Metagenomic Insights into the Impact of Nutrition on Human Gut Microbiota and Associated Disease Risk

Abstract

Metagenomic investigation of gut microbiome is a comprehensive and rapid technique for the analysis and diagnosis of numerous diseases. The gut microbiome is an intricate ecosystem, coordinated by the interaction of various microbes and the metabolites produced by them, which helps in developing and sustaining immunity and homeostasis. A healthy gut microbiome is driven by different factors, such as nutrition, lifestyle, etc. The current study examines the association of diet to gut microbiome dysbiosis and its role in various disease conditions. Gut microbiome data was collected from 73 patients and tested at BioAro Inc. lab, using shotgun metagenomics through next generation sequencing. It was then analyzed and compared with data from 20 healthy subjects from HMP database. An in-house bioinformatics pipeline (PanOmiQ) and Pathogen Fast Identifier were utilized for secondary analysis, while tertiary analysis was accomplished using R software. Results showed a higher number of opportunistic pathogen microorganisms in the gut microbiome of subjects consuming a meat diet, as compared to those consuming a plant diet. These opportunistic pathogens included Ruminococcus torques (>3.34%), Ruminococcus gnavus (>2.22%), and Clostridium symbiosum (>1.87%). The study also found a higher relative abundance of these pathogens in cancer patients, as compared to healthy subjects. We also observed a highly significant (p < 0.0001) correlation of a meat diet with obesity in comparison to the subjects on a plant diet and the healthy subjects. Our findings suggest that patients following a plant diet have a lower relative abundance of pathogens that are associated with cancer and obesity. These findings provide critical insight into how we can use shotgun metagenomics to study the composition and diversity of the gut microbiome and the effects of a diet on the gut microbiome and its role in metabolic diseases. This is the first report investigating gut microbiota using shotgun metagenomics, correlating with different diseases and diet followed, which might impact the presence of opportunistic pathogens or keystones species. Additionally, it can provide valuable insights to physicians and dietetic practitioners for providing personalized treatment or customizing a diet plan.

1. Introduction

Emerging progress in metagenomic sequencing has improved microbiome research, with both shotgun and 16S rRNA sequencing playing vital roles. Shotgun metagenomic sequencing enables unbiased, high-resolution profiling of microbial communities, capturing taxonomic and functional diversity across the domains of microbiome. In contrast to 16S rRNA gene sequencing, shotgun metagenomic sequencing offers an enhanced taxonomic resolution at the species and strain levels, along with the ability to profile functional gene content, thereby presenting itself as the method of choice for in-depth characterization of host-associated microbiomes and the investigation of their contributions to human health and disease [1]. The convoluted association between diet and human health has been a subject of extensive research, with recent focus shifting toward understanding how dietary habits influence the composition and functionality of the microorganisms in the gut [2]. The gut microbiome, a multifaceted ecology of microorganisms in the gastrointestinal tract, plays a critical role in maintaining host health by modulating various physiological processes. Emerging evidence suggests that the dietary patterns profoundly influence the gut microbiome composition, in turn impacting the host’s susceptibility to diseases and overall wellbeing [3,4,5].

The human gut is a shelter for trillions of microorganisms, which include bacteria, viruses, fungi, and archaea [5]. Nutritional components of the diet serve as substrates for their metabolism, influencing the abundance and diversity of gut microbial communities. For instance, diets rich in fiber promote the growth of beneficial gut bacteria, like Bifidobacterium sp. and Lactobacillus sp., whereas high-fat diets have been associated with a reduction in beneficial microorganisms or keystone species (microorganisms helping in maintaining the gut microbiome ecosystem). Moreover, high-fat diets may lead to an increase in pathogenic bacteria in the gut, specifically belonging to the phylum of Firmicutes (Ruminococcus spp. and Streptococcus spp.) [6,7].

The recent evidence suggests that modifications in gut microbiome composition due to nutritional factors can influence the growth and development of various chronic diseases [8,9]. Chronic conditions such as obesity, metabolic syndrome, inflammatory bowel diseases (IBD), cancer, and cardiovascular diseases have been associated with dysbiosis, which refers to an imbalance or disruption in the composition and metabolic capacity of the gut microbiota [10,11]. For instance, the consumption of the modern lifestyle diet, characterized by high fat, sugar, and processed foods, has been associated with an increased abundance of pathogenic species causing various diseases. This reflects an increased relative abundance of pathogens or microbial biomarkers associated with obesity, inflammation, cardiovascular disease, and insulin resistance—key risk factors contributing to chronic disease development [12,13].

Modification of the gut microbiome by dietetic changes opens avenues for targeted dietary interventions to promote health. Dietary modifications, such as increasing fiber intake, consuming fermented foods rich in probiotics, and adopting a diet rich in fruits, vegetables, and whole grains, have been shown to positively impact gut microbiome composition and function [14]. These dietary interventions have been associated with reduced inflammation, improved metabolic health, and enhanced immune function, highlighting the potential of diet–microbiome interactions in promoting overall wellbeing [15,16]. Abnormal changes in the gut microbiome (dysbiosis) impact host health and lead to various health problems. Moreover, databases like GMrepo focus on curated human gut microbiome data, emphasizing disease biomarkers and facilitating cross-dataset comparisons to identify consistent and inconsistent disease-associated microbial markers across various health conditions [17]. Additionally, a few bacteria like Clostridium symbiosum, Ruminococcus gnavus, Ruminococcus torques, Fusobacterium nucleatum, and Clostridium colicanis have been proposed as indicative markers in diseases such as obesity, colorectal cancer, and gastric cancer, highlighting the significance of gut microbial markers in disease diagnosis and understanding the relationship between gut microbiota and human health [18]. These published reports also highlight that the management of the gut microbiome through the use of probiotics, prebiotics, and synbiotics can help restore the balance of the microbiota and promote health [19,20,21].

Thus, in the present study, we delve into the available evidence elucidating the impact of diet on the gut microbiome, disease microbial markers, and health outcomes. The study aims to provide insights into the complex interplay between diet, gut microbiome, and human health. This is the first study comparing gut microbiota of human subjects following different diets and its association with different diseases, i.e., cancer or obesity. Understanding these relationships is essential for developing personalized dietary strategies aimed at optimizing gut microbiome composition and mitigating the risk of chronic diseases.

2. Materials and Methods

2.1. Sample Data Collection and Processing

This research was conducted in strict compliance with the ethical guidelines of the Health Research Ethics Board of Alberta (HREBA.CHC-25-0013). Informed consent was duly obtained from all participants, with robust measures implemented to ensure the protection of their privacy and confidentiality. As part of the BioGut program, clinical samples were received at the Genomics Facility of BioAro Inc. for next generation sequencing (NGS), and the results were archived in the BioAro Inc. data repository with anthropometric measures and the medical history of the participants. In this retrospective study, data were collected from the repository of BioAro Inc. for 73 participants with different age groups ranging from 10 to 80 years (mostly adults). Participant information, including diet, age, Body Mass Index (BMI), and gender, is clearly summarized in Supplementary Table S1. The table clearly defines the diet, age, BMI, and gender of the participants. The inclusion criteria in the study were availability of dietary information and informed consent, whereas exclusion criteria encompassed subjects on antibiotics medication or pregnant and breastfeeding mothers. For the purpose of our diet-based comparison study, we applied the Common Quantitative Standards to categorize and standardize dietary intake data across participants. This ensured that all diet-related variables were evaluated using consistent measurement units and classification criteria, allowing for reliable comparisons between individuals and groups [22,23,24]. Additionally, a total of 20 healthy adult subjects from the human microbiome project (HMP) site were selected as controls, with Body Mass Index (BMI) values ranging between 18.5 and 35 kg/m². Exclusion criteria included individuals currently receiving antibiotic treatment, pregnant or breastfeeding women, and those diagnosed with chronic gastrointestinal disorders such as Inflammatory Bowel Disease (IBD), Crohn’s disease, ulcerative colitis, or irritable bowel syndrome (IBS).

Stool samples for bio–gut analysis from participants were received at the BioAro Inc, Calgary, Alberta AB T2P 0L4, Canada. genomic facility and then frozen for storage at −80 °C for further processing. Microbial DNA was extracted from samples using the Zymo BIOMICS DNA Miniprep Kit (Zymo Research, Richmond, BC, USA, Cat. No. D4300) following the manufacturer’s instructions. Briefly, the protocol involved cell lysis using enzymatic digestion and bead beating, followed by DNA purification using the spin column. Purified DNA was eluted in DNase-free water [Figure 1].

2.2. Library Construction and Sequencing

The extracted DNA from each sample was quantified using a Qubit fluorometer [25] according to the manufacturer’s instructions. Library and sequencing was performed on the MGI platform (MGI Tech Co., Ltd., Shenzhen, China), a subsidiary of Complete Genomics, which specializes in high-throughput sequencing technologies. Each DNA sample was diluted to a pre-determined concentration (e.g., 10 ng/µL) using Tris-EDTA (TE) buffer to ensure equal representation in the library. This was followed by the fragmentation of DNA and the ligation of sequencing adapters containing MGI rapid-sequencing flow cells, complementary sequences, and unique barcode sequences to the DNA ends using T4 DNA Ligase in the MGIEasy FS DNA library prep kit, according to the manufacturer’s protocol. These adapters allow for library attachment to the sequencing flow cell and sample identification during sequencing. The prepared library was quantified using size-selection methods (e.g., gel electrophoresis or magnetic beads) and a Qubit fluorometer [25]. This ensures the sequencing of the desired library fragments. The fragment size distribution used an automated capillary electrophoresis system like Agilent Tapestation [26]. This step ensures the quality and quantity of the library for further sequencing process steps and verifies fragment size (350–500 bp) suitability for the MGI sequencing platform.

Further, the samples from the library were pooled together by combining the aliquots of the normalized library DNA from all microbiome samples into a single tube. The volume of each aliquot was proportional to the desired representation of each sample in the final sequencing data. The circular single-stranded DNA molecules were prepared by enzymatic circularization. Following the manufacturer’s instructions, this circularization method was completed using commercially available MGIEasy circularization module reaction kits. The circularized DNA were diluted to the recommended concentration for DNA Nano Ball (DNB) preparation and loaded onto the DNBSeq-G400RS Sequencing Flow Cell by following the manufacturer’s instructions. Finally, the paired-end shotgun sequencing was performed using the sequencing platform (MGI DNBSEQ-G400) to obtain reads from both ends of the DNA fragments. Shotgun sequencing was preferred over 16S sequencing as it enables the detection of a variety of microorganisms at the species level. The sequencing procedure was regulated by the addition of positive and negative controls in each run [Figure 1].

Concurrently, 16S sequencing and shotgun metagenomic sequencing are the most extensively utilized methodologies for the taxonomical profiling of microorganisms that offer diverse molecular and analytical advantages. The 16S sequencing depends on the amplification of homologous regions of the 16S ribosomal RNA gene for the genus-level identification of microbial taxonomy [27]. Conversely, shotgun metagenomic sequencing amplifies every DNA present in the sample, including the host genome, and identifies species- and strain-level microbial taxonomy. It also aids in the functional annotation of the genes involved in antibiotic resistance, virulence, and other metabolic processes [1,28].

2.3. Taxonomy Classification

The raw reads were processed through the in-house bioinformatics pipeline PanOmiQ developed at BioAro Inc., along with Pathogen Fast Identifier (PFI). It utilizes databases, which are rapidly curated and are a comprehensive platform for taxonomy identification and classification [29]. It has more than 27,000 microbial genomic DNA sequences for rapid identification. Initially, the raw reads were analyzed for their quality; low-quality reads and adaptors were trimmed. Further, host DNA and rRNA were removed from the sequences by assembling them with the human reference genome. The taxonomy classification was performed for species identification.

2.4. Statistical Analysis

The tertiary analysis with R statistical software package (v4.4.1) was utilized to perform statistical analysis. It explored potential variations in microbial community composition between the different patient groups. Non-parametric Fisher’s exact test was performed to calculate p < 0.05 with 95% confidence interval.

2.5. Virulence Gene Analysis

Virulence gene analysis was performed on all the samples using the MetaVF toolkit with default parameters. MetaVF [30] identifies virulence genes by aligning sequencing data to curated virulence factor databases. Detected virulence genes were linked to their respective microbial species detected in the samples. Network visualization was carried out using Cytoscape 3 [31].

3. Results and Discussions

Gut microbiome, as a characteristic human-associated niche, can be comprehensively analyzed using shotgun metagenomic sequencing to elucidate the host–microbial interactions. Effective computational methods aid in filtering host DNA contamination and enable accurate downstream analyses [32]. Despite the cost-efficiency of 16S sequencing, which makes it ideal for broad surveys and ecological comparisons, shotgun metagenomics delivers a more multi-faceted and exhaustive analysis of microbial communities, including functional characterization and species- or strain-level resolution [33]. Our shotgun metagenomic sequencing analysis found the bacterial fraction in different gut microbiome samples. The clean reads were mapped to the human reference genome for host sequence removal before taxonomic classification. The taxonomical composition and relative abundance of gut microbiota of plant and meat eaters have varied significantly. The gut microbiome of healthy controls, plant diet consumers, and meat diet consumers were dominated by four phyla: Bacteroidetes, Firmicutes, Actinobacteria, and Proteobacteria. The genus-level distribution of top 50 species is given in Figure 2. The species-level comparison of plant diet (PD) and meat diet (MD) consumers reveals an increased intestinal bacteria load in MD compared to PD consumers. There were 228 common bacterial species found in both groups, whereas 17 and 120 unique bacterial species were observed only in PD and MD, respectively [Figure 3]. Some of these unique species are considered to be opportunistic pathogens, such as Actinomyces spp. and Corynebacterium spp. Our analysis revealed that the MD consumer group had apparently a higher number of species, including pathogens such as Ruminococcus torques (>3.34%), Ruminococcus gnavus (>2.22%), and Clostridium symbiosum (>1.87%), than that of the PD group.

3.1. Opportunistic Pathogen Species and Disease Risk

Based on the anthropometric measurements of participants, each Body Mass Index (BMI) was calculated. Samples were categorized into normal weight, moderately obese, severely underweight, and severely obese on the basis of BMI. Based on the healthy history of the participants, the samples were also categorized into cancer and normal. The species-level comparison was also analyzed for different types of obese and normal samples, which revealed that 54 microbial species were commonly found in all categories. A total of 53, 27, 26, and 3 unique microorganisms were detected in normal, moderately obese, severely obese, and severely underweight samples, respectively [Figure 4]. The Fisher’s exact test performed using R software with 95% confidence interval for PD and MD group pathogens revealed a statistically significant observation (p = 0.02) in both the number and relative abundance of pathogens between the two categories. The presence of pathogens was compared among three groups, healthy control and PD and MD consumers. Additionally, pathogen comparison was also carried out among healthy control, obese patients, and cancer patients. It is observed that 24 pathogens were found in obese patients, followed by MD (22 pathogens), PD (17 pathogens), cancer subjects (13 pathogens), and healthy control subjects (5 pathogens) [Figure 5].

Non-metric multi-dimensional scaling analysis was performed using Phyloseq package in R software [34]. Figure 6 represents the Bray–Curtis dissimilarity among the samples, in which most of the MD group gut samples fall above 0.2, indicating the differences in species abundance compared with PD, with a p value of (0.005%). It is also observed that a higher abundance of R. gnavus (>2.22%) and R. torques (>3.34%) existed in most MD samples. R. gnavus is one of the microbial markers for various diseases such as cardiovascular disease (CVD), obesity, IBD, and mental health [35]. R. torques and R. gnavus contribute to IBD progression. Pathogens like C. symbiosum and Clostridium innoccum are common in both the PD and MD groups but are not found in healthy subjects. A reduced abundance of the species Akkermansia muciniphila was observed (0–0.15%) in obese patients. The relative abundance of irritable bowel syndrome (IBS) microbial markers such as Bifidobacterium spp., Lactobacillus spp., Veillonella spp., Ruminococcus spp., Clostridiales spp., and Prevotella spp. reveals a significant difference (non-significant in similarity) among MD, PD, and healthy subjects [Figure 7]. These species help in improving insulin sensitivity and regulate glucose hemostasis [36,37]. The Wilcoxon non-parametric test was employed to evaluate the statistically non-significant variation in the abundance of IBS-associated microbial markers between MD and PD consumer groups. The study also revealed that the relative abundance of beneficial bacterial families such as Lactobacillaceae (PD: 0.5–1.46%; MD: 0–0.02%) and Bifidobacteriaceae (PD: 2.97–14.43%; MD: 0.027–1.31%) was reduced, whereas Prevotella spp., Ruminococcus spp., and Clostridium spp. were increased in MD samples. Both these beneficial species have anti-inflammatory properties and help in maintaining intestinal metabolic health. Lactobacillus spp. helps inhibit the growth of certain pathogens, such as Staphylococcus aureus, Escherichia coli, and Clostridium difficile, in the gut microbiome [38]. Bifidobacterium spp. facilitates fiber digestion and produces saturated fatty acids and B vitamins. A lower abundance of these species might lead to diseases like gastrointestinal infections, gastric cancer, and colon neoplasms [39,40]. Moreover, the decreased presence of probiotic species belonging to the genus Lactobacillaceae has been associated with increased anxiety and altered cognitive behaviors [41]. The present analysis also detected bacteria belonging to genera Bacteroides spp., Roseburia spp., Eubacterium spp., and Methanobrevibacter spp. in the PD samples. The presence of the aforementioned bacteria was due to the consumption of foods like pulses (beans), and these might be responsible for somatic pain [42].

3.2. Short-Chain Fatty Acid Production

Short-chain fatty acid (SCFA)-producing bacteria were analyzed in both PD and MD samples. Genera known for SCFA production such as Faecalibacterium, Bifidobacterium, Coprococcus, Roseburia, and Akkermansia play a crucial role in maintaining gut health and overall metabolic balance. These bacteria ferment dietary fibers and resistant starches in the colon to produce SCFAs such as acetate, propionate, and butyrate. Additionally, they also influence glucose and lipid metabolism, support gut–brain signaling, and play a protective role against obesity, type 2 diabetes, colorectal cancer, and other chronic diseases. The presence and abundance of SCFA-producing bacteria are therefore considered to be a beneficial species of a healthy gut microbiome and are often associated with a reduced risk of metabolic and inflammatory disorder. A paired t-test comparing the relative abundance of these SCFA-producing bacteria in PD and MD groups yielded a t-value of 1.76 (df = 4, p = 0.153), indicating that the observed difference in mean abundance was not statistically significant at the 0.05 level. The t-test addresses differences in central tendency, whereas the correlation measures concordance in the samples of plant and meat diet observations. The significant correlation alongside a non-significant t-test reflects high relative agreement between groups without sufficient evidence for a systematic shift. The 95% confidence interval for the mean difference ranged from −1.51 to 6.77, suggesting substantial variability and insufficient evidence to reject the null hypothesis. However, a Pearson correlation analysis revealed a significantly higher abundance of SCFA-producing bacteria in PD samples compared to MD, with a correlation coefficient of r = 0.8814 (p = 0.04812) and a 95% confidence interval [Figure 8].

3.3. Virulence Genes Analysis

We performed a MetaVF [30] analysis on the meat- (MD) and plant-based (PD) groups, and the results were sorted with stringent thresholds of ≥80% coverage and ≥3× depth to identify high-confidence microbial virulence factor (VF) genes. The network visualization in Cytoscape 3 [31] provided a comprehensive view of VF gene distribution across samples [Figure 9]. Nodes representing samples and microbial species were interconnected based on shared VF genes. Additionally, node and edge colors represent the number of neighborhood connectivity and depth of virulence species, respectively. Edges represent the depth of the virulent species. The comparative analysis of virulence gene profiles between PD and MD samples revealed distinct microbial signatures. PD samples exhibited a higher prevalence of genes associated with commensal and fiber-degrading bacteria, such Streptococcus thermophilus, which are commonly linked to beneficial gut functions, including SCFA production and mucosal barrier support [43,44]. These findings are consistent with the higher intake of dietary fiber in PD individuals, which promotes the growth of symbiotic and metabolically favorable microbes [6].

In contrast, MD samples demonstrated an increased abundance of VF genes associated with opportunistic and pro-inflammatory bacteria, including Escherichia coli, Shigella spp., and Staphylococcus aureus. Many of these VF genes encode for factors such as adhesins, hemolysins, and toxins, which are implicated in epithelial invasion and immune activation. This pattern suggests a potentially higher pro-inflammatory microbial environment in the gut of MD individuals, possibly linked to lower fiber intake and increased availability of protein and fat substrates [45,46,47]. Particularly, in a subset of PD samples, no virulence genes were detected based on our predefined detection thresholds in the MetaVF toolkit. This absence may reflect a truly low virulence gene burden in these individuals, possibly due to the dominance of non-pathogenic, fiber-utilizing microbes. Alternatively, it may also indicate variability in microbial composition or sequencing depth. The lack of noticeable VF genes in these samples further supports the hypothesis that plant-based diets may foster a gut microbiome with reduced pathogenic potential.

One of the limitations of our study is the inability to perform pathway analysis due to the small cohort size. Functional profiling requires adequate statistical power to account for inter-individual variability, which is challenging with limited samples. This is particularly important for diet-based comparisons, where microbial functional differences may only emerge in larger populations. Future studies with expanded cohorts will be needed to generate robust pathway-level insights.

4. Conclusions

In summary, the pilot study investigated the gut samples of the subjects suffering from different diseases and following varying dietary patterns by using advanced shotgun metagenomic sequencing analysis. Taxonomical classification and relative abundance calculation was performed using PanOmiQ software https://panomiq.com/. The secondary analysis revealed that more than 100 unique species were detected only in the MD samples when compared with PD and healthy subjects. Among those unique species, 22 were pathogenic microorganisms. Additionally, the abundance of pathogenic microorganisms, i.e., R. gnavus and R. torques, were higher in MD samples in comparison to other groups. We also observed the presence of C. symbiosum and C. innoccum in most of the samples, and a higher abundance of these species was found in cancer patients and obese people. Overall, this is the first study depicting dietary factors influencing the presence of opportunistic pathogens in gut microbiome and perhaps leading to cancer and obesity. Pathogenic microorganisms are less common when a plant diet is consumed. The study provides insights into personalized dietary practices based on one’s health condition.