associated with community structure and metabolic potential of planktonic microbiota in plateau lakes. Front. Microbiol. 10, 2560 (2019). [DOI] [PMC free article] [PubMed]
  • 11.Ricotta, C. et al. Measuring the functional redundancy of biological communities: a quantitative guide. Methods Ecol. Evol.7, 1386–1395 (2016). [Google Scholar]
  • 12.Ramond, P., Galand, P. E. & Logares, R. Microbial functional diversity and redundancy: moving forward. FEMS Microbiol. Rev.49, fuae031 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Biggs, C. R. et al. Does functional redundancy affect ecological stability and resilience? A review and meta-analysis. Ecosphere11, e03184 (2020). [Google Scholar]
  • 14.Beauvais, M. et al. Functional redundancy of seasonal vitamin B12 biosynthesis pathways in coastal marine microbial communities. Environ. Microbiol.25, 3753–3770 (2023). [DOI] [PubMed] [Google Scholar]
  • 15.Galland, T. et al. Are redundancy indices redundant? An evaluation based on parameterized simulations. Ecol. Indic.116, 106488 (2020). [Google Scholar]
  • 16.Fässler, D., Heinken, A. & Hertel, J. Characterising functional redundancy in microbiome communities via relative entropy. Comput. Struct. Biotechnol. J.27, 1482–1497 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Galand, P. E., Pereira, O., Hochart, C., Auguet, J. C. & Debroas, D. A strong link between marine microbial community composition and function challenges the idea of functional redundancy. ISME J.12, 2470–2478 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Delgado-Baquerizo, M. et al. Lack of functional redundancy in the relationship between microbial diversity and ecosystem functioning. J. Ecol.104, 936–946 (2016). [Google Scholar]
  • 19.Louca, S. et al. Function and functional redundancy in microbial systems. Nat. Ecol. Evol.2, 936–943 (2018). [DOI] [PubMed] [Google Scholar]
  • 20.Louca, S., Parfrey, L. W. & Doebeli, M. Decoupling function and taxonomy in the global ocean microbiome. Science353, 1272–1277 (2016). [DOI] [PubMed] [Google Scholar]
  • 21.Ren, Z., Qu, X., Peng, W., Yu, Y. & Zhang, M. Functional properties of bacterial communities in water and sediment of the eutrophic river-lake system of Poyang Lake, China. PeerJ7, e7318 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Brasell, K. A. et al. Lake microbial communities are not resistant or resilient to repeated large-scale natural pulse disturbances. Mol. Ecol.30, 5137–5150 (2021). [DOI] [PubMed] [Google Scholar]
  • 23.Our Lakes, Our Future - enhancing lake health in Aotearoa https://ourlakesourfuture.co.nz/.
  • 24.Fierer, N. et al. Reconstructing the microbial diversity and function of pre-agricultural tallgrass prairie soils in the United States. Science342, 621–624 (2013). [DOI] [PubMed] [Google Scholar]
  • 25.Zeng, J. et al. Patterns and assembly processes of planktonic and sedimentary bacterial community differ along a trophic gradient in freshwater lakes. Ecol. Indic.106, 105491 (2019). [Google Scholar]
  • 26.D’Ambrosio, S. L. & Harrison, J. A. Methanogenesis exceeds CH4 consumption in eutrophic lake sediments. Limnol. Oceanogr. Lett.6, 173–181 (2021). [Google Scholar]
  • 27.Meier, D. et al. Hydrogen–independent CO2 reduction dominates methanogenesis in five temperate lakes that differ in trophic states. ISME Commun.4, ycae089 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.West, W. E., Creamer, K. P. & Jones, S. E. Productivity and depth regulate lake contributions to atmospheric methane. Limnol. Oceanogr.61, S51–S61 (2016). [Google Scholar]
  • 29.Gushulak, C. A. C. et al. Impacts of a century of land-use change on the eutrophication of large, shallow, prairie Lake Manitoba in relation to adjacent Lake Winnipeg (Manitoba, Canada). Freshw. Biol.69, 47–63 (2024). [Google Scholar]
  • 30.Keatley, B. E., Bennett, E. M., MacDonald, G. K., Taranu, Z. E. & Gregory-Eaves, I. Land-use legacies are important determinants of lake eutrophication in the anthropocene. PLOS ONE6, e15913 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Abell, J. M., Özkundakci, D., Hamilton, D. P. & Miller, S. D. Relationships between land use and nitrogen and phosphorus in New Zealand lakes. Mar. Freshw. Res62, 162–175 (2011). [Google Scholar]
  • 32.Quinn, J. M., Wilcock, R. J., Monaghan, R. M., McDowell, R. W. & Journeaux, P. R. Grassland farming and water quality in New Zealand. Tearmann Ir. J. Agric.-Environ. Res.7, 69–88 (2009). [Google Scholar]
  • 33.Schallenberg, M. & Sorrell, B. Regime shifts between clear and turbid water in New Zealand lakes: environmental correlates and implications for management and restoration. N. Z. J. Mar. Freshw. Res.43, 701–712 (2009). [Google Scholar]
  • 34.Özkundakci, D. & Lehmann, M. K. Lake resilience: concept, observation and management. N. Z. J. Mar. Freshw. Res.53, 481–488 (2019). [Google Scholar]
  • 35.Sauer, H. M., Hamilton, T. L., Anderson, R. E., Umbanhowar, C. E. Jr & Heathcote, A. J. Diversity and distribution of sediment bacteria across an ecological and trophic gradient. PLoS ONE17, e0258079 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Huang, W., Chen, X., Jiang, X. & Zheng, B. Characterization of sediment bacterial communities in plain lakes with different trophic statuses. MicrobiologyOpen6, e00503 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Baumann, K. et al. Microbial nitrogen transformation potential in sediments of two contrasting lakes is spatially structured but seasonally stable. mSphere7, e0101321 (2022). [DOI] [PMC free article] [PubMed]
  • 38.Cojean, A. N. Y. et al. Direct O2 control on the partitioning between denitrification and dissimilatory nitrate reduction to ammonium in lake sediments. Biogeosciences16, 4705–4718 (2019). [Google Scholar]
  • 39.Liu, C. et al. Migration and transformation of nitrogen in sediment–water system within storm sewers. J. Environ. Manag.287, 112355 (2021). [DOI] [PubMed] [Google Scholar]
  • 40.Wan, Y., Ruan, X., Zhang, Y. & Li, R. Illumina sequencing-based analysis of sediment bacteria community in different trophic status freshwater lakes. MicrobiologyOpen6, e00450 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Yang, W. et al. Community characteristics and functional gene response analysis of phosphorus-metabolizing bacteria in plateau saline lake sediments. Front. Environ. Sci. 10, 994104 (2022).
  • 42.Tranvik, L. J. et al. Lakes and reservoirs as regulators of carbon cycling and climate. Limnol. Oceanogr.54, 2298–2314 (2009). [Google Scholar]
  • 43.Thomson-Laing, G., Howarth, J. D., Vandergoes, M. J. & Wood, S. A. Optimised protocol for the extraction of fish DNA from freshwater sediments. Freshw. Biol.67, 1584–1603 (2022). [Google Scholar]
  • 44.Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinforma. Oxf. Engl.30, 2114–2120 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Li, D., Liu, C.-M., Luo, R., Sadakane, K. & Lam, T.-W. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinforma. Oxf. Engl.31, 1674–1676 (2015). [DOI] [PubMed] [Google Scholar]
  • 46.Yu, G. Seqmagick: sequence manipulation utilities. (CRAN, 2024).
  • 47.Hyatt, D. et al. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinforma.11, 119 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Steinegger, M. & Söding, J. MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets. Nat. Biotechnol.35, 1026–1028 (2017). [DOI] [PubMed] [Google Scholar]
  • 49.Feng, Y. et al. Metagenome-assembled genomes and gene catalog from the chicken gut microbiome aid in deciphering antibiotic resistomes. Commun. Biol.4, 1305 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Huerta-Cepas, J. et al. eggNOG 5.0: a hierarchical, functionally and phylogenetically annotated orthology resource based on 5090 organisms and 2502 viruses. Nucleic Acids Res.47, D309–D314 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Cantalapiedra, C. P., Hernández-Plaza, A., Letunic, I., Bork, P. & Huerta-Cepas, J. eggNOG-mapper v2: functional annotation, orthology assignments, and domain prediction at the metagenomic scale. Mol. Biol. Evol.38, 5825–5829 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Buchfink, B., Xie, C. & Huson, D. H. Fast and sensitive protein alignment using DIAMOND. Nat. Methods12, 59–60 (2015). [DOI] [PubMed] [Google Scholar]
  • 53.Vasimuddin, Md., Misra, S., Li, H. & Aluru, S. Efficient Architecture-Aware Acceleration of BWA-MEM for Multicore Systems. In Proc. IEEE International Parallel and Distributed Processing Symposium (IPDPS) 314–324 10.1109/IPDPS.2019.00041 (IEEE, 2019). [DOI]
  • 54.Shaw, J. & Yu, Y. W. Rapid species-level metagenome profiling and containment estimation with sylph. Nat. Biotechnol.43, 1348–1359 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Parks, D. H. et al. GTDB release 10: a complete and systematic taxonomy for 715,230 bacterial and 17,245 archaeal genomes. Nucleic Acids Res.54, D743–D754 (2026). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Xiong, X. et al. A catalog of microbial genes and metagenome-assembled genomes from the quail gut microbiome. Poult. Sci.102, 102931 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.R Core Team. R: A language and environment for statistical computing. (R Foundation for Statistical Computing (R Core Team, 2024).
  • 58.Oksanen, J. et al. The vegan package. Community Ecol. Package10, 631–637 (2007). [Google Scholar]
  • 59.Horikoshi, M. et al. Ggfortify: data visualization tools for statistical analysis results. (CRAN, 2022).
  • 60.Wickham, H. Ggplot2: Elegant Graphics for Data Analysis. (Springer, 2016).
  • 61.Krijthe, J. Rtsne: T-distributed stochastic neighbor embedding using a Barnes-Hut implementation. 0.17 10.32614/CRAN.package.Rtsne (CRAN, 2014). [DOI]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information (170.1KB, pdf)
Supplementary Data 1 (114.5KB, xls)
Supplementary Data 2 (60KB, xls)
Reporting Summary (1.6MB, pdf)
Transparent Peer Review file (594.2KB, pdf)
Source Data (269.7KB, xlsx)

Data Availability Statement

Raw sequence data, is available upon application to the Aotearoa Genomic Data Repository under dataset ID AGDR00072 and can be found at https://data.agdr.org.nz/discovery/NZ-00072. Access to these data is fully available upon application via the Aotearoa Genomic Data Repository application link and is reviewed by the authors with response to the application expected within days. Raw data access is managed through established cultural and ethical protocols because the data may contain information considered taonga (treasured) and potentially sensitive to Māori (indigenous people of New Zealand) communities associated with these lakes and organisms. Processed data from this study used to run the below scripts can be found at https://doi.org/10.6084/m9.figshare.30881612. Source data are provided with this paper.

The script for analysis can be found at https://github.com/jkpearmanbioinf/Lake-Metagenomic-Redundancy.

Articles from Nature Communications are provided here courtesy of Nature Publishing Group

RESOURCES