Advancing the Exercise-Microbiome Axis: A Methodological and Bioinformatic Roadmap from Short-Read Standards to Long-Read Frontiers and Multi-Omics Integration

So, Byunghun; Jang, Junho; Park, Jinhan; Park, Young-Min; Kwak, Dongmin; Kim, KyeongJin; Lee, Handol; Kang, Chounghun; Byunghun So; Junho Jang; Jinhan Park; Young-Min Park; Dongmin Kwak; KyeongJin Kim; Handol Lee; Chounghun Kang

doi:10.15857/ksep.2025.00570

Exerc Sci > Volume 34(4); 2025 > Article

So, Jang, Park, Park, Kwak, Kim, Lee, and Kang: Advancing the Exercise-Microbiome Axis: A Methodological and Bioinformatic Roadmap from Short-Read Standards to Long-Read Frontiers and Multi-Omics Integration

Review Article

Exerc Sci. 2025; 34(4): 406-421.

Published online: Nov 28, 2025

DOI: https://doi.org/10.15857/ksep.2025.00570

Advancing the Exercise-Microbiome Axis: A Methodological and Bioinformatic Roadmap from Short-Read Standards to Long-Read Frontiers and Multi-Omics Integration

Byunghun So MS¹

, Junho Jang MS¹

, Jinhan Park MS¹

, Young-Min Park PhD³

, Dongmin Kwak PhD⁴

, KyeongJin Kim PhD⁵

, Handol Lee PhD⁶

, Chounghun Kang PhD^1,^2,

¹Department of Health and Exercise Science, College of Graduate School, Inha University, Incheon, Korea

²Department of Physical Education, College of Education, Inha University, Incheon, Korea

³Division of Health and Kinesiology, Incheon National University, Incheon, Korea

⁴Department of Sports Science, Hanyang University, Ansan, Korea

⁵Department of Biomedical Sciences, College of Medicine, Inha University, Incheon, Korea

⁶Department of Environmental Engineering, Inha University, Incheon, Korea

Corresponding author: Chounghun Kang Tel +82-32-860-7880 Fax +82-32-865-0944 E-mail ck@inha.ac.kr

*This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2022S1A5A2A03052617).

Received Sep 19, 2025 Revised Nov 25, 2025 Accepted Nov 27, 2025

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

PURPOSE

Exercise–microbiome research is expanding rapidly, but methodological heterogeneity and technical limitations still hinder reproducibility and mechanistic interpretation. This review provides a comprehensive methodological roadmap to overcome these barriers.

METHODS

We conducted a structured literature search in PubMed, Web of Science, and Scopus for records published between 1998 and 2025 using predefined combinations of exercise and gut-microbiome terms. After deduplication, titles/abstracts and full texts were screened according to prespecified criteria, yielding 99 eligible studies on aerobic and resistance/anaerobic exercise and gut microbiota. Evidence is critically appraised across standard short-read 16S rRNA protocols, shotgun metagenomics, and emerging long-read sequencing, as well as metatranscriptomics, metabolomics, and associated bioinformatics pipelines. A PRISMA-style flow diagram summarizes the study-selection process.

RESULTS

Exercise across diverse modalities reshapes gut microbial diversity and community structure, frequently enriching taxa such as Akkermansia muciniphila and Veillonella and enhancing production of short-chain fatty acids (SCFAs). SCFAs strengthen the intestinal barrier, activate anti-inflammatory pathways, and supply energy substrates for colonocytes and exercising muscle. Long-read sequencing now enables species- and strain-level resolution beyond short V-region amplicons, while inclusion of the gut mycobiome and virome expands ecological scope. Multi-omics designs integrating metagenomics, metatranscriptomics, and metabolomics connect microbial composition with functional outputs and host metabolic adaptations.

CONCLUSIONS

The future of exercise–microbiome science lies not in enlarging static catalogs of responsive microorganisms but in constructing individual-level predictive models. Integrating long-read sequencing and multi-omics with standardized training metadata will enable precision exercise prescriptions and microbiome-targeted interventions, including tailored probiotics, synbiotics, and nutrition strategies. Adoption of these advanced methodologies can accelerate mechanistic insight and promote translation of exercise– microbiome research into athletic performance optimization and clinical practice.

Keywords: Exercise, Gut microbiota, Long-Read sequencing, Multi-Omics, Metagenomics, Bioinformatics, Sports nutrition

INTRODUCTION

The human body maintains bidirectional communication with its resident microorganisms, forming a complex symbiotic system. Collectively, these microbial communities that form assemblages throughout the human body and inhabit specific ecological niches are defined as the human microbiome [1]. Early studies suggested that microbial cells out-number human somatic cells by more than tenfold [2]. Subsequent revolutionary advances in sequencing technologies and bioinformatic analysis have lowered costs and enhanced resolution [3]. As a result, current estimates indicate that microbial cells are present in numbers similar to those of human somatic cells and may even be slightly more abundant [2,4]. The majority of these microorganisms inhabit the gastrointestinal tract, where they perform essential physiological roles such as modulating immune responses, extracting energy, generating metabolic end-products, and synthesizing vitamins and amino acids [4]. Because of these diverse activities, the gut microbiota is often referred to as a “ for-gotten organ,” serving as a central regulator of host homeostasis through the gut–brain axis and exerting broad effects on health and disease, as increasingly documented in recent studies [5].

Large-scale international projects such as the Human Microbiome Project (HMP) [6] and the European MetaHIT initiative [7] have greatly advanced the characterization of the human gut microbiota. These studies revealed that the human intestine harbors an exceptionally diverse microbial community, which can be classified into at least eleven major bacterial phyla [6,7]. Among them, Firmicutes, Bacteroidetes, Proteobacteria, and Actinobacteria typically account for roughly 90% of the gut microbiota, whereas phyla such as Fusobacteria and Verrucomicrobia are present at lower relative abundances [7]. The composition of the gut microbiota shows considerable inter-individual variation, largely influenced by factors such as ethnicity [8], habitual diet [9], antibiotic exposure, and environmental conditions [10] that are increasingly recognized as key determinants of host physiology and health [11]. Increasing evidence shows that physical exercise can modulate gut microbial structure and function, spurring interest in how exercise-altered microbiota interact with the human host [12,13]. Exercise has been reported to enhance microbial diversity and richness, and to selectively enrich species capable of producing metabolites such as short-chain fatty acids (SCFAs), which contribute to gut health and serve as an energy source for the host [14,15]. Moreover, exercise-induced changes in gut microbial composition may exert positive feedback on behavior by promoting exercise motivation and sustained participation [16].

Rapidly growing interest in exercise–microbiome research since 2015 has made findings from human studies appear inconsistent [17]. This in-consistency arises because multiple confounding variables exist, including differences in exercise participation, exercise type, and exercise intensity, and because environmental factors such as regional and population-specific conditions further limit the generalization of analytical results [15]. Moreover, even among healthy individuals, substantial biological variability is evident. For example, the relative abundance of the major gut phyla Bacteroidetes and Firmicutes can range from 10% to 90%, and certain bacterial taxa may account for 5% of fecal microbiota in one person but less than 0.01% in another [6]. Methodological limitations in microbiome analysis also contribute to these discrepancies. Most studies rely on stool samples, which cannot fully represent microbes residing in the small intestine, and sample preservation conditions, including RNAlater solution, freezing, or dry swabs, can introduce systematic biases in DNA and metabolite profiles [3]. In addition, the selection of 16S rRNA primers, such as V3–V4 or V4, or the choice between 16S amplicon sequencing and shotgun metagenomics can yield different community profiles from the same specimen [3]. Discrepancies, delays in database updates, and incomplete reference data in commonly used databases such as Greengenes [18], SILVA [19], and Ribosomal Database Project (RDP) [20] may classify identical sequences under different bacterial names. Furthermore, statistical interpretation is challenged by features such as sparsity, compositional data structure, and zero inflation, which restrict the direct use of conventional statistical methods [3]. Even when analyzing identical samples, differences in DNA extraction, amplification, sequencing, and bioinformatic pipelines can substantially affect outcomes [21].

The present study aims to provide a critical and forward-looking methodological roadmap for researchers investigating exercise–microbiome interactions. Specifically, it integrates representative case studies based on established short-read sequencing methods and proposes comprehensive strategies for applying long-read sequencing and multi-omics technologies to overcome the technical and analytical limitations identified above.

METHODS

This review synthesizes current evidence on how exercise shapes the gut microbiome and evaluates analytic workflows centered on next-generation sequencing (NGS), specifically 16S rRNA amplicon sequencing and shotgun metagenomics, for identifying and characterizing exercise-related microbial changes. We performed a structured literature search and applied a PRISMA-style selection workflow adapted for a narrative review. PubMed, Web of Science, and Scopus were searched for records from January 1998 through November 2025 using Boolean combinations of exercise terms (exercise, physical activity, training, aerobic, anaerobic, resistance, HIIT, endurance, athlete, fitness) and microbiome terms (gut microbiome/microbiota, intestinal microbiota, metagenomics, shotgun, 16S rRNA). Retrieval was supplemented by backward and forward citation chasing. Results were consolidated in EndNote and de-duplicated using normalized titles plus publication year. We identified 21,971 records and removed 6,748 exact duplicates, leaving 15,223 records for title/abstract screening. Studies were included if they reported exercise or physical activity interventions with gut microbiota outcomes measured by 16S rRNA or shotgun metagenomic sequencing. Full texts were sought for 990 reports, all of which were retrieved. Among these, 891 were excluded for the following reasons: duplicate dataset (n=10), off-scope study type (n=23), no gut-microbiome endpoint (n=13), and no exercise outcome (n=845). A total of 99 studies met inclusion criteria and were included in this narrative synthesis. The study-selection flow and stage-wise counts are presented in Supplementary Fig. 1.

RESULTS

I. The role of gut microbiota in disease modulation

The gastrointestinal microbiome contributes to immune homeostasis, pathogen control, and nutrient–energy metabolism [22,23]. Clinical and experimental evidence demonstrates that metabolic and cardiovascular diseases are often accompanied by pronounced dysbiosis, notably a reduction of butyrate-producing genera such as Faecalibacterium prausnitzii and Roseburia intestinalis as well as enrichment of pro-inflammatory taxa including Ruminococcus and Fusobacterium [24]. These compositional changes are linked to diminished microbial diversity, impaired epithelial barrier integrity, and greater insulin resistance [25-27]. Exercise interventions, especially regular aerobic activity, have been linked to a partial recovery of SCFA-producing taxa, improvement in glycemic control, and changes in body composition [15].

At the microbial metabolic level, exercise-induced environmental changes activate bacterial fermentation pathways that enhance SCFA biosynthesis. SCFA-producing genera such as Faecalibacterium, Roseburia, and members of Lachnospiraceae upregulate key enzymes including butyryl-CoA:acetate CoA-transferase, which converts butyryl-CoA and acetate into butyrate and acetyl-CoA in the final step of butyrate production [27,28]. Furthermore, exercise-induced lactate production in skeletal muscle can be transported to the intestinal lumen, where lactate-utilizing species such as Veillonella convert it into propionate via the methylmalonyl-CoA pathway [29]. These adaptations underlie the beneficial effects of exercise on SCFA production, epithelial barrier integrity, anti-inflammatory responses, and metabolic flexibility. In cardiovascular research, microbial metabolism of choline and L-carnitine to tri-methylamine N-oxide (TMAO) is implicated in atherosclerosis, while habitual physical activity is associated with improved vascular function, better lipid profiles, and lower TMAO production [30,31].

Fermentation of dietary fiber by gut bacteria produces SCFAs such as acetate, propionate, and butyrate. Each SCFA has distinct physiological effects. Butyrate is a major energy source for colonocytes and plays an anti-inflammatory role by inhibiting histone deacetylase. Acetate acts as a metabolic substrate for muscle and liver, and propionate supports gluconeogenesis and regulates local pH in the gut [32]. Enrichment of SC-FA-producing taxa such as Faecalibacterium has frequently been observed in individuals who undergo endurance training [29]. Despite consistent observations, there is considerable heterogeneity among published studies. Variability in factors such as timing of sample collection in relation to exercise, SCFA quantification methods (for example GC-MS, HPLC, or LC-MS), and the use of fecal versus systemic SCFA measurements makes it difficult to draw definitive mechanistic conclusions [33]. The adoption of long-read sequencing platforms, including PacBio and Oxford Nanopore, now permits more precise strain-level identification and analysis of functional genes involved in SCFA biosynthesis [34,35]. Attention to study design factors including participant sex, training status, and dietary intake, is also needed because such variables have significant effects on gut microbial dynamics [8,9,36,37].

Emerging research suggests that microbially derived metabolites, including SCFAs and certain fatty acid amides may influence physical activity and motivation through the gut–brain axis [16]. These effects may involve activation of TRPV1-expressing sensory neurons and alterations in dopamine signaling. However, direct pathway-specific evidence in human exercise contexts remains limited. It is essential for future studies to use standardized methods, integrate multi-omics data, and include concurrent assessments of relevant host phenotypes. Such approaches will help clarify how the gut microbiota and its metabolites shape adaptation to exercise and overall health outcomes.

2. Effects of exercise on gut microbiota and metabolic regulation

Physical activity reshapes gut-microbiota composition and function and is associated with lower inflammation, improved insulin sensitivity, and tighter energy control [15]. Moderate exercise regulates cytokines such as Interleukin-6 (IL-6), Interleukin-10 (IL-10), and Tumor Necrosis Factor-alpha (TNF-α), promotes M2 macrophage polarization, enhances skeletal-muscle glucose transporter type 4 (GLUT4) expression, and activates AMP-activated protein kinase (AMPK) [38,39]. These adaptations stimulate mitochondrial biogenesis [40] and fatty-acid oxidation through peroxisome-proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) and Peroxisome Proliferator-Activated Receptor Gamma (PPARγ), which together help prevent obesity and type 2 diabetes [39,41].

These host metabolic changes also modulate the intestinal microenvironment through alterations in gut motility, mucus secretion, and oxygen gradients. Specifically, exercise-induced increases in enterocyte oxygen consumption create steeper oxygen gradients along the colonic mucosa [25], lowering luminal oxygen concentrations and establishing conditions that preferentially favor obligate anaerobic SCFA-producing taxa such as Faecalibacterium prausnitzii, Roseburia species, and Akkermansia muciniphila [42]. Aerobic exercise generally increases the production of SCFAs (primarily acetate, propionate, and butyrate), lowers luminal pH, and strengthens epithelial tight junctions, thereby supporting gut homeostasis [43].

1) Consistent findings in animal models and agreed mechanistic insights

It is important to distinguish between primary physiological adaptations induced by voluntary exercise and secondary stress-mediated responses characteristic of forced exercise paradigms. Voluntary aerobic exercise drives beneficial microbial shifts through direct metabolic mechanisms. Increased oxygen consumption and transient shifts in intestinal oxygen gradients selectively favor obligate anaerobes such as SC-FA-producing Faecalibacterium and Roseburia. These taxa proliferate under conditions of enhanced substrate availability and reduced luminal oxygen, leading to increased butyrate and propionate production that support gut barrier function and systemic metabolic health [25,44]. In contrast, forced exercise paradigms such as prolonged treadmill running at high intensity induce additional stress-mediated pathways through activation of the hypothalamic–pituitary–adrenal axis [45]. Acute exercise exceeding certain intensity thresholds triggers the release of stress hormones including glucocorticoids and catecholamines [44]. Elevated stress hormones increase intestinal permeability and activate pro-inflammatory pathways, disrupting microbial homeostasis. Studies in mice have demonstrated that forced treadmill running increases the abundance of pro-inflammatory genera, while voluntary wheel running attenuates these effects [25,44]. Thus, while moderate voluntary exercise fosters adaptive microbial changes via direct metabolic modulation, forced exercise introduces confounding stress-mediated disruptions. Differentiating these pathways is critical for accurately attributing microbial changes to exercise-driven adaptations versus stress-induced dysbiosis.

Animal experiments consistently demonstrate that exercise modifies gut microbiota toward a more diverse and SCFA-enriched configuration. For example, eight weeks of treadmill running improved mitochondrial metabolism in mice, and fecal microbiota transplantation transferred similar metabolic benefits to sedentary recipients [46]. Across diverse protocols, exercise enriches SCFA-producing taxa such as Faecalibacterium prausnitzii and Roseburia spp., increases systemic SCFA levels, and enhances host energy and glucose metabolism [47]. These findings support a causal role of exercise in driving microbiota-mediated improvements in metabolic flexibility [46,47]. Animal models enable rigorous control of genetic, diet, and environment. They allow precise manipulation of exercise dose and permit invasive sampling for mechanistic testing [48]. However, their gut microbial composition and host physiology differ from humans, and forced-exercise paradigms may induce stress responses that confound microbiota changes [46]. To address these methodological limitations, voluntary exercise models such as wheel running are increasingly recommended in the literature, as they minimize stress-induced alterations and better capture primary physiological adaptations to physical activity [44,48].

Recent evidence demonstrates that voluntary exercise enriches short-chain fatty acid (SCFA)-producing taxa and improves metabolic flexibility [49,50]. In contrast, forced treadmill running significantly increases gram-negative endotoxin-producing bacteria including Proteobacteria and Tenericutes, potentially confounding mechanistic interpretations of direct exercise-induced effects [44]. Voluntary wheel running produces a different microbial profile with reduced Turicibacter abundance and enrichment of SCFA-producing genera such as Anaerotruncus. These changes correlate with improved insulin sensitivity, metabolic flexibility, and reduced systemic inflammation [51]. Therefore, the choice of exercise model, whether forced or voluntary, substantially influences microbiota composition and functional outcomes, and future investigations should strategically select paradigms based on research objectives regarding either standardized mechanistic control or physiologically representative metabolic adaptation.

2) Addressing heterogeneity in human studies: the confounding roles of diet, exercise prescription, and host factors

Human trials show more variable results. Some interventions, such as a 6-week treadmill program in overweight women [52] and ultra-endurance rowing, reported increases in α-diversity and butyrate producers with long-lasting functional gains [53]. Yet a 2023 systematic review concluded that more than half of human studies detected no significant changes in microbial diversity following exercise [53]. Likely explanations include uncontrolled or poorly quantified diet, heterogeneity in exercise type, intensity, and duration, small samples underpowered for multivariate endpoints, and baseline differences in age, adiposity, medication, and habitual physical activity [36]. Furthermore, most studies rely on fecal sampling, which does not fully capture microbes of the small intestine, and they use different sequencing methods and reference databases, limiting comparability [3].

Despite these challenges, human studies provide essential clinical relevance by reflecting real-world complexity and by allowing the direct examination of exercise effects on disease outcomes [53]. Future work should integrate standardized dietary control or validated dietary assessments, preregistered and adequately powered exercise protocols, harmonized sampling schedules, and multicenter designs using shared reference materials [54]. Reporting should distinguish community composition from functional readouts such as metabolite and gene profiles and should not rely solely on alpha diversity as an indicator of exercise effects [53]. Broader analytic frameworks that integrate beta diversity, functional gene and transcript profiles, and microbially derived metabolites within a multi-omics design are needed to capture exercise-induced changes in both microbial composition and function more comprehensively [55]. Additional host factors, including immune status, gut motility, hormon-al fluctuations, and psychosocial stress, may further confound microbiome responses to exercise and therefore warrant careful control or statistical adjustment in future studies [37].

3. Methods for analyzing gut microbiota

Microbiome research has progressed from early marker-gene surveys, which sequence taxonomic marker genes such as the 16S rRNA gene, to high-resolution and multi-omic profiling [56]. Exact sequence variants (ASVs), defined as unique DNA sequences differing by a single nucleotide, have largely replaced operational taxonomic units (OTUs) to improve species-level resolution and reproducibility in amplicon data [57]. Current best practices increasingly combine shotgun metagenomics with long-read or hybrid sequencing, including PacBio HiFi, Oxford Nanopore Technologies, and Illumina Complete Long Read, to achieve strain-level and functional characterization [34,58]. Modern workflows treat sequence counts as compositional data and integrate multiple analytic layers such as metagenomics, metatranscriptomics, and metabolomics to connect microbial functions with host physiology and exercise-related outcomes such as energy metabolism and immune regulation [59]. Conventional culture-based counts and targeted qPCR panels can still provide absolute quantification of specific organisms or genes, but they cannot capture the full microbial diversity of the gut and are now used mainly for validation [60].

1) The next frontier: long-read sequencing for high-resolution taxonomic profiling

Long-read sequencing, represented by PacBio HiFi and Oxford Nanopore Technologies (ONT), generates sequence reads of several to tens of kilobases and is emerging as a next-generation tool for high-resolution microbiome analysis [33,35]. By covering the entire 16S rRNA gene or even the full ribosomal RNA (rRNA) operon, it enables species- and sometimes strain-level taxonomic profiling that short-read V3–V4 amplicons cannot achieve. This ability is decisive for detecting subtle, strain-specific microbial changes and functional pathways, such as SCFAs synthesis or neurotransmitter metabolism, that may occur during acute bouts of exercise or across long-term training cycles. Long-read data can also be integrated with shotgun metagenomics and metabolomics to strengthen causal links between microbial composition, energy metabolism, immune regulation, and exercise motivation.

Despite these strengths, current exercise–microbiome studies rarely employ long-read sequencing, reflecting higher costs, the need for larger amounts of high-quality DNA, and platform-specific error profiles that require tailored bioinformatic pipelines and error correction [34]. PacBio HiFi offers very high base accuracy, whereas ONT provides ultralong reads and field portability [61]. Illumina Complete Long Read combines short-read accuracy with kilobase-scale contiguity [62]. Selecting among these options requires balancing cost, portability, and resolution while ensuring that downstream analysis can handle compositional data and correct residual errors [63]. Looking ahead, applying long-read sequencing in exercise research could transform the field by allowing precise identification of strain-level taxa that modulate VO₂ max, lactate threshold, inflammatory responses, and other performance-related outcomes. Its ability to resolve complex genomes and to capture functionally important but previously hidden microbial players positions long-read sequencing as a powerful, yet still underused, method poised to advance mechanistic and personalized exercise–microbiome studies. Table 1 summarizes key sequencing platforms with their read characteristics and practical considerations for exercise–microbiome applications.

Table 1.

Sequencing platform guide for exercise-microbiome research

Platform	Target for analysis	Read length	Error profile	Throughput/cost	Advantages for exercise research	Key bioinformatic challenges
Illumina Short-Read [62]	• 16S rRNA V3-V4 • Shotgun metagenomics • ITS for fungal mycobiome	150-300 bp	Low (<0.1%)	High/low	• Cost-effective for large cohort studies (well-validated pipelines)	Limited resolution at species and strain levels
PacBio HiFi [62]	• Full-length 16S • ITS • Shotgun metagenomics	10-25 kb	Very low (< 0.1%)	Medium/high	• Enables strain-level tracking of exercise-responsive microbes • Accurate profiling of fungal communities (suitable for multi-omics sampling)	Requires high DNA input
Oxford Nanopore [58]	• Full-length 16S • ITS • Shotgun metagenomics • Viral metagenomics	>50 kb	Medium raw accuracy (improving with duplex/consensus polishing)	Medium/medium	• Real-time, field-deployable sequencing for time-series training studies; detects DNA/RNA modifications; (compatible with viral and fungal metagenomics)	Requires accurate base-calling and quality control to manage higher raw error rates and run-to-run variability.
Illumina Complete Long Read [62]	• Shotgun (synthetic long read)	∼6–7 kb N50 (kilobase-scale)	Very low (<0.1%)	High/medium	• Combines short-read accuracy with kilobase-scale contiguity and low DNA input (facilitates multi-omics integration for exercise-related functional analysis)	Requires further pipeline development and still struggles with very long repetitive genomes.

2) Characterizing the ‘Dark Matter’: mycobiome and virome

The gut ecosystem includes not only bacteria but also a mycobiome, which consists of commensal and opportunistic fungi, and a virome, which is largely composed of bacteriophages that infect gut bacteria [64,65]. These additional microbial kingdoms interact with the bacterial community to shape nutrient flux, immune tone, and epithelial barrier function, providing ecological and functional dimensions that standard bacterial surveys alone cannot capture [66,67]. Emerging evidence indicates that intense exercise training significantly remodels the human fungal microbiome composition [66] though intervention studies linking virome shifts with exercise, immune regulation, and recovery remain limited. In contexts of metabolic disease, fungal-bacterial interactions are implicated in modulating immune and metabolic outcomes, but more robust human data are needed [68].

3) Function prediction: limitations and validation strategy

PICRUSt2 infers the functional potential of microbial communities by placing 16S rRNA–derived amplicon sequence variants (ASVs) into a reference phylogeny and predicting genes and pathways from curated databases such as KEGG Orthology (KO), Enzyme Commission numbers (EC), and MetaCyc [69]. This approach generates predictions rather than direct measurements and therefore cannot confirm the actual presence, transcriptional activity, or metabolic output of the predicted genes. Because strain-specific accessory functions and environment-dependent gene expression remain undetected, mechanistic claims or clinical conclusions should not rely on predictive data alone, and absolute quantification or strain-level interpretation is inappropriate [69].

Despite these limitations, PICRUSt2 has been used in some endurance exercise training studies which showed predicted increases in fatty-acid related pathways [70] and metabolic function only under certain conditions [53]. To enhance rigor, users should report uncertainty metrics such as the Nearest Sequenced Taxon Index (NSTI) and reference genome coverage and apply compositional data statistics such as centered log-ratio transformations. Although direct evidence linking PIC-RUSt2-predicted pathways with VO₂ max or lactate threshold is sparse, Durk et al. [71] reported a correlation between VO₂ max and gut microbiota composition in humans, and Soriano et al. [72] observed PIC-RUSt2 pathway shifts through the football season in non-concussed athletes. These findings suggest that predictive pathway analyses have the potential to relate microbial functions to exercise performance, if interventions include appropriate phenotyping and longitudinal sampling.

4) Blueprints for multi-omics integration

To understand how exercise shapes the gut microbiome and host metabolism in an integrated manner, samples for DNA, RNA, and metabolites should be collected from the same specimen and at matched time points during training cycles [3]. Recording key lifestyle factors, including habitual diet, sleep duration and quality, menstrual phase, medication use, and recent exercise load, in a standardized format is essential to control for confounding influences on microbial and metabolic profiles [3,52,73]. Statistical frameworks such as multiblock modeling and net-work analysis can integrate these data layers to reveal how microbial genes and metabolites relate to exercise performance indicators, including maximal oxygen uptake (VO₂ max), lactate threshold, power output, recovery rate, and heart-rate variability. For example, women with overweight or obesity undergoing a 6-week aerobic program showed metabolomic changes that correlated with fitness improvements [37,52]. However, only a few studies verify key findings with targeted laboratory analyses matched to specific training phases, an approach essential for establishing causal relationships.

5) Shotgun metagenomics and shotgun approaches

16S rRNA gene sequencing targets a bacterial and archaeal marker gene absent from host nuclear DNA, enabling selective profiling of microbial communities [74]. This approximately 1,500-bp gene contains nine hypervariable regions (V1–V9) flanked by conserved segments that provide a phylogenetic fingerprint for taxonomic classification [75]. Sequencing specific regions, most commonly V3–V4 or the full-length 16S gene, reduces cost and data requirements compared with shotgun metagenomics and remains a workhorse for assessing community structure and diversity. Full-length and long-read protocols, such as PacBio HiFi and Oxford Nanopore sequencing, now improve species- and sometimes strain-level resolution. Nevertheless, as a PCR-based approach it is subject to primer and amplification bias and provides limited direct functional information [76].

Shotgun metagenomics fragments and sequences all DNA in a sample to reconstruct the entire microbial community and detect taxa and functional genes beyond the reach of 16S rRNA surveys [77]. This method provides species- and often strain-level resolution and allows direct identification of genes and metabolic pathways [78]. Combined with long-read or hybrid assemblies, shotgun metagenomics now supports strain-resolved genome reconstruction and discovery of previously un-recognized microbes that may respond to specific exercise regimens and potentially contribute to metabolic adaptation [79]. Although it requires more complex laboratory and computational workflows, recent improvements in library preparation, sequencing cost, and assembly algorithms have made large-scale human exercise studies increasingly feasible [80,81]. A comparative summary of the major NGS-based methodologies and their key applications in exercise–microbiome research is provided in Table 2.

Table 2.

Principal NGS-based methodologies and their applications for exercise-related gut microbiome analysis

Approach	Primary measurement	Representative application	Principal outputs	Strengths	Limitations	Ref.
16S rRNA Amplicon sequencing	• Taxonomic composition • Genus level routinely (Species level possible in favorable cases)	• Composition and diversity comparisons • Pre- vs. Post intervention assessment • Large-cohort studies	• ASV table • Taxonomy table • Alpha and beta diversity metrics • Rarefaction curves	• Lower cost and high throughput • Standardized pipelines	• Limited species/strain-level resolution (prone to primer bias) • No direct functional genes information	[74]
Shotgun metagenomics & WGS	• Species level (Sometimes near strain level) • Direct gene content and pathways	• Species- and strain-level taxonomy, including discovery of novel taxa • Functional gene and metabolic-pathway analysis • Mechanistic microbiome-host interaction studies	• Species profiles • Gene families • Pathway abundance	• Joint view of taxonomy and function • Enables gene-level hypotheses	• Higher cost and computational complexity • Sensitive to sample quality and contamination	[77]
Metatranscriptomics	• Microbial gene expression and active pathways	• Real-time functional activity profiling • Pre-post exercise or time-course response studies	• Gene expression matrices • Pathway activity profiles	• Captures community activity in real-time • Suitable for time-course designs	• RNA instability and short half-life • Potential host RNA contamination • Requires paired metagenomics for full interpretation	[97]
Metabolomics (targeted)	• Quantification of predefined metabolites (e.g., SCFAs, bile acids, amino acids)	• Hypothesis-driven validation and quantification of key microbial metabolites	• Concentration tables for selected metabolites	• Direct functional readout • High sensitivity and accuracy	• Limited chemical coverage • Requires predefined panels	[98]
Metabolomics (untargeted)	• Broad small-molecule profiling without a predefined list	• Discovery of new metabolites and pathways • Guides later targeted assays	• Feature table with entative annotations	• Wide discovery potential	• Identification is challenging • Interpretation is complex and time-consuming	[99]
Mycobiome (ITS sequencing)	• Fungal community composition and diversity at species level	• Assessment of fungal shifts induced by exercise • Evaluation of fungi-driven metabolic or immune modulation	• ITS sequence variant tables • Fungal alpha- and beta-diversity metrics	• Standard DNA barcode for fungi • Species-level resolution across most fungal taxa	• Lower biomass than bacteria increase contamination risk • Primer and extraction bias; incomplete fungal reference databases	[64]
Virome (viral metagenomics)	Viral community structure, especially bacteriophages	• Characterization of exercise-induced phage–host dynamics • Identification of viral functions influencing microbial networks and host immunity	• Viral taxonomic profiles • Phage–host interaction networks	• Expands analysis beyond bacteria and fungi • Reveals phage-mediated microbial and metabolic regulation	• Requires specialized virus-like particle enrichment • Reference databases remain incomplete	[64]

6) Metatranscriptomics for active microbial functions

Metatranscriptomics profiles RNA transcripts to capture the active metabolic state of gut microbes and their immediate responses to environmental cues, including exercise. Because microbial RNA degrades rapidly and can mix with host RNA, rapid stabilization and careful extraction are essential [82]. Recent advances, such as long-read direct RNA sequencing and time-resolved sampling, permit more accurate detection of strain-specific transcripts and improve the ability to capture transient microbial responses during or immediately after training, al-though human exercise-intervention data remain scarce [83]. Interpreting metatranscriptomic data alongside metagenomic profiles helps distinguish genuine microbial expression from simple compositional shifts and strengthens causal inference [82]. In exercise–microbiome research, this approach holds promise for revealing how gut microbes acutely modulate pathways for energy production, SCFAs synthesis, and neurotransmitter precursors in response to endurance or resistance exercise, but confirmation in well-controlled training studies is still needed [84].

7) Metabolomics for functional readouts and host interaction

Metabolomics identifies small molecules such as SCFAs, bile acids, amino acids, and vitamins that represent the functional output of gut microbes [75]. These metabolites influence host energy balance, immunity, and recovery after exercise. High-resolution liquid chromatography–mass spectrometry (LC–MS) and gas chromatography–mass spectrometry (GC–MS) now enable both targeted quantification of predefined metabolites and untargeted discovery of novel compounds, while stable isotope tracing and spatial metabolomics reveal pathway fluxes and tissue distribution [84,85]. Such combined analyses are particularly valuable for connecting SCFA dynamics and amino-acid-de-rived metabolites to endurance capacity, muscle repair, and anti-inflammatory signaling during structured exercise programs [85].

4. A proposed best-practice workflow for microbiome data generation and analysis

Falling sequencing costs and maturing open-source toolchains have expanded microbiome research while highlighting how strongly pre-analytical and analytical decisions influence biological inference [86]. This section presents a sequence of decision points from sampling to computation, outlining common defaults, justified alternatives, and their implications for downstream analysis. Although examples refer mainly to short-read Illumina amplicon workflows, the principles extend to other platforms and to shotgun metagenomics. An integrated workflow summarizing critical decision points from sampling to multiomics integration is depicted in Fig. 1.

Fig. 1.

Study design and analytical pipeline for exercise–microbiome research. Schematic of the major steps and decision points in exercise–microbiome studies. The upper section summarizes sample handling (preservation, co-aliquoting for DNA/RNA/metabolites), sequencing-target selection (short- vs long-read; targeted or shotgun), and core bioinformatics workflows. The lower section outlines cross-omic integration with host training metrics (e.g., VO-2 max, lactate threshold, heart-rate variability [HRV]) and key design choices, including study design (cross-sectional vs. longitudinal with pre-, immediate-post-, and recovery sampling), trade-offs between sequencing depth and sample size, and analytic-target selection. HRV, heart-rate variability.

1) Sample collection and preservation

Sampling and early handling determine whether RNA- and metabolite-level analyses remain feasible and account for a large share of be-tween-study variability [3]. While immediate −80°C freezing provides broad compatibility with both 16S and shotgun designs and represents the gold standard for nucleic-acid preservation, this approach is logistically impractical for many field-based exercise studies. In field-based contexts or when metatranscriptomics or metabolomics analyses are planned, validated room-temperature preservatives offer viable alternatives that maintain nucleic acid and metabolite stability during transport. DNA/RNA Shield (OMNIgene.GUT) maintains nucleic acid stability for approximately one month at ambient temperature (4–25°C) with compositional deviations comparable to ethanol-preserved samples and refrigerated controls [37]. Both FTA cards, which preserve fecal samples for up to eight weeks at ambient temperature, and 95% ethanol provide alternatives with comparable long-term stability. In contrast, 70% ethanol should be avoided due to substantial microbiota alterations [87]. For short-term storage (≤24 hours), refrigeration at 4°C minimizes nucleic-acid degradation and metabolic profile changes compared to room-temperature or −20°C storage [88]. Investigators should document preservative chemistry and, where possible, evaluate potential compositional shifts with appropriate positive and negative controls [37].

Because gut microbial communities can fluctuate on an hourly to daily basis, researchers should design sampling schedules that capture both acute and chronic exercise-induced changes, including repeated or time-series collections within training microcycles [89]. Stool samples may not fully represent microbes from the small intestine or mucosa. Where study goals specifically include exercise-induced changes in nutrient absorption or barrier interactions, additional sampling of mucosal or intestinal lumen contents should be considered [3]. For multi-omics designs, pre-aliquoting aliquots from the same stool passage ensures that DNA, RNA, and metabolites originate from identical material, and re-cording the time-to-preservation, ambient conditions during storage and transport, and any freeze–thaw events facilitate robust cross-omic harmonization and batch effect correction [87].

2) DNA extraction and quantification

Extraction methods affect which taxa are recovered and at what integrity. Bead-beating with chemical lysis increases recovery of Gram-positive organisms but may shear long DNA fragments; gentler lysis reduces frag-mentation at the cost of under-representing robust cell walls [90]. When multi-omics integration is planned, RNA and metabolite extraction from the same aliquot is recommended to preserve cross-modal comparability [91]. Fluorescence-based quantification and inclusion of negative controls help separate true biological signals from technical noise [37].

3) 16S rRNA gene amplification and library preparation

Choosing between 16S rRNA amplicon sequencing and shotgun metagenomics depends on the desired taxonomic or functional resolution, budget, and sample size [37]. Amplicon sequencing remains cost-efficient for describing community structure but only infers function in-directly, whereas shotgun sequencing supports strain-level profiling and direct gene/pathway analysis at higher cost and depth. Primer set and variable region (e.g., V1–V3 vs. V3–V5 vs. V4) selection can markedly alter species-level classification [6]. Explicitly reporting primer choice and testing mock communities are essential to evaluate and mitigate amplification bias [3]. Within amplicon workflows, the primer set, targeted region, and read length must be chosen with known biases and the desired resolution in mind. As illustrated in Fig. 2, 16S rRNA analyses of exercise-microbiota studies typically reveal genus-level composition, diversity metrics, and exercise-associated taxa.

Fig. 2.

Illustrative 16S rRNA amplicon outputs for exercise–gut microbiome comparisons. Representative outputs from 16S rRNA amplicon analyses com-paring Aerobic, Anaerobic, Control, and Control (high) groups. (A) Genus-level taxonomic composition displayed as relative abundance. (B) Beta diversity based on Bray–Curtis dissimilarity (PCoA), PERMANOVA statistics (F, R², p) are indicated. (C) Beta diversity based on unweighted UniFrac distance (PCoA), PERMANOVA statistics are indicated. (D) LEfSe bar plot of differential taxa at the genus level with LDA score (log10) on the x-axis; y-axis labels are anonymized placeholders (Genus A–J) for illustration and correspond to genus names in applied analyses. Together, these panels illustrate how standard 16S workflows capture genus-level composition, beta-diversity structure, and differential taxa in exercise–microbiome studies.

4) Library strategy and sequencing depth

Sequencing strategy involves a trade-off between depth per sample and number of samples. Deeper libraries improve richness estimates and rare-taxon discovery, whereas large cohorts may prioritize sample number to increase statistical power [57]. Because exercise intervention studies often require time-series sampling, researchers should balance per-sample depth with the need for repeated measures across different exercise phases [37]. Normalization and pooling prevent run imbalance, and spike-in controls facilitate run calibration [92].

5) Computational processing and multi-omics integration

Post-sequencing, quality profiles guide trimming and truncation. Amplicon sequence variant (ASV) denoisers such as DADA2 or Deblur enhance precision and cross-study comparability. To ensure reliable results, chimera removal is essential [57]. Taxonomy is assigned with region-matched classifiers trained on curated databases (e.g., SILVA, GTDB), and phylogenies constructed by maximum-likelihood or fragment-insertion methods enable phylogeny-aware β-diversity [57]. Analytical outcomes can vary substantially depending on the chosen pipelines and reference databases [37]. Researchers should document software versions, parameter settings, and database update times, and whenever possible share code and processed data to enhance reproducibility [3]. Because microbiome tables are compositional, appropriate statistics, log-ratio transformations and compositional distance metrics such as Aitchison or PhILR, are recommended. Zero handling should be reported explicitly [93]. Multi-omics integration can link metagenomic, metatranscriptomic, and metabolomic layers with host exercise phenotypes using constrained ordination, multiblock projection (PLS/DIABLO), factor models (MOFA), or pathway-level mappings (KEGG/MetaCyc). DNA extraction should, where possible, permit parallel RNA and metabolite recovery to ensure compatibility across data types [94]. Integrating direct measures of metabolites and gene expression is especially important for exercise studies, where short-term metabolic flux and gene activation often determine training adaptation [3].

6) Reproducibility and reporting

Reproducibility requires transparent reporting of analysis code and software environments. This includes providing access to custom scripts or notebooks (for example, R scripts for statistical modeling or QIIME2 command scripts for microbiome analysis) and specifying the exact software versions used (for example, QIIME2 version 2025.7, R version 4.3.1 with DADA2 1.28.0) [57,95]. Depositing sequencing reads and essential metadata in public repositories and clearly listing key experimental de-tails such as DNA extraction kits, catalog numbers, and sequencing run parameters in the Methods section supports data re-use and cross-study comparison [3]. Consistent reporting of sampling frequency, sample type (stool versus mucosal), and exercise-related metadata (for example, work-load and recovery period) further improves the ability to synthesize findings across studies and enables meaningful meta-analysis [37].

DISCUSSION

Accumulating evidence shows that exercise beneficially reshapes the gut microbiota, with meaningful effects on host health and physical performance. Moderate-intensity aerobic training modifies microbial diversity and community structure in obese, hypertensive, and healthy animal models [11]. Reported changes include a reduction in Proteobacteria, often considered potential pathobionts, and an increase in beneficial genera such as Lactobacillus, which support gut barrier integrity and metabolic homeostasis [15]. Voluntary wheel running also counter-acts high-fat-diet–induced weight gain and metabolic impairment while enriching microorganisms that enhance fermentative energy metabolism, consistent with anti-obesity and metabolic benefits [96]. Collectively, these studies indicate that, despite individual variation in genetics and disease status, exercise acts as a key environmental factor that steers the gut ecosystem toward healthier states [15]. Interactions between the gut microbiome and the brain further underscore the systemic influence of physical activity. In animal experiments, exposure to an enriched environment that mimics regular exercise, combined with sup-plementation of A. muciniphila, a bacterium frequently enriched by aerobic training, alleviated diet-induced cognitive decline and restored spatial memory and object recognition [3]. Human studies support similar links. For example, in marathon runners, the genus Veillonella increased after competition and can convert exercise-generated lactate into propionate, a metabolite associated with improved endurance capacity [29]. More broadly, elite athletes typically display higher microbial diversity and greater functional potential than sedentary individuals, including metabolic profiles consistent with enhanced energy turnover and lower inflammatory tone [15]. These observations suggest that exercise-induced microbial and metabolic remodeling contributes to improved energy efficiency, immune balance, and neurocognitive function [15].

Building on these insights, the field should now move beyond cataloging exercise-associated taxa and instead focus on mechanistic and predictive frameworks. Future investigations need to integrate individual exercise characteristics, including training volume, intensity, and recovery period, with microbiome features to design personalized exercise prescriptions and microbiome-guided performance interventions [37]. To realize these applications, researchers should deploy advanced analytic tools capable of integrating multi-layered datasets, including high-resolution 16S rRNA and shotgun metagenomic profiling to define community structure and genetic potential, metatranscriptomics and metabolomics to identify active pathways, and machine-learning models to construct predictive and personalized analyses [37].

CONCLUSION

The future of exercise–microbiome research lies in creating individual-level predictive models that link high-resolution sequencing and multi-omics with standardized training metadata. Such models will enable precision exercise medicine, combining tailored training regimens with candidate probiotics, synbiotics, fiber formulations, and other targeted nutritional strategies to optimize metabolic health and athletic performance. Delivering this agenda will require mechanism-oriented longitudinal studies, interoperable analytic pipelines, and transparent reporting that facilitates replication across diverse cohorts, together with rigorous evaluation in randomized or crossover designs and careful safety monitoring. By embracing technological advances including long-read sequencing, multi-omics integration, and predictive modeling, future research can establish a new paradigm in which exercise science and microbiome research converge to promote health and enhance human performance.

Supplementary Material

ksep-2025-00570-Supplementary-Fig-1.pdf

Notes

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS

Conceptualization: C Kang; Data curation: YM Park, D Kwak, J Jang, J Park; Formal analysis: C Kang; Funding acquisition: C Kang; Methodology: C Kang, K Kim, H Lee, B So; Project administration: C Kang; Visualization: B So; Writing - original draft: B So; Writing - review & editing: C Kang, B So.

REFERENCES

1. Lynch SV, Pedersen O. The Human Intestinal Microbiome in Health and Disease. N Engl J Med. 2016;375(24):2369-79.

2. Sender R, Fuchs S, Milo R. Revised Estimates for the Number of Human and Bacteria Cells in the Body. PLoS Biol. 2016;14(8):e1002533.

3. Allaband C, McDonald D, Vázquez-Baeza Y, Minich JJ, Tripathi A, et al. Microbiome 101: Studying, Analyzing, and Interpreting Gut Microbiome Data for Clinicians. Clin Gastroenterol Hepatol. 2019;17(2):218-30.

4. Gilbert JA, Blaser MJ, Caporaso JG, Jansson JK, Lynch SV, et al. Current understanding of the human microbiome. Nat Med. 2018;24(4):392-400.

5. Dinan TG, Cryan JF. The Microbiome-Gut-Brain Axis in Health and Disease. Gastroenterol Clin North Am. 2017;46(1):77-89.

6. The Integrative Human Microbiome Project. Nature. 2019;569(7758):641-8.

7. Qin J, Li R, Raes J, Arumugam M, Burgdorf KS, et al. A human gut microbial gene catalogue established by metagenomic sequencing. Nature. 2010;464(7285):59-65.

8. Dwiyanto J, Hussain MH, Reidpath D, Ong KS, Qasim A, et al. Ethnicity influences the gut microbiota of individuals sharing a geographical location: a cross-sectional study from a middle-income country. Sci Rep. 2021;11(1):2618.

9. De Filippo C, Cavalieri D, Di Paola M, Ramazzotti M, Poullet JB, et al. Impact of diet in shaping gut microbiota revealed by a comparative study in children from Europe and rural Africa. Proc Natl Acad Sci U S A. 2010;107(33):14691-6.

10. Yatsunenko T, Rey FE, Manary MJ, Trehan I, Dominguez-Bello MG, et al. Human gut microbiome viewed across age and geography. Nature. 2012;486(7402):222-7.

11. Hughes RL. A Review of the Role of the Gut Microbiome in Personalized Sports Nutrition. Front Nutr. 2019;6:191.

12. Clarke SF, Murphy EF, O'Sullivan O, Lucey AJ, Humphreys M, et al. Exercise and associated dietary extremes impact on gut microbial diversity. Gut. 2014;63(12):1913-20.

13. Fritz P, Fritz R, Bóday P, Bóday Á, Bató E, et al. Gut microbiome composition: link between sports performance and protein absorption? J Int Soc Sports Nutr. 2024;21(1):2297992.

14. Hooper LV, Gordon JI. Commensal host-bacterial relationships in the gut. Science. 2001;292(5519):1115-8.

15. Clauss M, Gérard P, Mosca A, Leclerc M. Interplay Between Exercise and Gut Microbiome in the Context of Human Health and Performance. Front Nutr. 2021;8:637010.

16. Dohnalová L, Lundgren P, Carty JRE, Goldstein N, Wenski SL, et al. A microbiome-dependent gut-brain pathway regulates motivation for exercise. Nature. 2022;612(7941):739-47.

17. Puschhof J, Elinav E. Human microbiome research: Growing pains and future promises. PLoS Biol. 2023;21(3):e3002053.

18. DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL, et al. Greengenes, a chimera-checked 16S rRNA gene database and work-bench compatible with ARB. Appl Environ Microbiol. 2006;72(7):5069-72.

19. Pruesse E, Quast C, Knittel K, Fuchs BM, Ludwig W, et al. SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res. 2007;35(21):7188-96.

20. Cole JR, Wang Q, Fish JA, Chai B, McGarrell DM, et al. Ribosomal Database Project: data and tools for high throughput rRNA analysis. Nucleic Acids Res. 2014;42(Database issue):D633-42.

21. Dias CK, Starke R, Pylro VS, Morais DK. Database limitations for studying the human gut microbiome. PeerJ Comput Sci. 2020;6:e289.

22. Bäckhed F, Ley RE, Sonnenburg JL, Peterson DA, Gordon JI. Host-bacterial mutualism in the human intestine. Science. 2005;307(5717):1915-20.

23. Bengmark S. Ecological control of the gastrointestinal tract. The role of probiotic flora. Gut. 1998;42(1):2-7.

24. Carding S, Verbeke K, Vipond DT, Corfe BM, Owen LJ. Dysbiosis of the gut microbiota in disease. Microb Ecol Health Dis. 2015;26:26191.

25. Dmytriv TR, Storey KB, Lushchak VI. Intestinal barrier permeability: the influence of gut microbiota, nutrition, and exercise. Front Physiol. 2024;15:1380713.

26. Hong J, Fu T, Liu W, Du Y, Bu J, et al. Specific Alternation of Gut Microbiota and the Role of Ruminococcus gnavus in the Development of Diabetic Nephropathy. J Microbiol Biotechnol. 2024;34(3):547-61.

27. Saban Güler M, Arslan S, Ağagündüz D, Cerqua I, Pagano E, et al. Butyrate: A potential mediator of obesity and microbiome via different mechanisms of actions. Food Res Int. 2025;199:115420.

28. Munoz-Tamayo R, Laroche B, Walter E, Dore J, Duncan SH, et al. Kinetic modelling of lactate utilization and butyrate production by key human colonic bacterial species. FEMS microbiology ecology. 2011;76(3):615-24.

29. Scheiman J, Luber JM, Chavkin TA, MacDonald T, Tung A, et al. Meta-omics analysis of elite athletes identifies a performance-enhancing microbe that functions via lactate metabolism. Nat Med. 2019;25(7):1104-9.

30. Jie Z, Xia H, Zhong SL, Feng Q, Li S, et al. The gut microbiome in atherosclerotic cardiovascular disease. Nat Commun. 2017;8(1):845.

31. Marques FZ, Nelson E, Chu PY, Horlock D, Fiedler A, et al. High-Fiber Diet and Acetate Supplementation Change the Gut Microbiota and Prevent the Development of Hypertension and Heart Failure in Hypertensive Mice. Circulation. 2017;135(10):964-77.

32. Spiljar M, Merkler D, Trajkovski M. The Immune System Bridges the Gut Microbiota with Systemic Energy Homeostasis: Focus on TLRs, Mucosal Barrier, and SCFAs. Frontiers in Immunology. 2017;8.

33. Callahan BJ, Wong J, Heiner C, Oh S, Theriot CM, et al. High-throughput amplicon sequencing of the full-length 16S rRNA gene with single-nucleotide resolution. Nucleic Acids Res. 2019;47(18):e103.

34. Lang D, Zhang S, Ren P, Liang F, Sun Z, et al. Comparison of the two up-to-date sequencing technologies for genome assembly: HiFi reads of Pacific Biosciences Sequel II system and ultralong reads of Oxford Nanopore. Gigascience. 2020;9(12).

35. González A, Fullaondo A, Odriozola A. Why Are Long-Read Sequencing Methods Revolutionizing Microbiome Analysis? Microorganisms. 2025;13(8).

36. Pérez-Prieto I, Plaza-Florido A, Ubago-Guisado E, Ortega FB, Altmäe S. Physical activity, sedentary behavior and microbiome: A systematic review and meta-analysis. Journal of Science and Medicine in Sport. 2024;27(11):793-804.

37. Mancin L, Paoli A, Berry S, Gonzalez JT, Collins AJ, et al. Standardization of gut microbiome analysis in sports. Cell Rep Med. 2024;5(10):101759.

38. Callegari IOM, Rocha GZ, Oliveira AG. Physical exercise, health, and disease treatment: The role of macrophages. Front Physiol. 2023;14:1061353.

39. Gross B, Pawlak M, Lefebvre P, Staels B. PPARs in obesity-induced T2DM, dyslipidaemia and NAFLD. Nat Rev Endocrinol. 2017;13(1):36-49.

40. Hood DA, Uguccioni G, Vainshtein A, D'Souza D. Mechanisms of exercise-induced mitochondrial biogenesis in skeletal muscle: implications for health and disease. Compr Physiol. 2011;1(3):1119-34.

41. Wright DC, Han DH, Garcia-Roves PM, Geiger PC, Jones TE, et al. Exercise-induced mitochondrial biogenesis begins before the increase in muscle PGC-1alpha expression. J Biol Chem. 2007;282(1):194-9.

42. Fang J, Yan W, Sun X, Chen J. The role of exercise-induced short-chain fatty acids in the gut-muscle axis: implications for sarcopenia prevention and therapy. Front Microbiol. 2025;16:1665551.

43. Wong JM, de Souza R, Kendall CW, Emam A, Jenkins DJ. Colonic health: fermentation and short chain fatty acids. J Clin Gastroenterol. 2006;40(3):235-43.

44. Allen JM, Berg Miller ME, Pence BD, Whitlock K, Nehra V, et al. Voluntary and forced exercise differentially alters the gut microbiome in C57BL/6J mice. J Appl Physiol. (1985). 2015;118(8):1059-66.

45. Cook MD, Martin SA, Williams C, Whitlock K, Wallig MA, et al. Forced treadmill exercise training exacerbates inflammation and causes mortality while voluntary wheel training is protective in a mouse model of colitis. Brain Behav Immun. 2013;33:46-56.

46. Uchida M, Fujie S, Yano H, Iemitsu M. Aerobic exercise training-induced alteration of gut microbiota composition affects endurance capacity. J Physiol. 2023;601(12):2329-44.

47. Martin D, Bonneau M, Orfila L, Horeau M, Hazon M, et al. Atypical gut microbial ecosystem from athletes with very high exercise capacity improves insulin sensitivity and muscle glycogen store in mice. Cell Rep. 2025;44(4):115448.

48. Mohr AE, Mach N, Pugh J, Grosicki GJ, Allen JM, et al. Mechanisms underlying alterations of the gut microbiota by exercise and their role in shaping ecological resilience. FEMS Microbiol Rev. 2025;49.

49. Kim HJ, Kim YJ, Kim YJ, Baek JH, Kim HS, et al. Microbiota influences host exercise capacity via modulation of skeletal muscle glucose metabolism in mice. Exp Mol Med. 2023;55(8):1820-30.

50. Cook MD, Hogue T. Exercise and the Microbiome: Mechanistic Perspectives of the Impact of Exercise on the Gut-Vascular Axis. mSystems. 2021;6(4):e0065021.

51. Pereira SMS, José V, da Silva A, Martins KVC, Leite LB, et al. Can physical exercise modify intestinal integrity and gut microbiota composition? A systematic review of in vivo studies. Biol Sport. 2025;42(4):13-28.

52. Hintikka JE, Ahtiainen JP, Permi P, Jalkanen S, Lehtonen M, et al. Aerobic exercise training and gut microbiome-associated metabolic shifts in women with overweight: a multi-omic study. Sci Rep. 2023;13(1):11228.

53. Cullen JMA, Shahzad S, Dhillon J. A systematic review on the effects of exercise on gut microbial diversity, taxonomic composition, and microbial metabolites: identifying research gaps and future directions. Front Physiol. 2023;14:1292673.

54. Min L, Ablitip A, Wang R, Luciana T, Wei M, et al. Effects of Exercise on Gut Microbiota of Adults: A Systematic Review and Meta-Analysis. Nutrients. 2024;16(7).

55. Manor O, Dai CL, Kornilov SA, Smith B, Price ND, et al. Health and disease markers correlate with gut microbiome composition across thousands of people. Nat Commun. 2020;11(1):5206.

56. Jeske JT, Gallert C. Microbiome Analysis via OTU and ASV-Based Pipelines-A Comparative Interpretation of Ecological Data in WWTP Systems. Bioengineering. (Basel). 2022;9(4).

57. Xia Y, Sun J. Bioinformatic and statistical analysis of microbiome data: from raw sequences to advanced modeling with QIIME 2 and R. Springer Nature 2023.

58. Schatz MC. Nanopore sequencing meets epigenetics. Nat Methods. 2017;14(4):347-8.

59. Aguiar-Pulido V, Huang W, Suarez-Ulloa V, Cickovski T, Mathee K, et al. Metagenomics, Metatranscriptomics, and Metabolomics Approaches for Microbiome Analysis. Evol Bioinform Online. 2016;12(Suppl 1):5-16.

60. Jian C, Luukkonen P, Yki-Järvinen H, Salonen A, Korpela K. Quantitative PCR provides a simple and accessible method for quantitative microbiota profiling. PLoS One. 2020;15(1):e0227285.

61. Kim CY, Ma J, Lee I. HiFi metagenomic sequencing enables assembly of accurate and complete genomes from human gut microbiota. Nat Commun. 2022;13(1):6367.

62. Maghini DG, Kiguchi Y, Darling AE, Monahan LG, Halpern AL, et al. Illumina complete long read assay yields contiguous bacterial genomes from human gut metagenomes. mSystems. 2025;10(8):e0153124.

63. Kerkhof LJ. Is Oxford Nanopore sequencing ready for analyzing complex microbiomes? FEMS Microbiol Ecol. 2021;97(3).

64. Seed PC. The human mycobiome. Cold Spring Harb Perspect Med. 2014;5(5):a019810.

65. Matijašić M, Meštrović T, Paljetak H, Perić M, Barešić A, et al. Gut Microbiota beyond Bacteria-Mycobiome, Virome, Archaeome, and Eukaryotic Parasites in IBD. Int J Mol Sci. 2020;21(8).

66. Wang Y, Chen J, Ni Y, Liu Y, Gao X, et al. Exercise-changed gut mycobiome as a potential contributor to metabolic benefits in diabetes prevention: an integrative multi-omics study. Gut Microbes. 2024;16(1):2416928.

67. Pargin E, Roach MJ, Skye A, Papudeshi B, Inglis LK, et al. The human gut virome: composition, colonization, interactions, and impacts on human health. Front Microbiol. 2023;14:963173.

68. Wang L, Zhang K, Zeng Y, Luo Y, Peng J, et al. Gut mycobiome and metabolic diseases: The known, the unknown, and the future. Pharmacol Res. 2023;193:106807.

69. Douglas GM, Maffei VJ, Zaneveld JR, Yurgel SN, Brown JR, et al. PIC-RUSt2 for prediction of metagenome functions. Nat Biotechnol. 2020;38(6):685-8.

70. Yoon EJ, Lee SR, Ortutu BF, Kim JO, Jaiswal V, et al. Effect of Endurance Exercise Training on Gut Microbiota and ER Stress. Int J Mol Sci. 2024;25(19).

71. Durk RP, Castillo E, Márquez-Magaña L, Grosicki GJ, Bolter ND, et al. Gut Microbiota Composition Is Related to Cardiorespiratory Fitness in Healthy Young Adults. Int J Sport Nutr Exerc Metab. 2019;29(3):249-53.

72. Soriano S, Curry K, Sadrameli SS, Wang Q, Nute M, et al. Alterations to the gut microbiome after sport-related concussion in a collegiate football players cohort: A pilot study. Brain Behav Immun Health. 2022;21:100438.

73. Barton W, Cronin O, Garcia-Perez I, Whiston R, Holmes E, et al. The effects of sustained fitness improvement on the gut microbiome: A longitudinal, repeated measures case-study approach. Transl Sports Med. 2021;4(2):174-92.

74. Regueira-Iglesias A, Balsa-Castro C, Blanco-Pintos T, Tomás I. Critical review of 16S rRNA gene sequencing workflow in microbiome studies: From primer selection to advanced data analysis. Mol Oral Microbiol. 2023;38(5):347-99.

75. Rodicio Mdel R, Mendoza Mdel C. [Identification of bacteria through 16S rRNA sequencing: principles, methods and applications in clinical microbiology]. Enferm Infecc Microbiol Clin. 2004;22(4):238-45.

76. Beghini F, McIver LJ, Blanco-Míguez A, Dubois L, Asnicar F, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. Elife. 2021;10.

77. Jovel J, Patterson J, Wang W, Hotte N, O'Keefe S, et al. Characterization of the Gut Microbiome Using 16S or Shotgun Metagenomics. Front Microbiol. 2016;7:459.

78. Ranjan R, Rani A, Metwally A, McGee HS, Perkins DL. Analysis of the microbiome: Advantages of whole genome shotgun versus 16S amplicon sequencing. Biochem Biophys Res Commun. 2016;469(4):967-77.

79. Pérez-Cobas AE, Gomez-Valero L, Buchrieser C. Metagenomic approaches in microbial ecology: an update on whole-genome and marker gene sequencing analyses. Microb Genom. 2020;6(8).

80. Usyk M, Peters BA, Karthikeyan S, McDonald D, Sollecito CC, et al. Comprehensive evaluation of shotgun metagenomics, amplicon sequencing, and harmonization of these platforms for epidemiological studies. Cell Rep Methods. 2023;3(1):100391.

81. Peterson D, Bonham KS, Rowland S, Pattanayak CW, Klepac-Ceraj V. Comparative Analysis of 16S rRNA Gene and Metagenome Sequencing in Pediatric Gut Microbiomes. Front Microbiol. 2021;12:670336.

82. Abu-Ali GS, Mehta RS, Lloyd-Price J, Mallick H, Branck T, et al. Metatranscriptome of human faecal microbial communities in a cohort of adult men. Nat Microbiol. 2018;3(3):356-66.

83. Ramos SF, Siguenza N, Zhong W, Mohanty I, Lingaraju A, et al. Metatranscriptomics Uncover Diurnal Functional Shifts in Bacterial Trans-genes with Profound Metabolic Effects. Cell Host Microbe. 2025;33(7):1057-72.

84. Aya V, Pardo-Rodriguez D, Vega LC, Cala MP, Ramirez JD. Integrating metagenomics and metabolomics to study the gut microbiome and host relationships in sports across different energy systems. Scientific Reports. 2025;15(1):19.

85. Weng CC, Suarez C, Cheang SE, Couture G, Goodson ML, et al. Quantifying Gut Microbial Short-Chain Fatty Acids and Their Isotopomers in Mechanistic Studies Using a Rapid, Readily Expandable LC-MS Platform. Anal Chem. 2024;96(6):2415-24.

86. Metzker ML. Sequencing technologies - the next generation. Nat Rev Genet. 2010;11(1):31-46.

87. Song SJ, Amir A, Metcalf JL, Amato KR, Xu ZZ, et al. Preservation Methods Differ in Fecal Microbiome Stability, Affecting Suitability for Field Studies. mSystems. 2016;1(3).

88. Gratton J, Phetcharaburanin J, Mullish BH, Williams HR, Thursz M, et al. Optimized Sample Handling Strategy for Metabolic Profiling of Human Feces. Anal Chem. 2016;88(9):4661-8.

89. Leeming ER, Louca P, Gibson R, Menni C, Spector TD, et al. The complexities of the diet-microbiome relationship: advances and perspectives. Genome Med. 2021;13(1):10.

90. Wesolowska-Andersen A, Bahl MI, Carvalho V, Kristiansen K, Sicheritz-Pontén T, et al. Choice of bacterial DNA extraction method from fecal material influences community structure as evaluated by metagenomic analysis. Microbiome. 2014;2:19.

91. Gorzelak MA, Gill SK, Tasnim N, Ahmadi-Vand Z, Jay M, et al. Methods for Improving Human Gut Microbiome Data by Reducing Variability through Sample Processing and Storage of Stool. PLoS One. 2015;10(8):e0134802.

92. Asnicar F, Thomas AM, Passerini A, Waldron L, Segata N. Machine learning for microbiologists. Nature Reviews Microbiology. 2024;22(4):191-205.

93. Lyu R, Qu Y, Divaris K, Wu D. Methodological Considerations in Longitudinal Analyses of Microbiome Data: A Comprehensive Review. Genes (Basel). 2023;15(1).

94. Eicher T, Kinnebrew G, Patt A, Spencer K, Ying K, et al. Metabolomics and Multi-Omics Integration: A Survey of Computational Methods and Resources. Metabolites. 2020;10(5).

95. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010;7(5):335-6.

96. Zhang L, Zou WY, Hu YY, Wu HH, Gao Y, et al. Voluntary wheel running ameliorated the deleterious effects of high-fat diet on glucose metabolism, gut microbiota and microbial-associated metabolites. Food Science and Human Wellness. 2024;13(3):1672-84.

97. Ojala T, Kankuri E, Kankainen M. Understanding human health through metatranscriptomics. Trends Mol Med. 2023;29(5):376-89.

98. Dettmer K, Aronov PA, Hammock BD. Mass spectrometry-based metabolomics. Mass spectrometry reviews. 2007;26(1):51-78.

99. Holmes E, Li JV, Athanasiou T, Ashrafian H, Nicholson JK. Understanding the role of gut microbiome–host metabolic signal disruption in health and disease. Trends in microbiology. 2011;19(7):349-59.