Spatial scaling of metagenomic diversity reveals ecological disruption in the gut microbiome of gout patients

Abstract

Gout, a painful inflammatory arthritis, is characterized by hyperuricemia and monosodium urate crystal deposition, with growing evidence linking its pathogenesis to gut microbiome dysbiosis. However, traditional diversity metrics fail to capture the complex spatial organization of microbial communities. This study addresses this gap by applying the novel metagenomic Diversity-Area Relationship (m-DAR) model to investigate scaling laws in the gout microbiome—quantifying how metagenomic diversity changes with the number of individuals sampled. Our analysis of gut microbiomes from gout patients and healthy controls revealed fundamental ecological disruptions. We found that gout microbiomes exhibited significantly altered scaling patterns: they showed greater inter-individual dissimilarity (higher z-values) at the level of rare genes (q = 0), but weaker scaling of dominant genes (q = 1–3) compared to healthy controls. Crucially, the maximal accrual diversity (MAD) was substantially lower in gout patients, indicating a severely constrained potential for total microbial gene diversity. Furthermore, profiling of metagenomic functional gene clusters (MFGCs) uncovered widespread functional perturbations, including increased diversity scaling for carbohydrate-active enzymes (CAZy) but decreased scaling in essential metabolic pathways (KEGG, KO). These results demonstrate that the gout gut microbiome is defined by a loss of ecological structure, featuring reduced homogeneity in dominant taxa, expanded rare biosphere variation, and an overall collapsed diversity capacity. This work introduces an ecological framework for characterizing dysbiosis in gout that complements traditional diversity metrics and may inform the development of microbiome-based therapeutic strategies. Further research is needed to translate these ecological patterns into clinical applications.

Introduction

Gout is a debilitating inflammatory arthritis caused by the deposition of monosodium urate (MSU) crystals in joints and tissues, a direct consequence of hyperuricemia. Its prevalence is higher in males and is increasingly associated with modern dietary patterns characterized by high intake of fructose, purines, and fats, but low in fiber—a combination known to disrupt gut microbial balance and promote hyperuricemia¹. Gout is not only a painful joint condition but also a significant risk factor for systemic comorbidities, including hypertension and cardiovascular disease, underscoring the critical need for effective management strategies.

The primary pathophysiological driver of gout is an imbalance in uric acid homeostasis, resulting from either its overproduction or underexcretion. Humans lack the enzyme uricase, which in other mammals converts uric acid into more soluble allantoin, leading to inherently higher serum uric acid (SUA) levels. While the kidneys excrete approximately two-thirds of the body’s uric acid, the intestinal tract clears the remaining third, a process where the gut microbiome plays a pivotal role. Compelling recent evidence shows that specific gut bacteria can metabolize uric acid, utilizing it to synthesize nucleotides for bacterial DNA replication, thereby directly reducing systemic uric acid pools^2–4. This positions the gut microbiome as a compelling therapeutic target for managing hyperuricemia and gout.

Emerging research consistently highlights a state of gut microbial dysbiosis as a key feature of gout pathogenesis^31–33. Comparative analyses reveal that the gut microbiota of gout patients is compositionally distinct from that of healthy individuals. A recurring finding is the depletion of beneficial, anti-inflammatory bacteria such as Faecalibacterium prausnitzii and Bifidobacterium pseudocatenulatum, which are crucial for producing the short-chain fatty acid (SCFA) butyrate. Concurrently, there is often an enrichment of potentially harmful taxa like Bacteroides caccae and Bacteroides xylanisolvens^5,6. This dysbiosis is linked to impaired uric acid metabolism and systemic inflammation²². The gut microbiome influences gout through several interconnected mechanisms: (1) by modulating host purine metabolism and reducing intestinal purine absorption⁷, (2) by enhancing uric acid excretion, as demonstrated by species like Akkermansia muciniphila that regulate uric acid transporters in the gut and kidneys⁸, and (3) by mitigating immune-inflammatory responses via bacterial metabolites, particularly butyrate, which can suppress histone deacetylase and alleviate acute gouty arthritis⁹.

Functional metagenomic profiling has further delineated the microbiome’s role, revealing that gout patients exhibit an enrichment of microbial genes involved in fructose and mannose metabolism and lipid A (a component of endotoxin) biosynthesis. In contrast, there is a marked reduction in the abundance of genes responsible for urate degradation and the production of SCFAs⁵. SCFAs, including butyrate, acetate, and propionate, exert anti-inflammatory effects and are known to promote renal uric acid excretion,their depletion likely contributes significantly to the inflammatory milieu and hyperuricemia in gout. Furthermore, the gut microbiome of gout patients shows a pattern of enriched carbohydrate metabolism but diminished purine metabolism—a profile that is conversely observed in patients effectively treated with Febuxostat, underscoring the gut microbiota’s central role in the metabolic dysregulation of gout¹⁰.

Therapeutically, strategies aimed at correcting this dysbiosis—such as probiotics, prebiotics, and dietary interventions like the Mediterranean diet—show promise in modulating the microbiota, reducing SUA, and improving clinical outcomes^11,12. However, a fundamental question remains: is there a consistent, core gout microbiome, or is the dysbiosis highly individualized? The answer is crucial for designing effective, broad-spectrum versus personalized interventions.

To address this, we apply the Diversity-Area Relationship (DAR) model to metagenomic data from gout patients. The DAR model, an extension of the Species-Area Relationship, allows us to partition metagenomic diversity—encompassing both metagenomic genes (MGs) and functional clusters (MFGCs)—across spatial scales^13–16. Unlike standard diversity metrics, DAR explicitly quantifies how diversity is structured within individuals versus among individuals^{21,24–28,30,31}. This provides a powerful lens to determine whether gout is characterized by a loss of a common, healthy core microbiome or by a highly heterogeneous and personalized dysbiotic state. Our objectives are threefold: (1) to characterize the spatial scaling of metagenomic diversity in gout, (2) to identify key taxa and functional pathways driving these diversity patterns, and (3) to evaluate the potential for generalized versus personalized microbiome-based therapies. By applying the DAR framework, we aim to uncover novel ecological principles governing the gout microbiome and provide a refined foundation for future therapeutic development.

Materials and methods

Datasets

The datasets used in this research originate from a study by Chu et al⁵., which recruited 102 male acute gout patients (SUA, 543.1 ± 128.4 μmol/L) and 86 age-matched male healthy controls (SUA, 358.6 ± 52.3 μmol/L) for the study between May 2016and September 2018. They analyzed 307 fecal samples from 102 gout patients and 86 healthy individuals. Additionally, 70, 40, and 9 samples were collected from gout patients in the treatment follow-up cohort (referred to as the ‘discovery cohort’) at three timepoints post-baseline (weeks 2, 4, and 24). Further details regarding the cohorts, including healthy controls and patients, are available in the original publication by Chu et al⁵. To ensure balanced sample sizes for robust analysis, 84 healthy control samples, 99 gout patient samples, and 69 samples from the two-week discovery cohort were selected. The metagenomic shotgun sequencing data were retrieved from the CNGB Nucleotide Sequence Archive (CNSA) under accession code CNP0000284. Analysis of metagenomic sequencing, including the quantification of metagenomic genes (MGs) and metagenomic functional gene clusters (MFGCs), was performed using the “EasyMetagenome” pipeline (Liu et al., 2021, 2023).

The therapeutic intervention in the original study consisted of standardized urate-lowering therapy (benzbromarone, allopurinol, or febuxostat) combined with anti-inflammatory drugs (colchicine, celecoxib, etoricoxib, betamethasone, or voltaren), along with dietary counseling (low-purine diet, increased water intake, alcohol restriction). Fecal samples were collected at baseline and at three follow-up timepoints: week 2 (n = 70), week 4 (n = 40), and week 24 (n = 9) post-treatment initiation. For the present m-DAR analysis, we selected the week 2 timepoint (Gout_2W, n = 69 after quality filtering) as the primary follow-up group because it represents the largest and most statistically robust follow-up cohort. The week 4 and week 24 timepoints had substantially smaller sample sizes, which would introduce imbalance and reduce statistical power for m-DAR modeling.

For deeper functional insights into the genes, sequences were aligned against three key databases: the Carbohydrate-Active EnZymes database (CAZy), the Gene Ontology (GO) database, and the Kyoto Encyclopedia of Genes and Genomes (KEGG), which includes KEGG Orthology (KO) and Pathway modules. The CAZy database categorizes enzymes involved in breaking down, modifying, or creating glycosidic bonds, providing a detailed classification of their structural and functional domains. The KEGG database integrates molecular-level data from genome sequencing and high-throughput experiments to elucidate biological systems at various levels, from cellular to ecosystem scales. Meanwhile, the GO knowledgebase, the largest repository of gene function information, offers standardized annotations for genes across biological processes, molecular functions, and cellular components. KEGG was used as a reference resource for gene and protein annotation, and is cited here according to the KEGG guidelines. No KEGG pathway diagrams or images were reproduced. KEGG was used solely as a reference database for functional annotation and is cited accordingly (Kanehisa et al. 2000; 2016).

Ethics approval and consent to participate

All methods were carried out in accordance with relevant guidelines and regulations. The present study is a secondary analysis of publicly available, de-identified human gut metagenomic datasets; no new human samples were collected and no additional experiments involving human participants were performed by the authors. Ethical approval and informed consent for participant recruitment, sample collection and sequencing in the original gout cohorts were obtained by the investigators of the primary study and are reported in detail by Chu et al⁵.. Because only de-identified data already in the public domain were used, no additional institutional review board approval or informed consent was required for this work.

Estimation of metagenome diversity

Analogous to the estimation of species diversity of macrobial assemblages, Ma and Li¹⁷ used Hill numbers ^36–39 to define different measures of diversity of MGs and MFGCs:

1

In Eq. 1, G is the number of MGs or MFGCs, p_i is the relative abundance of the i-th MG or MFGC, and q is the order number of diversity. When q = 0, ⁰D = G, the number (“richness”) of MGs or MFGCs. When q = 1, ¹D is MG or MFGC diversity weighted proportionately by gene or functional gene cluster frequency. When q = 2, ²D is MG or MFGC diversity weighted by dominant (more abundant) genes or functional gene clusters. When q = 3, ³D is weighted more heavily than ²D by dominant genes or functional gene clusters¹⁷.

Measures of metagenome diversity based on MGs use single genes, but metagenome diversity of MFGCs can differ depending on whether clusters are based on metabolic functions (KEGG) or protein functions^34,35. Ma and Li¹⁷ also distinguished two types of MFGC diversity, depending on how individual gene abundance information is used in defining the functional clusters. Type I MFGCs ignore the abundance of individual genes and only count the number of genes in a cluster (analogous to incidence data in macrobial species diversity calculations. In contrast, Type II MFGCs depend on both the number of genes and their relative abundances. .

Fitting m-DAR models and constructing m-DAR diversity profiles

Following Ma¹³, who extended the classic species-area relationship (SAR) to a diversity-area relationship (DAR), Ma & Ellison¹⁶ used a basic power law (PL) model to define the metagenomic diversity-area relationship (m-DAR):

2

In Eq. 2, ^qD is metagenome diversity (Eq. 1) of order q; A (“area”) is the number of subjects whose metagenome diversity are sampled; and c and z are fitted parameters. As applied to microbial metagenomes, c can be thought of as the estimated diversity of any single individual, whereas the diversity scaling parameter z is the rate of increase in metagenome diversity with increasing number of individuals sampled.

We follow Plotkin et al. (2000) and Ulrich and Buszko (2003)^40,41 in modifying Eq. 2 to include a third parameter, d:

3

In this “power law with exponential cutoff” (PLEC) model, d < 0 and exp(dA) eventually overwhelms the exponential function at very large values of A, leading to an asymptotic value of ^qD. Using this exponential decay term makes sense because there are a finite number of people and thus a finite diversity of metagenomes.

We use log-transformed versions of Eqs. 2 and 3

4

5

to estimate the parameters of the models because their computation is simpler; z is scale-invariant in Eq. 4; and the ecological interpretation of z as a scaling parameter is preserved in Eq. 5. On a log–log plot, z is the slope of the linearized functions. Fitting of Eqs. 4 and 5 to the data was evaluated using the linear correlation coefficients (r) and associated P values.

A key distinction exists between the human microbiome and microbial communities in natural ecosystems. In natural biomes, such as those found in soil, water, or plant-animal systems, spatial organization and environmental gradients (e.g., soil layers, altitude, or temperature) often provide a clear structure for microbial assemblages. In contrast, human microbiomes lack such inherent spatial ordering, as human hosts do not follow a natural gradient or hierarchy. To address this challenge, we developed a methodological approach to analyze microbial diversity in the absence of a predefined spatial framework.

First, we generated all possible permutations of sample subject orderings to account for the lack of natural spatial structure. From these permutations, we randomly selected 100 orderings for metagenomic genes (MGs) and metagenomic functional gene clusters (MFGCs). For each selected ordering, we fitted the metagenomic diversity-area relationship (m-DAR) models (Eqs. 4, 5). We then filtered out poorly fitting models (those with P > 0.05) and, for the power law with exponential cutoff (PLEC) models (Eq. 5), excluded any models with biologically implausible parameters (e.g., A_max < 0). Finally, we calculated the average values of the estimated parameters from the remaining models to derive the overall m-DAR model parameters.

The relationship between the diversity order (q) and the scaling parameter (z) from Eq. 2 defines the m-DAR profile. This profile is analogous to the diversity scaling observed in species diversity studies¹³ and provides a framework for understanding how microbial diversity scales across different sample groupings in the absence of natural spatial ordering.

Metagenomic maximal accrual diversity of metagenome

Ma¹³ derived the maximal accrual diversity (MAD) in a cohort or population based on the DAR-PLEC model (Eq. 3) as:

6

for which the number of individuals (A_max) reaching the maximum diversity (D_max) is estimated as:

7

We then define the m-MAD profile as the set of D_max values corresponding to different diversity orders q. ^qD_maxcan be interpreted as a proxy for the so-called potential (“dark”) diversity: species (OTUs) or metagenomes (MG or MFGC) that are absent locally but present in regional or global species pools (^16,18,42,43).

Pair-wise diversity overlap

Given the assumption that the “areas” corresponding to different sampled individuals are approximately equal, the scaling parameter (z) from the basic metagenomic diversity-area relationship (m-DAR) model (Eq. 2 or 4) can be utilized to estimate the pair-wise diversity overlap (PDO). The PDO, denoted as g, quantifies the proportion of new diversity introduced by a second individual relative to the first. It is defined as:

8

where z is the scaling parameter from the basic m-DAR model (Eq. 2 or 4). The value of g ranges from 0 to 1, where:

g = 0 indicates no overlap between the two individuals (i.e., z = 1), meaning their microbial communities are entirely distinct.

g = 1 indicates complete overlap (i.e., z = 0), meaning their microbial communities are identical.

The m-PDO profile is the collection of g values across different diversity orders (q). This profile provides an approximation of the similarity between pairs of human metagenomes, offering insights into the degree of shared microbial diversity between individuals.

Results

Metagenomic diversity-area relationships (m-DAR)

The analysis of metagenomic diversity-area relationships (m-DAR) using both the basic power law (PL) model (Eqs. 2, 4) and the power law with exponential cutoff (PLEC) model (Eqs. 3, 5) revealed significant insights into the scaling of microbiome diversity across different study groups. Both models demonstrated a satisfactory fit to the microbiome data (P < 0.05; Table S1-S5, Fig. 1).

Fig. 1.

Representative log–log plots of the diversity-area relationship (DAR) for metagenomic genes (MGs) and each functional database at q = 0. The x-axis shows log(n), where n is the number of samples (individuals); the y-axis shows log(D), where D is the alpha diversity (q = 0). Each panel corresponds to a specific group (Healthy, Gout, or Gout_2W) and a specific data type (MGs, CAZy, GO, KO, or KEGG Pathway). For each combination, one representative permutation out of the 100 random orderings is shown. The blue line represents the fitted power law (PL) model, and the green line represents the fitted power law with exponential cutoff (PLEC) model. Corresponding goodness-of-fit statistics (R and p-values) for all permutations are reported in Tables 1 and S1–S5.

At the metagenomic genes levels (MGs), the scaling parameter z showed a clear dissociation across diversity orders (Table 1; Fig. 2). At q = 0 (rare genes), gout patients had significantly higher z than controls (0.455 vs. 0.434, p < 0.001), indicating greater inter-individual variability in the rare gene repertoire. In contrast, at q ≥ 1 (dominant genes), gout patients exhibited lower z (e.g., at q = 2: 0.269 vs. 0.306, p = 0.012) and higher pairwise diversity overlap (g), meaning that the most abundant genes are more similar across gout patients. Thus, gout is characterized by a shared core of altered dominant taxa together with a highly personalized rare biosphere.

Table 1.

Summarizes the differences in DAR model parameters between the control and gout cohorts (summarized from Table S1-S6).

Database	Comparison	q = 0	q = 1	q = 2	q = 3	Interpretation
Metagenomic Genes
z	Gout vs. Control	↑ (***)	NS	↓ (*)	↓ (***)	Rare genes more variable; dominant genes more shared in gout
Dmax	Gout vs. Control	NS	↓ (***)	↓ (***)	↓ (***)	Reduced overall diversity capacity
CAZy
z	Gout vs. Control	NS	NS	NS	NS	No consistent difference in scaling
Dmax	Gout vs. Control	NS	NS	↓ (***)	↓ (***)	Reduced maximal diversity of rare CAZy functions
GO
z	Gout vs. Control	↓ (**)	NS	NS	NS	Rare GO terms less variable in gout
Dmax	Gout vs. Control	↓ (***)	NS	↓ (***)	↓ (***)	Reduced maximal diversity for rare GO terms
KO
z	Gout vs. Control	↑ (NS)	NS	NS	NS	Trend toward higher rare-KO variability
Dmax	Gout vs. Control	↓ (***)	↓ (***)	↓ (***)	↓ (***)	Consistent reduction across all orders
KEGG Pathway
z	Gout vs. Control	↓ (***)	NS	NS	NS	Rare pathways less variable in gout
Dmax	Gout vs. Control	NS	↓ (***)	↓ (***)	↓ (***)	Reduced maximal diversity for dominant pathways

*Arrow direction indicates the value in gout relative to controls. Significance: (***) p < 0.001, (**) p < 0.01, (*) p < 0.05, NS = not significant.

Fig. 2.

Summary of m-DAR parameters for metagenomic genes (MGs) across diversity orders and statistical comparisons between groups. Violin plots combined with box plots display the distribution of the scaling parameter z, pairwise diversity overlap g, and maximal accrual diversity D_max for each diversity order (q = 0, 1, 2, 3). Each distribution is derived from the 100 permutation runs for the respective parameter. Results are shown for the three study groups: Healthy (blue), Gout (red), and Gout_2W (green). Pairwise group comparisons that reached statistical significance (Wilcoxon rank-sum test, p < 0.05) are marked with asterisks (* p < 0.05, ** p < 0.01, *** p < 0.001); non-significant comparisons are marked as “NS”. The plots integrate the parameter estimates from Table S1 and the permutation test results from Supplementary Table S6.

The pair-wise diversity overlap (PDO) profile exhibited an inverse pattern compared to the m-DAR profile (Table S1). This is because the z parameter in the m-DAR profile quantifies the dissimilarity of neighboring individuals, whereas the g parameter in the PDO profile quantifies the overlap or similarity between individuals. This inverse relationship highlights the complementary nature of these two metrics in describing microbiome diversity. Additionally, the z-value decreases as the diversity order increases from q = 0 to q = 3, while the pairwise diversity overlap (PDO) increases (Table S1) in both healthy and gout-affected microbiome. This indicates that gout microbiome metagenomic genes are more similarity in common or dominate gene diversity (q = 1, 2, 3), but dissimilarity in rare gene diversity (q = 0).

The PLEC model indicated the existence of an asymptote for MG diversity, as evidenced by the negative d values and the corresponding D_max ​values presented in Table S1. Additionally, the model estimated the number of individual subjects (A_max​) required to reach this asymptote. The maximal accrual diversity (MAD) of MGs in the gut metagenome was 4.6 × 10⁶ for the total samples at q = 0, representing the estimated gene richness for this cohort. Maximal accrual diversity (D_max) was significantly reduced in gout for all dominant orders (p < 0.001), whereas no significant difference was observed at q = 0 (Table S6). For example, at q = 1, the MAD for the Control group was 393,429.1, compared to 291,143.8 and 268,791.0 for the Gout and Gout_2W groups, respectively. This suggests that the Control group had a higher potential for gene diversity accumulation compared to the gout-affected groups.

Metagenome functional gene cluster (MFGC) diversity-area relationships

Functional annotation against CAZy, GO, KO, and KEGG databases provided complementary insights (Tables S2–S5; Figures S1–S4).

For carbohydrate‑active enzymes (CAZy) (Table S2, Figure S1), gout patients showed higher z and D_max at q = 0 compared to controls (0.026 vs. 0.025 for z, not significant; D_max 399.0 vs. 394.9, p < 0.001). This suggests an expanded and more variable rare repertoire of carbohydrate metabolism, consistent with enrichment of Bacteroides and Prevotella species and known upregulation of fructose/mannose metabolism in gout.

For KO (KEGG Orthology) (Table S4, Figure S3), a striking dissociation mirrored the taxonomic pattern: at q = 0, gout patients had higher z (0.174 vs. 0.163, p = 0.053, trend) and higher D_max (15393.8 vs. 13368.3, p < 0.001); at q ≥ 1, gout patients showed lower z and lower D_max (e.g., at q = 1: D_max 2189.9 vs. 2291.2, p < 0.001). This indicates that rare functional orthologs are highly personalized, whereas the dominant core functions (e.g., urate degradation, SCFA synthesis) are consistently depleted across patients.

For KEGG pathways (Table S5, Figure S4), at q = 0 gout patients had lower z (0.015 vs. 0.028, p < 0.001) and similar D_max, indicating that the rare pathway repertoire is less variable in gout. This suggests that the specific metabolic pathways altered (purine metabolism, LPS biosynthesis) are consistently dysregulated, leaving little room for personalized variation. The GO database showed only minor differences (Table S3, Figure S2), with controls having slightly higher z at q = 0 (0.213 vs. 0.186, p = 0.001), reflecting that broad functional categories are less perturbed than more specific annotations.

Collectively, the functional m-DAR patterns demonstrate that gout dysbiosis involves both a shared core of disrupted dominant functions and an individualized rare functional landscape.

Additionally, as described in above section about the metagenomic gene (MG) scaling parameter, the z-value of MFGC decreases as the diversity order increases from q = 0 to q = 3, while the pairwise diversity overlap (PDO) increases in both healthy and gout-affected microbiome. This indicates that gout microbiome MFGC are more similarity in common or dominate MFGC diversity (q = 1, 2, 3), but dissimilarity in rare MFGC diversity (q = 0).

Discussion

This study applied the m-DAR framework to characterize the ecological scaling properties of the gut microbiome in gout. The results reveal that gout‑associated dysbiosis is not merely a loss of diversity but a structured disruption that separates rare from dominant components. The key finding is a consistent dissociation: at the rare level (q = 0), gout patients show increased inter‑individual variability (higher z), whereas at the dominant level (q ≥ 1), they show reduced variability and lower maximal accrual diversity (lower z and D_max). This pattern is observed for both taxonomic (MGs) and detailed functional (KO) profiles, and it provides a unified answer to the central research question.

Core versus individualized dysbiosis

The evidence supports a dual nature. The shared core is evidenced by lower z and higher PDO at q ≥ 1, together with reduced D_max. This shared core offers a target for broad-spectrum interventions. In contrast, the individualized rare biosphere is reflected by higher z and lower PDO at q = 0, meaning each patient harbors a unique set of low-abundance genes and functions. This personalization may explain heterogeneity in disease progression, treatment response, and comorbidities, and it implies that precision microbiome-based therapies may be necessary. These findings align with previous studies linking gout to gut dysbiosis, characterized by reduced microbial diversity and imbalances in key bacterial taxa^5,19.

Functional database-specific patterns

The contrasting patterns between CAZy (expanded carbohydrate metabolism) and KEGG pathways (contracted core metabolism) highlight that different functional compartments respond differently to the gout milieu. The expansion of CAZy diversity likely reflects adaptation to a high-carbohydrate, low-fiber diet, while the contraction of core metabolic pathways (e.g., SCFA synthesis, urate degradation) directly contributes to hyperuricemia and inflammation^23,29.

The complete z(q) profile provides a comprehensive picture of microbiome ecological structure. The pattern observed in gout—higher z at q = 0 combined with lower z at q ≥ 1 compared to controls—can be interpreted through the following unified framework: High z at q = 0 indicates greater inter-individual variation in rare species (q = 0 treats all genes equally, weighting rare and common equally), suggesting that the rare biosphere is highly individualized in gout patients. Low z at q ≥ 1 indicates reduced inter-individual variation in dominant species (q ≥ 1 weights genes by their abundance), suggesting that gout patients share a common set of altered dominant taxa. This pattern reveals that gout dysbiosis has dual characteristics: (1) a disrupted but consistent core of dominant taxa that is shared across patients, and (2) extensive variability in the rare biosphere that is highly personalized. This interpretation resolves the apparent contradiction in z-value patterns and provides a more nuanced understanding of gout-associated dysbiosis than traditional diversity metrics alone.

The power law with exponential cutoff (PLEC) model revealed an asymptote in metagenomic gene (MG) diversity, with the Control group exhibiting higher maximal accrual diversity (MAD) values across all diversity orders compared to the gout-affected groups. This indicates that healthy individuals have a greater potential for gene diversity accumulation, further highlighting the diminished microbial richness in gout patients. The negative d-values in the PLEC model suggest that MG diversity approaches a saturation point, with the Control group reaching higher saturation levels than the gout groups. This finding is consistent with the hypothesis that gout-associated dysbiosis limits the gut microbiome’s capacity to sustain diverse microbial communities.

The observed reductions in microbiome diversity and functional gene scaling in gout patients align with previous findings linking gut dysbiosis to abnormal uric acid metabolism, systemic inflammation, and metabolic dysfunction^5,19. The depletion of beneficial bacteria, such as SCFA-producing taxa (e.g., Bifidobacterium and Lactobacillus), and the enrichment of harmful bacteria (e.g., Prevotella copri and Collinsella aerofaciens) may exacerbate gout symptoms by impairing uric acid homeostasis and promoting inflammation.

The partial recovery of microbial diversity in the Gout_2W group following a two-week therapeutic intervention suggests that targeted interventions, such as probiotics or dietary modifications, may help mitigate gout-associated dysbiosis. However, the incomplete restoration of diversity highlights the need for longer-term studies to evaluate the sustainability and efficacy of such interventions. Promisingly, recent studies have demonstrated the potential of microbiome-based therapies, including probiotics and traditional Chinese medicine (TCM), to modulate the gut microbiome and improve clinical outcomes in gout patients^11,20.

A central question of this study was whether gout-associated dysbiosis reflects a consistent, shared microbial signature or a highly individualized disruption. The m-DAR framework, by partitioning diversity scaling across rare and dominant components, allows us to address this question directly. At q ≥ 1 (dominant genes and functions), gout patients exhibited lower z-values and higher pairwise diversity overlap (PDO) compared to healthy controls, alongside reduced maximal accrual diversity (Dmax). These patterns indicate that the most abundant taxa and functions—such as the depletion of butyrate-producing species and enrichment of Bacteroides and Prevotella, as well as the loss of urate-degrading pathways—are consistently altered across patients, defining a shared dysbiotic core. In contrast, at q = 0 (rare components), gout patients showed higher z-values and lower PDO, revealing that the rare microbial repertoire is highly personalized, with each patient harboring a unique set of low-abundance taxa and functions. Thus, gout dysbiosis is dual in nature: a common, targetable core coexists with extensive inter-individual variation in the rare biosphere. The existence of a shared core suggests that broad-spectrum interventions (e.g., probiotics containing Faecalibacterium, prebiotics to enhance short-chain fatty acid production, or dietary modulation to suppress Bacteroides) could be effective across many patients. Simultaneously, the personalized rare biosphere implies that precision microbiome-based therapies, tailored to each patient’s unique microbial profile, may be necessary to fully restore gut homeostasis. A combined strategy targeting both the shared core and the individualized rare components therefore represents a promising path forward for microbiome-based management of gout.

Limitations

Several limitations must be acknowledged. The cross-sectional design prevents causal inference; longitudinal studies are needed to determine whether dysbiosis precedes or follows gout. The study cohort was exclusively male, which controlled for sex but limits generalizability to females. The small sample sizes at weeks 4 and 24 precluded temporal trajectory analysis. Direct comparison with other diseases (e.g., IBD, diabetes, obesity) is hampered by differences in sequencing depth, sample size, and protocols. Future harmonized multi-disease studies are required to contextualize the severity of gout-associated dysbiosis. Finally, while we used 100 permutations and verified convergence, the permutation approach averages over all possible orderings and may not capture biologically meaningful spatial structures that could exist.

Finally, direct comparison of m-DAR parameters between our study and other disease cohorts (e.g., inflammatory bowel disease, type 2 diabetes, obesity) is hampered by differences in sample size, sequencing depth, and experimental protocols. Future studies employing standardized pipelines across multiple disease cohorts are needed to assess the relative severity of gout‑associated dysbiosis and to prioritize microbiome‑based therapeutic development.

Conclusion

This study highlights the significant alterations in gut microbiome diversity and functional gene scaling associated with gout, providing a novel framework for understanding the spatial dynamics of microbial communities in this disease. The application of the DAR model offers valuable insights into the scaling patterns of metagenomic diversity and underscores the potential of microbiome-based interventions for gout management. By restoring microbial balance and enhancing metabolic homeostasis, such interventions could pave the way for more effective and personalized treatment strategies. Future research should focus on elucidating the causal relationships between dysbiosis and gout, as well as exploring the therapeutic potential of targeted microbiome modulation.

Author contributions

JSZ and HJC conceived the study, QYT and HJC performed the data analysis, JSZ, QYT, HJC, LLW and WS wrote and revised the manuscript. All authors approved the submission.

Funding

This work was supported by the second batch of Provincial Science and Technology Plan (Basic Research Plan-Natural Science Foundation) Project of Gansu Province in 2024 (No: 24JRRJ001), Yunnan Provincial Basic Research Project (202301AT070380).

Data availability

The shotgun metagenomic sequencing datasets re-analysed in this study are publicly available in the CNGB Nucleotide Sequence Archive (CNSA) under accession code CNP0000284 (Chu et al., 2021), accessible at [https://db.cngb.org/data_resources/?query=CNP0000284] (https:/db.cngb.org/data_resources/?query = CNP0000284) **. Any additional processed data tables and analysis scripts generated during the current study are available from the corresponding author on reasonable request.**

Declaration

Competing interests

The authors declare no competing interests.

Ethics

This study is a secondary analysis of fully de-identified metagenomic sequencing data from a previously published cohort⁵. No new samples were collected and no additional clinical information was obtained by the authors. According to the policies of the Ethics Committee of the People’s Hospital of Dingxi City and the Second Affiliated Hospital of Lanzhou University Medical School, the need to obtain ethical approval for this secondary analysis of anonymised public data was waived. The need to obtain informed consent from the original participants was likewise waived by the Ethics Committee of the People’s Hospital of Dingxi City and the Second Affiliated Hospital of Lanzhou University Medical School.