Method for determining alteration of gut microbiome

Abstract

The present invention relates to a method for determining whether an individual has or is at risk of a change in intestinal bacterial diversity.

Description

Technical Field

[0001] The present invention relates to a method for determining whether an individual has or is at risk of having an alteration in the diversity of intestinal bacteria. Background Art

[0002] It is estimated that the number of microorganisms inhabiting the human gastrointestinal tract equals or exceeds that of host cells. Some studies even suggest that they can outnumber host cells by 10 to one and outnumber host genes by more than 100 to one. These microorganisms associated with the digestive tract are commonly referred to as the gut microbiota, a complex and dynamic ecosystem comprising thousands of bacterial species, including commensal, beneficial, and pathogenic bacteria.

[0003] The gut microbiome is often compared to an organ due to its specific biochemical interactions with the host and its systemic integration into host biology. Like any other organ, the human gut microbiome plays a crucial role in health and disease. While a normal state of the microbiome appears to ensure homeostasis, an imbalance or partial loss of this microbiome has been associated with numerous diseases, including obesity, inflammatory bowel disease (IBD), irritable bowel syndrome (IBS), type 2 diabetes, non-alcoholic liver disease, cardiometabolic diseases, allergies, and malnutrition.

[0004] Numerous factors, both exogenous and endogenous, influence the composition of the gut microbiota. These include diet, host genotype, antibiotic history, age, and sex. However, recent studies suggest that environmental and host factors explain less than 20% of the variation in microbial composition, suggesting that stochastic factors and ecological imperatives play a significant role in gut microbial assembly. A factor known to influence species distribution and ecosystem diversity is the history of community assembly. Indeed, the order in which members appear in an ecosystem can influence its evolution, a phenomenon known as the priority effect, which has been studied in animal models but remains underexplored in humans due to the lack of large-scale, longitudinally intensive datasets. In summary, multiple factors contribute to the substantial inter-individual variability of the human gut microbiome across the human lifespan. This substantial inter-subject variability, combined with its individual stability, provides evidence for a better understanding of the underlying ecological signatures to guide microbiome-based preventive or therapeutic approaches.

[0005] However, assessing and understanding gut bacterial diversity remains extremely difficult due to the lack of reliable scientific tools or comprehensive methodologies applicable to most people. Summary of the Invention

[0006] The present invention arises from the unexpected discovery by the inventors that the ratio of Faecalibacterium to Bacteroidetes is associated with the diversity of the intestinal microbiota and can be used as a relevant biomarker to detect or predict changes in the intestinal microbiota.

[0007] Therefore, the present invention provides a method for determining whether an individual has or is at risk of having an alteration in gut bacterial diversity by determining the ratio of Faecalibacterium population / Bacteroides population in a microbial sample from the individual.

[0008] Such determinations are useful, inter alia, for assessing whether an individual is at risk for developing conditions such as obesity, inflammatory bowel disease (IBD), irritable bowel syndrome (IBS), type 2 diabetes, non-alcoholic liver disease, cardiometabolic disease, allergies, and malnutrition. DETAILED DESCRIPTION

[0009] As used herein, the term "subject" shall mean a mammal, preferably any human subject. In the context of the present invention, the subject may be healthy or unhealthy. The subject may be an adult, an elderly person, a child or a newborn.

[0010] As used herein, the expression "alteration in gut bacterial diversity" describes a disruption among the species of organisms present in an individual's gut microbiota. Such alterations may lead to intestinal dysbiosis, defined as an imbalance in the composition of the microbiota, which is associated with the pathogenesis of intestinal and extraintestinal diseases.

[0011] As used herein, "microbiome" refers to the microbial flora found in the gut. In microbiology, flora refers to the collective term for the bacteria and other microorganisms found in an ecosystem (e.g., a part of the body of an animal host). The "gut microbiome" consists of all the species found in an individual's gut.

[0012] As used herein, the term "relative abundance" refers to the relative quantification of multiple taxonomic characteristics of a microbial population, for example, expressed as a ratio or %, for example, taxon A 50%, taxon B 50% (the taxa are equally abundant); taxon A 40%, taxon B 30%, taxon C 30% (taxon A, the most abundant population).

[0013] As used herein, the term "Faecalibacterium" refers to the bacterial genus Faecalibacterium. Its only known species is Faecalibacterium prausnitzii. z ii).

[0014] As used herein, the term "Bacteroides" refers to the bacterial genus Bacteroides. Species of this genus are Bacteroides fragilis, Bacteroides vulgatus, and Bacteroides stercoris, particularly Bacteroides fragilis.

[0015] The "ratio of Faecalibacterium population / Bacteroides population" refers to the number of bacteria belonging to the genus Faecalibacterium in the microbiome (or in a microbiome sample) divided by the number of bacteria belonging to the genus Bacteroides in the microbiome (or in a microbiome sample). The number of bacteria belonging to the genus Faecalibacterium or Bacteroides can be estimated or quantified by techniques well known to those skilled in the art. Such techniques include, but are not limited to, PCR amplification, sequencing (such as 16S RNA), or shotgun metagenomics present in a microbiome sample.

[0016] As used herein, a "microbiota sample" refers to any biological sample that can be used to detect the presence and composition of the intestinal microbiota.

[0017] In one aspect, the present invention provides a method for determining whether an individual has or is at risk for an alteration in gut bacterial diversity by determining the relative abundance of Faecalibacterium and Bacteroides in a microbiota sample.

[0018] In an embodiment, the method comprises:

[0019] (i) determining the ratio of Faecalibacterium population to Bacteroides population in a microbiota sample from the individual,

[0020] (ii) If the ratio of Faecalibacterium population / Bacteroides population is less than or equal to 1 / 4, the individual is determined to have or be at risk of having an alteration in intestinal bacterial diversity.

[0021] In embodiments, Faecalibacterium and Bacteroides populations are the two most abundant populations in an individual's gut microbiome. Indeed, as the inventors have shown, most of the changes in gut bacterial diversity occur in populations enriched in Bacteroides, with Faecalibacterium and Bacteroides species being the most heavily weighted.

[0022] In a preferred embodiment, Bacteroides is the most abundant species in the individual's gut microbiome as determined by relative abundance.

[0023] According to the present invention, "the most abundant species" means the dominant genera among all genera that are usually detected.

[0024] In embodiments, Bacteroides is the most abundant genus species compared to other commonly detected genera (e.g., Faecalibacterium, Roseburia, Prevotella, Akkermansia, Bifidobacterium, Blautia, Coprococcus, Ruminococcus, Alistipes, Parabacteroides, Clostridium, and Lachnospira).

[0025] The abundance of each bacterial genus and species can be determined by various techniques commonly used by those skilled in the art, such as DNA or RNA sequencing.

[0026] Therefore, an embodiment of the present invention relates to a method as defined above, comprising a step prior to determining the composition of the gut microbiome.

[0027] In particular, embodiments of the present invention relate to a method as defined above, comprising the following prior steps: a) quantifying the genus Bacteroides and at least one or more species selected from the group consisting of Faecalibacterium, Roseburia, Prevotella, Akkermansia, Bifidobacterium, Blautia, Coprococcus, Ruminococcus, Alternaria, Parabacteroides, Clostridium, and Lachnospira; and b) determining the relative abundance of each genus species. In a preferred embodiment, the method as defined above is performed only when Bacteroides is the most abundant genus species.

[0028] In an embodiment, the present invention relates to a method as defined above, further comprising:

[0029] (iii) providing dietary advice to an individual at risk of altered gut bacterial diversity that is suitable for increasing the ratio of Faecalibacterium populations to Bacteroides populations, or administering a diet or treatment to the individual to promote restoration of gut bacterial diversity.

[0030] In an embodiment, the present invention relates to a method as defined above, further comprising:

[0031] (iii) administering to the individual:

[0032] - a diet rich in fiber,

[0033] - a diet containing one or more probiotic species,

[0034] - food supplements containing bacteria of the genus Faecalibacterium,

[0035] - a diet rich in Faecalibacterium bacteria,

[0036] -- A diet rich in riboflavin (vitamin B2), and / or

[0037] - Treatments to reduce oxidation and / or inflammation.

[0038] In an embodiment, the present invention relates to a method as defined above, further comprising:

[0039] (iii) administering a fiber-rich diet to the individual.

[0040] In an embodiment, the present invention relates to a method as defined above, further comprising:

[0041] (iii) administering to the individual a diet comprising one or more probiotic species.

[0042] In an embodiment, the present invention relates to a method as defined above, further comprising:

[0043] (iii) administering to said individual a food supplement comprising a bacterium of the genus Faecalibacterium,

[0044] In an embodiment, the present invention relates to a method as defined above, further comprising:

[0045] (iii) administering to the individual a diet enriched in Faecalibacterium bacteria.

[0046] In an embodiment, the present invention relates to a method as defined above, further comprising:

[0047] (iii) administering a diet rich in riboflavin (vitamin B2) to the individual.

[0048] In an embodiment, the present invention relates to a method as defined above, further comprising:

[0049] (iii) administering to the individual a therapy for reducing oxidation and / or inflammation.

[0050] In an embodiment, the invention relates to a method as defined above, wherein the diet enriched in fiber is a diet enriched in inulin, pectin and / or fructooligosaccharides (FOS).

[0051] In an embodiment, the invention relates to a method as defined above, wherein the diet administered to the individual comprises a fermented dairy product or a plant-based milk substitute.

[0052] In an embodiment, the present invention relates to a method as defined above, wherein said probiotic is selected from the group consisting of: Streptococcus, Lactococcus and Bifidobacterium.

[0053] In an embodiment, the present invention relates to a method as defined above, wherein the probiotic is selected from the group consisting of: Streptococcus thermophilus, Lactococcus lactis subsp. lactis, Bifidobacterium lactis, Bifidobacterium adolescentis, Bifidobacterium longum, Bifidobacterium breve, Bifidobacterium Bifidum, Bifidobacterium pseudocatenulatum, Lacticaseibacillus rhamnosus and Lacticaseibacillus paracasei.

[0054] In an embodiment, the present invention relates to a method as defined above, wherein said probiotic is Streptococcus thermophilus.

[0055] In an embodiment, the present invention relates to a method as defined above, wherein the probiotic is Lactococcus lactis subsp. lactis.

[0056] In an embodiment, the present invention relates to a method as defined above, wherein said probiotic is Bifidobacterium lactis.

[0057] In an embodiment, the present invention relates to a method as defined above, wherein the probiotic is Bifidobacterium adolescentis.

[0058] In an embodiment, the present invention relates to a method as defined above, wherein the probiotic is selected from the group consisting of: Streptococcus thermophilus CNCM 1-3862, Lactococcus lactis subsp. lactis CNCM 1-1631 and Bifidobacterium lactis CNCM I-2494.

[0059] CNCM I-1631 refers to the strain deposited on October 24, 1995, at the Collection Nationale de Cultures de Microorganismes (CNCM) (Institut Pasteur, 25-28 Rue du Docteur Roux, 75724 Paris Cedex 15, France) in accordance with the Budapest Treaty, with the reference number CNCM I-1631.

[0060] CNCMI-2494 refers to the strain deposited with the CNCM on June 20, 2000 under the Budapest Treaty, with the reference number CNCMI-2494.

[0061] CNCMI-3862 refers to the strain deposited with the CNCM on October 31, 2007 under the Budapest Treaty, with the reference number CNCMI-3862.

[0062] In an embodiment, the present invention relates to a method as defined above, wherein the individual has or is at risk of an alteration in the diversity of intestinal bacteria if the ratio of the Faecalibacterium population / the Bacteroides population is lower than or equal to 1 / 5, 1 / 10, 1 / 20, 1 / 30, 1 / 40, 1 / 50, 1 / 60, in particular lower than or equal to 1 / 64.

[0063] In an embodiment, the present invention relates to a method as defined above, wherein the individual has or is at risk of an alteration in gut bacterial diversity if the ratio of Faecalibacterium population / Bacteroides population is ≤ 1 / 4 and > 1 / 64.

[0064] It is expected that dietary intervention will help restore intestinal diversity when the ratio of Faecalibacterium to Bacteroides is ≤ 1 / 4 and > 1 / 64. For example, dietary intervention can be, but is not limited to, administering to an individual a diet comprising one or more probiotics, and / or a diet rich in Faecalibacterium bacteria and / or rich in fiber and / or rich in riboflavin (vitamin B2).

[0065] In an embodiment, the present invention relates to a method as defined above, wherein the individual has or is at risk of an alteration in gut bacterial diversity if the ratio of Faecalibacterium population / Bacteroides population is ≤ 1 / 64.

[0066] It is expected that when the Faecalibacterium / Bacteroides ratio is ≤ 1 / 64, a treatment may be needed to restore intestinal diversity. For example, the treatment may be, but is not limited to, administering to the individual a therapy to reduce oxidation and / or inflammation and / or antibiotics that target Bacteroides bacteria.

[0067] The microbiota sample may be a stool sample or a mucosal sample, preferably a stool sample.

[0068] In an embodiment, the invention relates to a method as defined above, comprising the step of isolating the microbiota sample prior to step (i). Isolation of the microbiota sample can be achieved by a non-invasive procedure, such as collecting a stool sample, or a more invasive procedure, such as collecting a mucosal sample by colonoscopy or biopsy.

[0069] In an embodiment, the present invention relates to a method as defined above, comprising the step of analyzing the microbiota sample before step (i). Analysis of the microbiota sample can be achieved by various techniques, for example, PCR amplification of genomic sequences present in the sample or by plating on a selective medium.

[0070] All embodiments and features given above are referenced to apply to the other following aspects of the invention.

[0071] In another aspect, the present invention also relates to the use of the ratio of Faecalibacterium population to Bacteroides population in a microbiota sample as a biomarker for determining whether an individual has or is at risk of having altered intestinal bacterial diversity.

[0072] In another aspect, the present invention therefore provides a method of determining whether an individual has or is at risk of intestinal dysbiosis by determining the relative abundance of Faecalibacterium and Bacteroides in a microbiota sample, the method comprising:

[0073] (i) determining the ratio of Faecalibacterium population to Bacteroides population in a microbiota sample from the individual,

[0074] (ii) If the ratio of Faecalibacterium population / Bacteroides population is less than or equal to 1 / 4, the individual is determined to have intestinal dysbiosis or to be at risk of intestinal dysbiosis.

[0075] In another aspect, the present invention also relates to the use of the ratio of Faecalibacterium population to Bacteroides population in a microbiota sample as a biomarker for determining whether an individual suffers from or is at risk of intestinal dysbiosis.

[0076] The present invention will be further illustrated by the following non-limiting figures and examples. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 The American Gut Project groups partitions based on similarity. Each partition centroid is mapped to a branch.

[0078] Figure 2 Longitudinal analysis of DMM partitions within branches. a. The network between partitions shows how individuals change partitions over time. Node size indicates the stability of the partition, and the strength reflects the corresponding branch. Arrow width indicates the proportion of change between each node. To reduce visual noise, only edges representing more than 5% of partition events (including switches and non-switches) are shown; these edges collectively represent 81% of the events.

[0079] Figure 3 The density of the Faecalibacterium:Bacteroides ratio within the Bacteroides-rich clade. The vertical black line shows the average Faecalibacterium:Bacteroides ratio for AGP participants who switched to another clade over time (h), and the Faecalibacterium:Bacteroides ratio as a function of the Bacteroides-rich clade.

[0080] Figure 4 Figure 2. Faecalibacterium:Bacteroides ratio in gut microbiome samples from AGPs as a function of gut microbiome alpha diversity (Shannon index); samples were partitioned into clades enriched in Bacteroides. Microbiome DMM partitions were classified according to the Faecalibacterium:Bacteroides ratio threshold (dashed line) defined by longitudinal data analysis.

[0081] Figure 5 Individual characteristics of partition M1. The bar chart shows the proportion of participants who selected that AGP questionnaire answer, compared with those who chose the M1 partition of the gut microbiome, among all participants with AGP. Partition M1, representing the highest alpha diversity and the structural root of the branch, was used as the reference for multinomial logistic regression analysis.

[0082] Figure 6Absolute model weights for predictor variables used in models describing partitions dominated by Bacteroides . Partitions were clustered according to the Faecalibacterium:Bacteroides ratio. Heatmap displays the log odds ratios (ORs) of predictor variables as a function of partition. Significant ORs are indicated by black dots (FDR < 10%). Dichotomous predictor variables (e.g., self-reported disease status) were coded from 0 to 1 (e.g., "diagnosed by a healthcare professional" was coded as 1, "I do not have this condition" was coded as 0, and other categories were treated as missing values). Color intensity > 0.0 to 2.5 indicates that the partition was associated with a higher frequency of disease. For diet, such as food group frequency, it was coded from 0 (never) to 1 (daily). Color intensity < 0.0 to -2.5 indicates that the partition was associated with a lower frequency of antibiotic intake. Antibiotic history was considered an ordered factor ranging from "I have not taken antibiotics in the past year" to "the week" of stool sample collection. Color intensity > 0.0 to 2.5 indicates that the partition was associated with a shorter time since the last antibiotic dose.

[0083] Example

[0084] method

[0085] American Gut Project (AGP) dataset

[0086] In the AGP project, stool samples were collected at home and transported at room temperature, followed by microbial DNA extraction and 16S rRNA amplicon sequencing, all performed according to previously described methods (McDonald et al., mSystems, 3:e00031-18, 20188). The inventors used redbiom (McDonald et al., mSystems, 4:e00215-e00219, 2019) to obtain data from Qiita (Gonzalez et al., Nature methods, 15:796-98, 2018). As of December 5, 2019, the database had 20,454 stool sample identifiers available under the Deblur-Illumina-16S-V4-100nt-fbc5b2 analysis environment. Analysis was performed according to previously described methods (Cotillard et al., 2021). In brief, bioinformatics analysis was performed using QIIME 2019.10, bloom sequences were deleted as previously described (Amir et al., mSystems, 2:e00199-16, 2017), and the taxonomy was specified using the GreenGenes database (v 13.5). The inventors retained samples with >= 1000 reads. According to the criteria in Cotillard et al., 2021, 1579 samples were defined as technical outliers and excluded. The genus count matrix containing 16021 samples was analyzed.

[0087] Compartmentalization and branching of the gut microbiome

[0088] In AGP, 30 genera with the highest read volume were extracted for downstream analysis. Using the Dirichlet's multinomial mixture (DMM) model established based on microbiome data, the samples were partitioned (Holmes et al., PloS one 7:e30126,2012). The inventors used five subsets in the entire data set to train the DMM model to reduce population bias. Each subset was sampled from the entire data set and stratified by each combination of gender, birth geographic region (for the AGP data set, birth region) and age group (by layer, n is up to 30). Subsampling is not limited to one sample per subject. It is worth noting that the proportion of participants with at least two samples in the participants is less than 5%. DMM modeling is performed on each subset that constitutes our training set, and the best model is selected using BIC, Laplace minimum (Laplace minima) and majority voting principle (that is, the optimal value of the k parameter is selected more frequently). Partition uniformity is evaluated using the θ index extracted from the DMM model. Low values of θ correspond to highly variable partitions. The entire dataset was modeled using the DMM, using the corresponding genera and species from the training dataset in the remaining dataset. Hierarchical clustering based on Jensen-Shannon distance (Ward method) was used to compare the genera and species α weights for each DMM component in the AGP dataset.

[0089] The inventors used the PHATE algorithm (phateR version 1.0.7) to evaluate the potential structure of microbiome branches on the AGP dataset (relative abundance of genus and species), where the γ parameter was set to zero (Moon et al., Nature Biotechnology 37:1482-92, 2019) for visualization purposes and all other parameters were set to default values. Based on the DMM genus and species α weights, the inventors extracted the microbiome branches.

[0090] Longitudinal analysis

[0091] To assess patterns of stability within partitions and switching between partitions over time, the inventors extracted data from the AGP dataset from 745 participants who provided at least two samples, resulting in 2,998 samples. Partition stability was assessed by comparing the proportion of individuals who remained in their originally assigned partition with the proportion of individuals who remained in their randomly assigned partition. Randomization was performed 100 times to calculate confidence intervals for each partition. The 95th percentile of stability achieved for the randomly assigned partitions was retained as the significance threshold.

[0092] Health, demographics, and diet related to subdivisions and branches

[0093] To assess whether the health, demographic, and dietary metadata of participants were associated with gut microbiome partitioning and branching in AGP, we fitted a polynomial log-linear model using the nnet R package (version 7.3-15) with a neural network. Gut microbiome partitioning was used as the response variable, and metadata was used as the predictor variable. The partition with the highest α diversity, as assessed by the Shannon index, was defined as the reference.

[0094] Categorical variables (e.g., region of birth) were one-hot encoded. Dichotomous predictor variables (e.g., self-reported disease status) were coded from 0 to 1 (e.g., “diagnosed by a healthcare professional” was coded as 1, “I do not have this condition” was coded as 0, and other categories were treated as missing values). Continuous predictor variables (e.g., age and BMI) were scored from 0 to 1. Ordinal predictor variables (e.g., food group frequency) were coded from 0 (never) to 1 (daily). Antibiotic history was treated as an ordinal factor ranging from “I did not take antibiotics in the past year” to “during the week” of stool sample collection. Briefly, 100 health, demographic, and dietary predictor variables were used to construct the model. Missing values were replaced by the mean value calculated for each predictor variable.

[0095] Log odds ratios (log odds ratios) of predictor variables and gut microbiome compartments and their respective p-values were extracted from the resulting models using the broom helpers R package (version 1.2.1). False discovery rates were calculated for the predictor variables. All statistical analyses were performed in silico using R software, version 3.6 (Team 2021).

[0096] result

[0097] To identify partitions, or possible ecological states, of the human gut microbiome, a Dirichlet polynomial mixture (DMM) partitioning method was used, as this approach has been widely used in previous gut microbiome studies. The potential energy of heat diffusion within the affinity-based trajectory embedding (PHATE) algorithm was applied to further explore this possibility. PHATE is considered a visualization method for discovering latent structure (e.g., transformations) in high-dimensional data while preserving both global and local structure. This method has previously been applied to gut microbiome data to detect "branching" (Moon et al., 2019). The partitioning and ranking method was applied to the American Gut Project dataset, which consists of approximately 16,000 stool samples based on 16S rRNA gene amplicon sequencing, which were correlated with multiple demographic, lifestyle, health, and dietary variables and included longitudinal sampling from a subset of individuals. Using DMM-based modeling at the genus-species level, the best model fit was obtained for 19 partitions in the American Gut Project database. High homogeneity within partitions (except for partition 19) based on high prediction confidence is shown, as well as consistent Shannon alpha diversity between the training and remaining sets.

[0098] Then, a PHATE plot of the AGP dataset was generated, and three main branches were generated. Comparing the proportions of Bacteroides and Prevotella distinguished the two main branches. The projection of the centroids of the 19 DMM partitions on this PHATE plot resulted in the partitions being arranged along global branches. Notably, a decreasing gradient of alpha diversity was observed from the partitions dominated by Clostridiales to the tips of the Bacteroides or Prevotella branches: the M1 partition was the most diverse, while the least diverse Bacteroides or Prevotella partitions (M11 and M18, respectively) formed the tips of their branches ( Figure 1 ).

[0099] Data from 745 participants from the AGP cohort, from whom at least two samples were collected over time with a mean interval of 12.5 (IQR [1.1-73.0]) days, were then analyzed. Each of the 2998 associated gut microbiome profiles was associated with a partition and branch, and the changes between the two time points were reported. To test whether local partitioning was randomly stratified or ecological state, networks were constructed based on longitudinal data with stability as a function of state, while instability was marked by a high incidence of observed switching ( Figure 2 ). It is important to note that this network is performed in an unsupervised manner (i.e., without integrating Figure 1 ).

[0100] The percentage of individuals remaining in their initial partition (average, 42%) consistently outperformed the frequency calculated using randomly generated events (~10% upper CI 95% limit), suggesting that the partitions may be relatively stable states. Among the 18 partitions, partitions M15 and M14 were the most stable within the Bacteroides and Prevotella clades, respectively. In contrast, some partitions were connected by a higher incidence of network-based switching (e.g., between M3 and M5 and between M2 and M6). These switches occurred within Bacteroides-rich clades, with Faecalibacterium and Bacteroides having the highest weights in these partitions. These partitions were further evaluated to determine whether they were artifacts of overpartitioning. Using alpha diversity and the proportions of the two most abundant genera (Faecalibacterium and Bacteroides) as markers of ecosystem composition within Bacteroides-rich clades, it is reasonable to assume that overpartitioning would occur equally across and within the tested partitions over time. By plotting these parameters over time, it was observed that compositional shifts were limited in the absence of individual switching between partitions. At the same time, where switching between partitions did occur, compositional changes were scattered and overlapping.

[0101] We further investigated the dynamics of the Faecalibacterium:Bacteroides ratio during partition switching between M3 / M5 and M2 / M6, and observed that this ratio varied more when switching between partitions than when fluctuating within partitions. This observation suggests that the difference in Faecalibacterium:Bacteroides ratios was larger between samples from participants who switched between partitions than from samples from participants who remained stable within partitions over time, supporting our hypothesis that partitions may represent local stable states. Notably, in subjects who switched partitions over time, the median Faecalibacterium:Bacteroides ratio was almost identical between the two groups of partitions (M2 / M6 and M3 / M5), at 0.24 and 0.26, respectively, which could suggest that a Faecalibacterium:Bacteroides ratio of 1:4 could be a potential marker of instability in the gut microbiome for the Bacteroides-rich microbiome clade ( Figure 3 Based on this 1:4 ratio, M8, M4, and M2 were classified as the high Faecalibacterium:Bacteroides group. In partition M15, more than 50% of the samples had a Faecalibacterium:Bacteroides ratio below 1:64. Then, partition M15 was classified as the low Faecalibacterium:Bacteroides group. The other partitions (M6, M3, M10, and M11) belonging to the Bacteroides-rich clade were classified as the low Faecalibacterium:Bacteroides group.

[0102] For partitions with a medium or low proportion of Faecalibacterium:Bacteroides, α diversity was significantly correlated with them. Notably, their Shannon index was higher ( Figure 4 ).

[0103] The AGP dataset (approximately 16,000 stool samples) was analyzed to identify associations between branches / partitions and host-related factors, including dietary habits, lifestyle, region of origin, age, BMI, bowel movement frequency, sex, disease, and antibiotic history. Given the high-dimensional nature of the data, a multinomial logistic regression was fitted to 100 predictor variables (i.e., factors) collected via the AGP main questionnaire (FDR < 0.1). To gain insight into factors that could explain the reduced diversity of communities along branches, the most central and diverse partition (M1) was used as a reference in the logistic regression, which consisted mainly of female participants with a high frequency of vegetable consumption (daily) and low exposure to antibiotics ( Figure 5 For each predictor variable, the model returns an odds ratio indicating the strength of association for a given partition compared to a reference partition.

[0104] Individuals with low and medium proportions from the Faecalibacterium:Bacteroides group were mainly associated with lower consumption of vegetables, fruits, and plant diversity in their diet. The low proportion of Faecalibacterium:Bacteroides was mainly associated with higher antibiotic intake and disease ( Figure 6 ).

[0105] Overall, our data suggest that partitions exhibit both common and divergent features with respect to partition-dependent environmental and host factors.

[0106] Ratios less than or equal to 1 / 4 and >1 / 64 may be targets for dietary intervention, whereas <1 / 64 is the target for therapeutic approaches.

Claims

1. A method for determining whether an individual has or is at risk for an altered gut bacterial diversity by determining the relative abundance of Faecalibacterium to Bacteroides in a microbiota sample, and providing dietary advice to the at-risk individual adapted to increase the Faecalibacterium / Bacteroides population ratio.

2. A method for determining whether an individual has or is at risk of having an altered gut bacterial diversity by determining the relative abundance of Faecalibacterium and Bacteroides in a microbiota sample, the method comprising: (i) determining the ratio of Faecalibacterium population to Bacteroides population in a microbiota sample from the individual, (ii) If the ratio of Faecalibacterium population / Bacteroides population is less than or equal to 1 / 4, the individual is determined to have or be at risk of having an alteration in intestinal bacterial diversity.

3. A method for determining whether an individual has or is at risk of developing intestinal dysbiosis by determining the relative abundance of Faecalibacterium and Bacteroides in a microbiota sample, the method comprising: (i) determining the ratio of Faecalibacterium population to Bacteroides population in a microbiota sample from the individual, (ii) If the ratio of Faecalibacterium population / Bacteroides population is less than or equal to 1 / 4, the individual is determined to have intestinal dysbiosis or to be at risk of developing intestinal dysbiosis.

4. The method according to any one of claims 1 to 3, wherein Faecalibacterium and Bacteroides species are the two most abundant species in the intestinal microbiota of the individual.

5. The method according to any one of claims 1 to 4, wherein the microbiota sample is a stool sample.

6. The method according to any one of claims 1 to 5, further comprising: (iii) administering to the individual a diet or treatment to promote restoration of intestinal bacterial diversity.

7. The method according to any one of claims 1 to 6, further comprising: (iii) administering to the individual: - a diet rich in fiber, - a diet containing one or more probiotic species, - food supplements containing bacteria of the genus Faecalibacterium, - a diet rich in Faecalibacterium bacteria, -- A diet rich in riboflavin (vitamin B2), and / or - Treatments to reduce oxidation and / or inflammation.

Images