Scholarly article on topic 'Identification of an Intestinal Microbiota Signature Associated With Severity of Irritable Bowel Syndrome'

Identification of an Intestinal Microbiota Signature Associated With Severity of Irritable Bowel Syndrome Academic research paper on "Clinical medicine"

CC BY-NC-ND
0
0
Share paper
Academic journal
Gastroenterology
OECD Field of science
Keywords
{"Functional Bowel Disorder" / Bacteria / Microbiome}

Abstract of research paper on Clinical medicine, author of scientific article — Julien Tap, Muriel Derrien, Hans Törnblom, Rémi Brazeilles, Stéphanie Cools-Portier, et al.

Background & Aims We have limited knowledge about the association between the composition of the intestinal microbiota and clinical features of irritable bowel syndrome (IBS). We collected information on the fecal and mucosa-associated microbiota of patients with IBS and evaluated whether these were associated with symptoms. Methods We collected fecal and mucosal samples from adult patients who met the Rome III criteria for IBS at a secondary/tertiary care outpatient clinics in Sweden, as well as from healthy subjects. The exploratory set comprised 149 subjects (110 with IBS and 39 healthy subjects); 232 fecal samples and 59 mucosal biopsy samples were collected and analyzed by 16S ribosomal RNA targeted pyrosequencing. The validation set comprised 46 subjects (29 with IBS and 17 healthy subjects); 46 fecal samples, but no mucosal samples, were collected and analyzed. For each subject, we measured exhaled H2 and CH4, oro-anal transit time, and the severity of psychological and gastrointestinal symptoms. Fecal methanogens were measured by quantitative polymerase chain reaction. Numerical ecology analyses and a machine learning procedure were used to analyze the data. Results Fecal microbiota showed covariation with mucosal adherent microbiota. By using classic approaches, we found no differences in fecal microbiota abundance or composition between patients with IBS vs healthy patients. A machine learning procedure, a computational statistical technique, allowed us to reduce the 16S ribosomal RNA data complexity into a microbial signature for severe IBS, consisting of 90 bacterial operational taxonomic units. We confirmed the robustness of the intestinal microbial signature for severe IBS in the validation set. The signature was able to discriminate between patients with severe symptoms, patients with mild/moderate symptoms, and healthy subjects. By using this intestinal microbiota signature, we found IBS symptom severity to be associated negatively with microbial richness, exhaled CH4, presence of methanogens, and enterotypes enriched with Clostridiales or Prevotella species. This microbiota signature could not be explained by differences in diet or use of medications. Conclusions In analyzing fecal and mucosal microbiota from patients with IBS and healthy individuals, we identified an intestinal microbiota profile that is associated with the severity of IBS symptoms. Trial registration number: NCT01252550.

Academic research paper on topic "Identification of an Intestinal Microbiota Signature Associated With Severity of Irritable Bowel Syndrome"

Accepted Manuscript

Identification of an Intestinal Microbiota Signature Associated With Severity of Irritable Bowel Syndrome

Julien Tap, Muriel Derrien, Hans Törnblom, Rémi Brazeilles, Stéphanie CoolsPortier, Joël Doré, Stine Störsrud, Boris Le Nevé, Lena Öhman, Magnus Simrén

PII: S0016-5085(16)35174-5

DOI: 10.1053/j.gastro.2016.09.049

Reference: YGAST 60743

To appear in: Gastroenterology Accepted Date: 29 September 2016

Please cite this article as: Tap J, Derrien M, Tornblom H, Brazeilles R, Cools-Portier S, Doré J, Storsrud S, Le Nevé B, Ohman L, Simrén M, Identification of an Intestinal Microbiota Signature Associated With Severity of Irritable Bowel Syndrome, Gastroenterology (2016), doi: 10.1053/j.gastro.2016.09.049.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Identification of an Intestinal Microbiota Signature Associated With Severity of Irritable Bowel Syndrome

Short title

Intestinal microbiota and irritable bowel syndrome severity Authors

1 2* 1* 3 4 1

Julien Tap12, Muriel Derrien1 , Hans Tómblom3,4\ Rémi Brazeilles1, Stéphanie Cools-Portier1, Joël Doré2, Stine Stórsrud3, Boris Le Nevé1, Lena Ohman3,5,6#, Magnus Simrén3,4,7#

Affiliations

1 Danone Nutricia Research, Avenue de la Vauve, Palaiseau, France

2 INRA MetaGenoPolis, Jouy en Josas, France.

3Department of Internal Medicine & Clinical Nutrition, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden.

4 University of Gothenburg Centre for Person-Centered Care (GPCC), Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

5 Department of Microbiology and Immunology, Sahlgrenska Academy at University of Gothenburg, Sweden

6 School of Health and Education, University of Skóvde, Skóvde, Sweden

7 Center for Functional GI and Motility Disorders, University of North Carolina, Chapel

Hill, NC, United States

Both authors contributed equally to this work, # shared senior authorship

Grant Support

This research was supported by the Swedish Medical Research Council (grants 13409, 21691 and 21692), AFA Insurance, The Marianne and Marcus Wallenberg Foundation, University of Gothenburg, Centre for Person-Centered Care (GPCC), Sahlgrenska Academy, and University of Gothenburg and by the Faculty of Medicine, University of Gothenburg, VINNOVA as well as by Danone Nutricia Research.

Abbreviations

IBS: Irritable Bowel Syndrome, OTU: Operational Taxonomic Unit, IBS-SSS: IBS Symptom Severity Score, BSF: Bristol Stool Form, OATT: Oro-Anal Transit Time, JSD: Jensen Shannon Divergence.

Correspondence

Pr Magnus Simren: magnus.simren@medicine.gu.se Dr Muriel Derrien : muriel.derrien@danone.com Conflict of Interest

B. Le Nevé, M. Derrien, R. Brazeilles, S. Cools-Portier and J. Tap are employees of Danone Research. M Simrén has received unrestricted research grants from Danone and AstraZeneca, and served as a Consultant/Advisory Board member for AstraZeneca, Danone, Novartis, Almirall, Albiroe, Shire, Nestlé, Glycom and Chr Hansen, and been on the speaker's bureau for Takeda, Tillotts, Shire, Almirall, Menarini and Danone. L. Ohman has received a financial support for research by Danone Research and served as a consultant for Genetic Analyses (Lecture fee(s): Abbvie, Takeda). J. Doré has received a financial support for research by Danone Research, Pfizer and PiLeJe, and served as a consultant/Advisory Board member for Danone Research and AlphaWasserman, Enterome Bioscience and MaaT Pharma.

H. Tornblom served as a consultant/Advisory Board member for Almirall, Allergan, Danone and Shire, and has been on the speakers' bureau for Tillotts, Takeda, Shire and Almirall. S. Storsrud has nothing to declare.

Author Contributions

LO, HT, BLN, MS contributed to the study concept and design. HT collected the clinical samples. JT, MD, LO, SCP, BLN, SS, HT contributed to the acquisition of data. JT, MD, LO, RB, BLN, SS, MS analyzed and interpreted the data. JT and RB made statistical analyses. JT, MD, BLN, LO, HT and MS drafted the manuscript. MS, BLN, JD provided administrative, technical or material support. All authors contributed to the critical revision of the manuscript and gave approval of the final version of the manuscript.

Acknowledgements.

We acknowledge Anne Druesne and Chenhong Zhang for technical assistance with DNA extraction and Johan van Hylckama Vlieg for valuable discussions. Martin Balvers and Jolanda Lamberts are deeply acknowledged for their bioinformatics support in microbiota analysis. The quantitative PCR used was operated under Yakult License (YIFSCAN technology). Graphical summary (Figure 6) use icons made by Crea Nostra (www.creanostra.fr).

Patient consent Obtained

Ethics approval: this study was approved by the Regional Ethical Review Board at the University of Gothenburg.

Provenance and peer review

Not commissioned; externally peer reviewed.

Data sharing statement

Data from this study are available upon request.

ABSTRACT

Background and Aims: We have limited knowledge about the association between the composition of the intestinal microbiota and clinical features of irritable bowel syndrome (IBS). We collected information on the fecal and mucosa-associated microbiota of patients with IBS and evaluated whether these were associated with symptoms.

Methods: We collected fecal and mucosal samples from adult patients who met the Rome III criteria for IBS at secondary or tertiary care outpatient clinics in Sweden, as well as from healthy subjects. The exploratory set comprised 149 subjects (110 with IBS and 39 healthy subjects); 232 fecal samples and 59 mucosal biopsy samples were collected and analyzed by 16S rRNA targeted pyrosequencing. The validation set comprised 46 subjects (29 with IBS and 17 healthy subjects); 46 fecal samples, but no mucosal samples, were collected and analyzed. For each subject, we measured exhaled H2 and CH4, oro-anal transit time, and severity of psychological and gastrointestinal symptoms. Fecal methanogens were measured by quantitative PCR. Numerical ecology analyses and a machine learning procedure were used to analyze the data.

Results: Fecal microbiota demonstrated covariation with mucosal adherent microbiota. Using classical approaches, we found no differences in fecal microbiota abundance or composition between patients with vs without IBS. A computational statistical technique-like machine learning procedure allowed us to reduce the 16S rRNA data complexity into a microbial signature for severe IBS, consisting of 90 bacterial operational taxonomic units. We confirmed the robustness of the intestinal microbial signature for severe IBS in the validation set. The signature was able to discriminate between patients with severe symptoms, patients with mild/moderate

symptoms, and healthy subjects. Using this intestinal microbiota signature, we found IBS symptom severity to be negatively associated with microbial richness, exhaled CH4, presence of methanogens, and enterotypes enriched with Clostridiales or Prevotella species. This microbiota signature could not be explained by differences in diet or use of medications.

Conclusions: In analyzing fecal and mucosal microbiota from patients with IBS and healthy individuals, we identified an intestinal microbiota profile that associates with the severity of IBS symptoms. Trial registration number: NCT01252550

Keywords: functional bowel disorder, bacteria, microbiome

Trial registration number: NCT01252550

INTRODUCTION

Irritable bowel syndrome (IBS) is the most prevalent functional gastrointestinal disorder in western societies. It affects about 11% of the adult population and strongly impairs quality of life, social function, work productivity and brings substantial

costs to health-care services'. The etiology of IBS remains poorly understood and the

search for biomarkers is ongoing2.

It is now well accepted that IBS is a disorder involving multiple pathophysiological mechanisms where composition of gut microbiota has been proposed as one of the potentially important factors3,4. Since the first study that investigated the fecal microbiota composition of IBS patients and healthy subjects using a molecular based approach5, many studies have followed using targeted approaches6, specifically quantitative PCR 6,7 More recently, the use of advanced tools has allowed a better

689 10 1112

overview of gut microbiota composition689, function10, and metabolites production1112 in IBS. Even though gut microbiota alterations seem to exist in IBS, no uniform gut microbiota pattern in IBS has been demonstrated. The existing inconsistencies among currently available data in IBS may be attributed to several factors including heterogeneity of gut microbiota profiling methods, inherent individual microbiota variability and differences in inclusion criteria, as well as sample size. This highlights the difficulty to find robust microbiota markers associated with IBS clinical parameters3, and shows the need for large and well characterized cohorts to obtain valid and reliable analyses of the association between clinical symptoms and microbiota composition and function in IBS.

Fecal microbiota has been the target of most studies due to its convenient accessibility. However, mucosal microbiota is of great interest given its proximity to host cells as host-microbe interactions have been proposed to be of relevance for

symptom generation in IBS4. Although the analysis of mucosal microbiota involves more invasive sampling methods, several studies have investigated colonic mucosal

samples obtained by sigmoidoscopy, either following bowel cleansing13 or using unprepared biopsies14-16. The use of unprepared biopsy samples offers the

advantage to avoid perturbation of the microbiota composition1'. Large cohort studies in IBS patients combining paired microbiota samples originating from fecal and mucosa samples are rare, but a few smaller studies using next generation sequencing exist13,16.

Thus, there is currently a need for an improved understanding of gut microbiota composition in IBS patients and for the potential role played by the gut microbiota in the generation of IBS symptoms. In this study, we therefore aimed to determine the fecal and mucosa-associated microbiota, and the link to clinical symptoms in a large and well-characterized cohort of IBS patients.

MATERIAL AND METHODS Subject recruitment and study design

Adult patients, aged 18-65 years and fulfilling the Rome III criteria18 for IBS were prospectively included at a secondary/tertiary care outpatient clinic (Sahlgrenska University Hospital, Sweden). The diagnosis was based on a typical clinical presentation and additional investigations if considered necessary by the gastroenterologist (HT or MS). Classification into IBS subtypes according to the Rome III criteria was done based on Bristol Stool Form (BSF) scale characteristics: IBS with constipation (IBS-C), IBS with diarrhea (IBS-D), mixed IBS (IBS-M) or unsubtyped IBS (IBS-U)18. Exclusion criteria included the use of probiotics or antibiotics during the study period or within one month before the inclusion, another diagnosis that could explain the GI symptoms, severe psychiatric disease as the dominant clinical problem, other severe diseases, and a history of drug or alcohol abuse. By use of advertisement, a healthy control group was recruited and checked by interview and a questionnaire to exclude chronic diseases and any current GI symptoms.

All participants gave their written informed consent to participate after verbal and written information about the study. The Regional Ethical Review Board at the University of Gothenburg approved the study prior to the start of subject inclusion.

Subject characterization

Demographic information and body mass index were obtained in all subjects. IBS patients reported their current use of medications and completed questionnaires in order to characterize their symptom severity and bowel habits: The IBS Severity Scoring System (IBS-SSS)19, the Hospital Anxiety and Depression (HAD) scale20, a

4-day food diary21 and a two-week stool diary based on the Bristol stool form scale22. IBS severity was based on validated cut-off scores on IBS-SSS (mild IBS: IBS-SSS <175, moderate IBS: IBS-SSS=175-300, severe IBS: IBS-SSS>300)19.

The oro-anal transit time (OATT) (radiopaque marker study)23 and the amount of exhaled H2 and CH4 after an overnight fast, (i.e. not after intake of any substrate), was also determined in IBS patients (see supplementary online-only material for more details).

Fecal and mucosal samples collections and DNA extraction

Fecal samples were collected from 195 subjects in RNA later solution (Ambion, Courtaboeuf, France). For most of the IBS subjects, two fecal samples were collected (average 26 ± 16 days between the two samples). A first set, referred to as the exploratory set, was composed of 149 subjects (110 IBS and 39 healthy subjects) from which 232 fecal samples and 59 biopsy samples were collected and analyzed. A second set, referred to as the validation set, was composed of 46 subjects (29 IBS and 17 healthy subjects) from which 46 fecal samples, but no biopsy samples were collected and analyzed (Table S1). The division into two study sets was based on sampling date.

Fecal DNA was extracted using mechanical lysis (Fastprep® FP120 (ThermoSavant) followed by phenol/chloroform-based extraction as previously described24. Biopsies from the sigmoid colon were obtained from 59 subjects (39 IBS and 20 healthy subjects). The biopsies were taken 25-35 cm proximal of the anus during an unprepared sigmoidoscopy. Once collected, biopsies were immediately placed in liquid nitrogen and stored at -80°C until further a nalysis. Mucosal adherent microbiota DNA was isolated using a adapted protocol based on Godon and colleagues (for

more detail, see supplementary online-only material)25. Routine histopathology of biopsies confirmed the absence of active inflammation.

Microbial composition assessment

Hypervariable 16S rRNA regions (V5-V6) were amplified using primers 5'-AGGATTAGATACCCTGGTA-3' and 5'-CRRCACGAGCTGACGAC-3'. Sequencing was performed by DNA Vision SA (Charleroi, Belgium) on a 454 Life Sciences Genome Sequencer FLX instrument (Roche) using titanium chemistry. Raw reads quality filtering and trimming, OTU (operational taxonomic units) clustering and taxonomic assignment were all performed using the LotuS v1.32 pipeline26 (for more details, see supplementary online-only material).

Detection of Methanobacteriales by quantitative PCR

Fecal extracted DNA were subjected to quantitative PCR using primers targeting the Methanobacteriales order Mtb857F (5'-CGWAGGGAAGCTGTTAAGT-3') and Mtb1196R (5'-TACCGTCGTCCACTCCTT-3') as has been described previously27. PCR results were then translated as presence or absence of dominant Methanobacteriales in fecal sample using 106 rRNA genes copies per gram of wet fecal content as threshold as defined by the lowest detected value in the standard curve.

Numerical ecology and statistical analysis

Alpha diversity (diversity within samples) was assessed using numbers of observed OTUs rarefied at the same sequencing depth (3,500 sequences per sample in this study) using the vegan R package28. Beta diversity (diversity between samples) was assessed by square root Jensen-Shannon divergence metrics29, referred to as Jensen-Shannon distance (JSD) in the manuscript, as well as Bray-Curtis distance. Enterotype stratification was identified in fecal samples (one sample per individual) using previously described methods with the Dirichlet multinomial mixture models (DMM) 30

A machine learning procedure to identify a microbial signature for IBS severity was implemented using L1 regularized logistic regression31 (Least Absolute Shrinkage and Selection Operator or LASSO) using the LIBLINEAR library32 validated through a ten-fold independent cross-validation. Features selection and data transformation was processed as previously described33. The performance of prediction models was assessed for its discriminative ability using area under the receiver operating characteristic curve (AU-ROC) on exploratory and validation datasets. OTUs selected by machine learning were further characterized by their prevalence in healthy subjects and severe IBS patients (IBS-SSS>300), and by their phylogenetic specificity (for more detail, see supplementary online-only material).

Co-inertia analysis was used to identify the relationship between the fecal and biopsy microbiota datasets, as well as the relationship between a microbial signature for IBS severity and the clinical data, using the ade4 R package34. The overall similarity between datasets was then measured by the RV coefficient34. Relation between variables was assessed using non-parametric tests (Wilcoxon test and Spearman correlation test) for continuous variables (e.g. bacterial relative abundance, microbiota alpha and beta-diversity, IBS-SSS, HAD, age, BMI, exhaled gas, stool

frequency and consistency) and chi squared test for categorical variables (e.g. enterotypes, IBS subtypes, presence or absence of methanogens). A Monte Carlo permutation test (99 permutations) was used to assess the robustness of the RV-coefficient. The statistical approaches (e.g. Co-inertia and Monte-Carlo) applied to microbiota analysis have previously been thoroughly described35. In case of multiple testing, all p-values were adjusted by Benjamini-Hochberg false discovery rate (FDR) correction36. All statistical analyses were carried out with R software and described in supplementary online-only material)

Access to study data

All authors had access to the study data and reviewed and approved the final manuscript. Source codes used in this study are available from GitHub (http://github.com/tapj/IBSMicrobiota).

RESULTS

Clinical characteristics of IBS patients and healthy subjects

The study cohort was divided into two study sets based on sampling date. The first set, referred to as the exploratory set (149 subjects: 110 IBS and 39 healthy), were recruited between April 2010 and May 2012. The second study set, referred to as the validation set (46 subjects: 29 IBS and 17 healthy), were recruited between June 2012 and November 2013. Detailed clinical and demographic characteristics for the exploratory and validation sets are summarized in Table 1 and Figure S1. There were no differences between IBS patients or healthy subjects from the two study sets regarding gender, age, BMI or clinical parameters. Apart from expected differences between IBS subtypes in stool characteristics and OATT, also gender differences were seen (more females in IBS-C, Pearson's Chi squared test, p<0.05), but otherwise other clinical and demographic were similar in IBS subtypes. Regarding IBS subtypes distribution, no differences were observed, neither between severe and other IBS, nor between the exploratory and validation sets (Table S2).

Fecal and mucosal microbiota are structurally distinct but highly correlated

Paired mucosal and fecal microbiota samples were analyzed from 59 study subjects in the exploratory set (Table S1). Paired comparisons of 16S rRNA gene sequencing analysis revealed that fecal microbiota harbored more Firmicutes and Actinobacteria, while mucosal microbiota was enriched in Bacteroidetes and Proteobacteria (p<0.05) (Figure 1A). Alpha-diversity of mucosal adherent microbiota, as measured by the number of OTUs, was significantly lower than that of fecal microbiota (p<0.05) (Figure 1B). Using a co-inertia approach based on JSD metrics, mucosal microbiota

was significantly associated with the fecal microbiota (RV=0.71, p<0.001), as depicted in Figure 1C-D.

Microbiota diversity in fecal and biopsy samples

Microbiota diversity in both fecal and biopsy samples was analyzed in the exploratory set using classical ecological descriptive approaches including alpha-diversity (richness as measured by number of OTUs, 97% identity) and beta-diversity metrics (JSD and Bray-Curtis distance using bacterial genus relative abundance). Neither microbiota richness, nor microbiota variability, differed between groups (healthy subjects, IBS patients, IBS Rome III subtypes, IBS severity) in the exploratory set. This was true for both fecal and biopsy samples (Figure S2).

Enterotype stratification in healthy subjects and IBS patients

Microbiota clustering was performed on 232 fecal samples from IBS patients (including two samples for most patients) and healthy subjects in the exploratory set (Table S1) using Dirichlet multinomial mixture model (DMM).

Microbiota separated optimally into three distinct microbiota communities as assessed by Bayesian information criterion (BIC) and Laplace parameter (Figure S3 A, B). The three identified microbiota communities were similar to the previously described gut microbial enterotypes29. One enterotype was enriched in Bacteroides (16% of samples), one enriched in Prevotella (14% of samples) and one enriched in Clostridiales (70% of samples) (Figure 2A and S3C). Notably, distribution of enterotypes correlated with microbial richness, with Bacteroides-enterotyped subjects harboring the lowest richness, as compared to Prevotella and Clostridiales-enterotyped subjects (p<0.05) (Figure 2B).

Next, we analyzed enterotype distribution and association with clinical parameters. After correcting for multiple comparisons, no significant association was observed between enterotypes and clinical parameters including age, BMI, HAD anxiety, HAD depression, exhaled H2 and CH4 or bowel habits (stool consistency and frequency) (Figure S4A). However, enterotype distribution was associated with OATT in both healthy subjects and IBS patients (Figure S4A). Subjects with the Clostridiales enterotype exhibited longer transit time than subjects with Prevotella and Bacteroides enterotypes (p<0.05). Further, the Prevotella enterotype was significantly more prevalent in men (p<0.05) (Figure S4B). Enterotype distribution differed between healthy subjects and IBS patients, with the enterotype Bacteroides being more frequent in IBS subjects whereas the Prevotella enterotype was more common in healthy subjects (Figure S4C) (p<0.05). Enterotype distribution was also associated with bowel habits (Pearson's Chi-squared test, p<0.05). IBS-D and IBS-M patients had a higher prevalence of Bacteroides enterotype compared to IBS-C patients and healthy subjects (Fig. 2C). Regarding IBS symptom severity (IBS-SSS), the prevalence of the Prevotella enterotype gradually decreased as symptom severity increased (p<0.05) (Figure 2D).

Prevalence of methanogens in healthy subjects and IBS patients

In the exploratory set, microbiota enterotyping was complemented by detection of fecal methanogens by quantitative PCR, with specific emphasis on Methanobacteriales, which was detected in 33% of fecal samples with a similar prevalence in healthy subjects and IBS patients (Figure S5). A significant association was observed between exhaled CH4 and presence of fecal methanogens (p<0.05) (Figure S5A). Detection of methanogens was dependent on enterotypes as 90% of

individuals who harbored Methanobacteriales belonged to the Clostridiales enriched enterotype and less than 5% of Bacteroides enriched enterotypes had detectable level of Methanobacteriales (Figure S5B).

The presence of methanogens was associated with IBS subtype distribution (p<0.05). The proportion of IBS-D patients with undetectable Methanobacteriales was higher (~40%) than in IBS-C patients (~10%) (Figure S5C). IBS symptom severity (IBS-SSS) was not associated with the presence of Methanobacteriales (p > 0.05, Figure S5D). IBS-D patients exhaled less CH4 than IBS-M patients (Figure S5E), and heathy subjects exhaled less CH4 than IBS patients with mild symptoms (Figure S5F).

Identification of a microbial signature for IBS symptom severity

We further explored the association between IBS symptom severity and fecal microbiota composition. Compared to mild and moderate IBS subjects, a significantly higher number of OTUs (n=100) could discriminate IBS subjects with severe symptoms from healthy subjects (Figure 3A). Using bootstrapping method, we observed that the analysis was sensitive to randomness. When comparing two batches of 30 individuals randomly selected from the exploratory set, up to 50 bacterial OTUs could differ between the two batches driven by chance. In other words, compared to the total number of differential OTUs, half of the significant observations between tested groups could be detected by chance (Figure 3A). To overcome the issues related to randomness, we explored fecal microbiota in association with IBS symptom severity using an additional and more robust statistical approach, based on machine learning (LASSO). This allowed us to decrease the OTUs complexity by combining them into a consensus microbial signature from an ensemble of classifiers which discriminated patients with severe IBS from patients

with moderate or mild IBS and healthy subjects. Based on IBS symptom severity, 90 OTUs out of 2,911 total OTUs were selected by the machine learning procedure. The predictive power of this signature was quantified by AU-ROC analyses. Cross-validation of the microbial signature for IBS severity obtained with the exploratory set was performed against the fecal samples validation set (AUC of 0.74). The signature based on fecal samples in the exploratory set was also efficient to classify mucosal samples according to IBS severity (AUC of 0.82) (Figure 3B). Next, we assessed the OTUs microbial signature for IBS severity obtained from the exploratory set in the validation set (n=46 individuals) which included 13 severe IBS patients (Figure 3B). In the validation set, an AUC of 0.64 was obtained, suggesting that the OTUs identified as a microbial signature for IBS severity were robust.

Taxonomical characterization of the gut microbial signature for IBS severity

In order to further characterize the gut microbial signature for IBS severity, we analyzed the taxonomy of the 90 OTUs selected by the machine learning procedure (Figure 4). To assess the phylogenetic distribution of the microbial signature for IBS severity, we performed a principal coordinate analysis of the OTUs originating from the whole fecal microbiota dataset (n=2,911 OTUs), using nucleotide identity between their respective representative sequences (Figure 4A). Notably, there were no phylogenetic lineage specific of the microbial signature for IBS severity, but instead an overlapping taxonomy between the microbial signature for IBS severity and the whole microbiota dataset, suggesting that those 90 OTUs associated with IBS severity are taxonomically as diverse as the OTUs from the whole microbiota dataset. When examining only the OTUs extracted from the microbial signature for IBS severity, the dominant families of the gut microbiota were represented (ie.

Lachnospiraceae, Ruminococcaceae, Bacteroidaceae) but a large proportion (>25%) was not assigned at the family level (Figure 4B). Then, the 90 IBS severity discriminating OTUs were ranked according to their average weight in the model (Table S3). In this model, OTUs with negative weight had a positive association with IBS severity. The proportion of OTUs that displayed positive or negative weight in the model was similar within Firmicutes and Bacteroidetes phyla (Figure 4C). The number of genus-unassigned OTUs from the Firmicutes (notably in the Ruminococcaceae family) increased when they were associated with IBS severity (Pearson's Chi-squared test, p<0.05). In the Firmicutes phylum, including known and dominant genera such as Faecalibacterium, Oscillibacter, Blautia and Coprococcus OTUs were positively associated with IBS of moderate severity or healthy status. Finally, we compared the prevalence of the 90 OTUs from the signature for IBS severity between healthy subjects and patients with severe IBS. The prevalence of OTUs that were positively associated to healthy status (i.e. positive weight in the model signature) was significantly higher in the microbiota of healthy subjects as compared to severe IBS patients (Wilcoxon test, p<0.05, Figure 4D).

Gut microbial signature for IBS severity and association with clinical and microbial parameters

To evaluate the robustness of the microbial signature for IBS severity in relationship to other clinical parameters, we investigated the relative abundance of OTUs along with clinical data in a co-inertia analysis as depicted in Figure 5. Clinical parameters included age, BMI, HAD anxiety, HAD depression, exhaled CH4 and H2, IBS subtypes, stool consistency (BSF) and frequency, and OATT. The first two co-inertia components explained more than 50% of co-variation between the IBS severity signature microbial OTUs and clinical parameters.

IBS severity (IBS-SSS) was confirmed to be the most important factor contributing to variation in the full dataset along the first co-inertia component PC1 (Figure 5A). In addition, OTUs that had a positive weight in the microbial signature were more prevalent in microbiota of healthy subjects and were positively associated with PC1 (Figure 5B). This suggests that OTUs selected by the machine learning procedure allowed ranking patients along the IBS severity scale (Figure 5C).

As expected, anxiety and depression were positively associated with IBS severity and this association was reflected on the gut microbial signature (Spearman rho correlation with PC1 of -0.4 and - 0.32 respectively, p<0.05, Table S4). Also, exhaled CH4 concentration (Figure 5A and D) was associated with both components PC1 (rho=0.36, p <0.05) and PC2 (rho=0.44, p<0.05). IBS-C and exhaled CH4 concentrations were the most important factors explaining the variation along the second co-inertia component PC2 (Figure 5A and E). This suggests that exhaled CH4 was primarily associated with slower transit and secondly with less severe symptoms. PC2 was to a lesser extent associated with OATT (rho=0.54, p<0.05) and stool consistency (rho=-0.36, p<0.05). This suggests that the machine learning procedure, which was originally set up to discriminate severe IBS from mild, moderate IBS and healthy subjects, selected additional OTUs to identify IBS-C patients with high concentrations of exhaled CH4. However, the microbial signature for IBS severity was poorly explained by age, BMI and H2 concentrations as depicted in Figure 5A.

We then tested a posteriori these two first co-inertia components against microbiota parameters. These included microbial richness, absence or presence of Methanobacteriales, and enterotype stratification (Figure 5C, D, F respectively), which were altogether significantly associated with the first component PC1

(Wilcoxon test, p< 0.05). When plotting IBS-SSS and microbial richness, the gut microbial signature for IBS severity was linked with lower microbial richness (Figure 5C), lower levels of exhaled CH4 (Figure 5D), and Bacteroides-enriched enterotype (Figure 5F). This suggests that selected OTUs from the machine learning procedure discriminated patients along a symptom severity gradient together with enterotype stratification.

Gut microbial signature for IBS severity and association with diet and use of medications

111 individuals (89 IBS and 22 healthy) followed a 4-day food diary to assess nutrient intake. Average daily intakes were calculated for energy, proportion of fat, carbohydrates, fiber and protein, and total intake of FODMAPs (Table S5). By using a co-inertia analysis followed by a Monte Carlo test (see supplementary online-only material), the overall association between dietary data and the gut microbial signature for IBS severity was tested, and no significant association could be detected (RV=0.10, p>0.05), indicating that the microbial signature for IBS severity is independent of overall nutrient intake, as well as intake of FODMAPs. Similar analyses regarding the influence of medications (Table S6), including laxatives or bulking agents, acid suppressants (mainly proton pump inhibitors), antidiarrheals and antidepressants drugs did not reveal any significant associations with the microbial signature for IBS severity.

DISCUSSION

In this study we have characterized both fecal and mucosal microbiota in the so far largest cohort of IBS patients and healthy subjects. Using a machine learning approach, we demonstrate that IBS symptom severity is associated with a distinct

fecal microbiota signature that is also detected in the intestinal mucosa. This signature is also associated with microbial richness, exhaled CH4, presence of Methanobacteriales and enterotype stratification, assessed by using Dirichlet multinomial mixture model, as well as stool consistency and transit time. Lower microbial richness and exhaled CH4, as well as reduced prevalence of Methanobacteriales and Prevotella enterotype were observed in severe IBS subjects. Interestingly, the prevalence of Prevotella enterotype decreased as the severity of symptoms increased, in parallel to increased prevalence of Bacteroides enterotype. A graphical summary of the main findings is demonstrated in Figure 6.

We explored the data using a combination of approaches that are well described for microbiota analysis, i.e. univariate and multivariate analyzes. However, when using these classical ecological approaches, no clear differences were observed between healthy subjects and IBS patients, or between IBS subtypes as defined by the Rome III criteria. No differences were detected in fecal microbiota between IBS and healthy subjects regarding alpha-diversity (microbial richness), or beta-diversity (pairwise JSD distance comparisons) at any taxonomy level (from phyla to species). Similar findings were observed for mucosal microbiota. A trend towards a reduction in richness of fecal microbiota was observed in IBS subjects. Rajilic-Stojanovic and coworkers did not report change in alpha-diversity but an almost two-fold increase of the ratio of major bacterial phyla Firmicutes:Bacteroidetes in 62 IBS patients (Rome II criteria) compared to 46 healthy subjects8. In a previous study from our group, although no overall differences in microbiota composition between 37 IBS (Rome II criteria) patients and 20 healthy subjects were observed, nevertheless two specific IBS sub-clusters with altered fecal microbiota composition were identified. Interestingly, the two IBS sub-clusters accounted for 60% of IBS patients in the study,

and harbored an increased Firmicutes:Bacteroidetes ratio compared to that of healthy subjects and other IBS patients9. Other studies, as us, however, did not report significant differences between IBS and healthy subjects neither in fecal nor in

1 5 37

small intestinal microbiota or reported conflicting results153'. Recently , a larger cohort from Pozuelo et al. similar to our study, reported in contrary a trend towards a higher abundance of Bacteroidetes in 113 IBS (Rome III criteria) compared to 66 healthy subjects38, as well as a lower richness in IBS subjects, probably driven by the high proportion of IBS-D subjects in that cohort. These discrepancies could be due to several factors including heterogeneity of IBS cohorts (i.e proportion of IBS subtypes, with a majority of one subtype, and differences in symptom severity), as well as the use of different methods (machine learning versus descriptive statistics) and 16S rRNA gene variable regions used to analyze microbiota (Table S7 and references within), and also absence of validation cohorts. However, differences between IBS patients and healthy subjects were observed in our study based on enterotype distribution, with healthy subjects being more likely to belong to the Prevotella enterotype than IBS patients, with concomitant increase in prevalence of Bacteroides enterotype in IBS.

We also explored paired mucosal and fecal microbiota data. The distinction between fecal and mucosal colonic microbiota in terms of composition, was already detectable at the phylum level, with increased proportions of Bacteroidetes and Proteobacteria in biopsies, while higher abundance of Firmicutes and Actinobacteria were observed in fecal samples. Our data are also in part supported by Rangel et al., who observed different microbiota composition between mucosal and fecal samples in both healthy subjects and IBS patients, although the microbial diversity in biopsies and fecal samples from their study was lower among IBS patients16. In our study, the difference

in microbiota composition between fecal samples and biopsies was accompanied by lower richness in biopsies as compared with fecal samples in both healthy subjects and IBS patients, which is consistent with previous reports39,40. Although microbiota from paired mucosal and fecal samples were structurally distinct and with different alpha-diversity, their respective microbial beta-diversity co-varied, which is consistent with the data obtained from the rhesus macaque microbiota41. A larger number of biopsy samples would be needed to decipher whether mucosal microbiota is more strongly associated with IBS severity than fecal microbiota.

In this study, we complemented the microbiota composition analysis by quantification of microbial groups able to produce CH442, so called methanogens, which prevent H2 accumulation in the gut. Production of CH4 using H2 is restricted to Archaea, with the order Methanobacteriales containing Methanobrevibacter smithii and Methanosphaera stadtmanae as the dominant methanogens in human42 43. Evidence has emerged suggesting that CH4 is linked to constipation44, and it has been reported that Methanobrevibacter smithii is more commonly found in patients with IBS-C45. In our study, 33% of study subjects harbored detectable levels of Methanobacteriales, with no difference between healthy subjects and IBS patients. Interestingly, there was a significant correlation between exhaled CH4 and detection of Methanobacteriales by quantitative PCR, which is in line with previous findings45. In our study, the presence of Methanobacteriales differed according to IBS subtypes. IBS-D patients had more undetectable Methanobacteriales compared to the other IBS subtypes. Moreover, Methanobacteriales detection was positively associated with microbial richness in the enterotype Clostridiales in our study, which is in accordance with the study by Vandeputte et al.,46. This enterotype was also associated with longer transit time, which is consistent with another recent study47.

We explored the data using a robust statistical analysis based on a machine learning algorithm, as large microbial datasets generated from sequencing technologies might generate over-fitting and over-estimation48. Recently, there has been a growing interest in the use of machine learning techniques to detect microbiota signature in health and diseases33,49. In our study, the LASSO procedure identified 90 bacterial OTUs that could be used as a composite gut microbial signature for IBS severity. The signature was robust as it still remained valid after cross-validation and test in the validation cohort. The microbial signature was enriched in taxonomically diverse phylotypes. At the family level, similar proportion of OTUs within Lachnospiraceae and Ruminococcaceae were associated with health or IBS severity. OTUs known to be associated with health were more prevalent in healthy subjects than in severe IBS patients. These OTUs include Faecalibacterium, Oscillibacter, Blautia and Coprococcus species, which were previously reported to belong to the healthy microbiota phylogenetic core50,51. The phylogenetic core may represent co-evolved species within the gut microbiome that support essential gut microbial functions51. Finally, the microbial signature for IBS severity was tested against clinical and microbial parameters. IBS severity was thus confirmed to be the strongest factor associated with the microbial signature along with the presence of methanogens, exhaled CH4, enterotype and microbial richness. Overall, clinical parameters other than IBS severity were not strongly associated with the microbial signature. As recent studies reported transit time as a strong confounding factor for microbiota composition46,47, it seems important to stress that OATT and IBS subtypes were not confounding factors for the microbial signature for IBS severity.

A limitation of the present study that could prevent extrapolation of the results to the general IBS population is that all patients were included at a secondary/tertiary

referral center. It is well known that IBS patients seen in referral centers have more severe GI and psychological symptoms as well as reduced quality of life, and therefore do not reflect the general IBS population. As it is well recognized that diet and intake of medications are two factors that shape gut microbiota52,47, we further examined whether the microbial signature for IBS severity was influenced by these two factors. Our analysis on global intake of nutrients (protein, carbohydrate, fat and calorie intake), and intake of FODMAPs did not support a relationship between these food categories and our gut microbiota signature for IBS severity. Regarding intake of medications, although patients with severe IBS as expected were more likely to be treated with antidepressants than patients with milder forms of IBS, the gut microbiota signature for IBS severity was neither significantly associated with intake of antidepressants, nor with intake of other groups of medications. To conclude, by using a large cohort and classical ecological approaches, we could not observe differences between healthy subjects and IBS patients. However, machine learning approach allowed identifying a gut microbial signature for IBS severity, which could also be reproduced in a validation cohort. Importantly, due to its relatively low sensitivity, this microbiota signature cannot be used as a clinical predictor of the IBS severity, but as a way to explore relevant features (i.e. OTUs), which deserve to be explored in future IBS microbiota studies. Our study highlights the heterogeneity of IBS patients, and the difficulty to stratify patients based on microbiota profile when using only classical ecological approaches. The use of machine learning has allowed us to circumvent the issues related to large microbial dataset and to better explore the microbial data. We were able to identify several interesting links between gut microbiota and the clinical profile.

REFERENCES

1. Almquist E, Tornblom H, Simrén M. Practical management of irritable bowel syndrome: a clinical review. Minerva Gastroenterol. Dietol. 2015.

2. Sood R, Law GR, Ford AC. Diagnosis of IBS: symptoms, symptom-based criteria, biomarkers or "psychomarkers"? Nat. Rev. Gastroenterol. Hepatol. 2014;11:683-691.

3. Rajilic-Stojanovic M, Jonkers DM, Salonen A, et al. Intestinal microbiota and diet in IBS: causes, consequences, or epiphenomena? Am. J. Gastroenterol. 2015;110:278-287.

4. Ohman L, Simrén M. Intestinal microbiota and its role in irritable bowel syndrome (IBS). Curr. Gastroenterol. Rep. 2013;15:323.

5. Kassinen A, Krogius-Kurikka L, Makivuokko H, et al. The fecal microbiota of irritable bowel syndrome patients differs significantly from that of healthy subjects. Gastroenterology 2007;133:24-33.

6. Salonen A, Vos WM de, Palva A. Gastrointestinal microbiota in irritable bowel syndrome: present state and perspectives. Microbiology 2010;156:3205-3215.

7. Simrén M, Barbara G, Flint HJ, et al. Intestinal microbiota in functional bowel disorders: a Rome foundation report. Gut 2013;62:159-176.

8. Rajilic-Stojanovic M, Biagi E, Heilig HGHJ, et al. Global and deep molecular analysis of microbiota signatures in fecal samples from patients with irritable bowel syndrome. Gastroenterology 2011 ;141:1792-1801.

9. Jeffery IB, O'Toole PW, Ohman L, et al. An irritable bowel syndrome subtype defined by species-specific alterations in faecal microbiota. Gut 2012;61:997-1006.

10. Durbán A, Abellán JJ, Jiménez-Hernández N, et al. Instability of the faecal microbiota in diarrhoea-predominant irritable bowel syndrome. FEMS Microbiol. Ecol. 2013;86:581-589.

11. Shankar V, Homer D, Rigsbee L, et al. The networks of human gut microbe-metabolite associations are different between health and irritable bowel syndrome. ISME J. 2015;9:1899-1903.

12. Duboc H, Rainteau D, Rajca S, et al. Increase in fecal primary bile acids and dysbiosis in patients with diarrhea-predominant irritable bowel syndrome. Neurogastroenterol. Motil. 2012;24:513-e247.

13. Codling C, O'Mahony L, Shanahan F, et al. A molecular analysis of fecal and mucosal bacterial communities in irritable bowel syndrome. Dig. Dis. Sci. 2010;55:392-397.

14. Ng SC, Lam EFC, Lam TTY, et al. Effect of probiotic bacteria on the intestinal microbiota in irritable bowel syndrome. J. Gastroenterol. Hepatol. 2013;28:1624-1631.

15. Durbán A, Abellán JJ, Jiménez-Hernández N, et al. Structural alterations of faecal and mucosa-associated bacterial communities in irritable bowel syndrome. Environ. Microbiol. Rep. 2012;4:242-247.

16. Rangel I, Sundin J, Fuentes S, et al. The relationship between faecal-associated and mucosal-associated microbiota in irritable bowel syndrome patients and healthy subjects. Aliment. Pharmacol. Ther. 2015;42:1211-1221.

17. Harrell L, Wang Y, Antonopoulos D, et al. Standard colonic lavage alters the natural state of mucosal-associated microbiota in the human colon. PloS One 2012;7:e32545.

18. Longstreth GF, Thompson WG, Chey WD, et al. Functional bowel disorders. Gastroenterology 2006; 130:1480-1491.

19. Francis CY, Morris J, Whorwell PJ. The irritable bowel severity scoring system: a simple method of monitoring irritable bowel syndrome and its progress. Aliment. Pharmacol. Ther. 1997;11:395-402.

20. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr. Scand. 1983;67:361-370.

21. Bohn L, Storsrud S, Liljebo T, et al. Diet low in FODMAPs reduces symptoms of irritable bowel syndrome as well as traditional dietary advice: a randomized controlled trial. Gastroenterology 2015;149:1399-1407.e2.

22. Heaton KW, O'Donnell LJ. An office guide to whole-gut transit time. Patients' recollection of their stool form. J. Clin. Gastroenterol. 1994;19:28-30.

23. Tornblom H, Van Oudenhove L, Sadik R, et al. Colonic transit time and IBS symptoms: what's the link? Am. J. Gastroenterol. 2012;107:754-760.

24. Matsuki T, Watanabe K, Fujimoto J, et al. Use of 16S rRNA Gene-Targeted Group-Specific Primers for Real-Time PCR Analysis of Predominant Bacteria in Human Feces. Appl. Environ. Microbiol. 2004;70:7220-7228.

25. Godon JJ, Zumstein E, Dabert P, et al. Molecular microbial diversity of an anaerobic digestor as determined by small-subunit rDNA sequence analysis. Appl. Environ. Microbiol. 1997;63:2802-2813.

26. Edgar RC. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 2013;10:996-998.

27. Yu Y, Lee C, Kim J, et al. Group-specific primer and probe sets to detect methanogenic communities using quantitative real-time polymerase chain reaction. Biotechnol. Bioeng. 2005;89:670-679.

28. Oksanen J, Blanchet F, Kindt R, et al. vegan: community ecology package. R package version 1.17-1, 2010. R Package Version 2010:1-17.

29. Arumugam M, Raes J, Pelletier E, et al. Enterotypes of the human gut microbiome. Nature 2011 ;473:174-180.

30. Holmes I, Harris K, Quince C. Dirichlet multinomial mixtures: generative models for microbial metagenomics. PloS One 2012;7:e30126.

31. Tibshirani R. Regression shrinkage and selection via the lasso: a retrospective. J. R. Stat. Soc. Ser. B Stat. Methodol. 2011;73:273-282.

32. Fan R-E, Chang K-W, Hsieh C-J, et al. LIBLINEAR: A Library for Large Linear Classification. J. Mach. Learn. Res. 2008;9:1871-1874.

33. Zeller G, Tap J, Voigt AY, et al. Potential of fecal microbiota for early-stage detection of colorectal cancer. Mol. Syst. Biol. 2014;10:766.

34. Dray S, Chessel D, Thioulouse J. Co-inertia analysis and the linking of ecological data tables. Ecology 2003;84:3078-3089.

35. Tap J, Furet J-P, Bensaada M, et al. Gut microbiota richness promotes its stability upon increased dietary fibre intake in healthy adults. Environ. Microbiol. 2015;17:4954-4964.

36. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B Methodol. 1995;57:289-300.

37. Dlugosz A, Winckler B, Lundin E, et al. No difference in small bowel microbiota between patients with irritable bowel syndrome and healthy controls. Sci. Rep. 2015;5:8508.

38. Pozuelo M, Panda S, Santiago A, et al. Reduction of butyrate- and methane-producing microorganisms in patients with Irritable Bowel Syndrome. Sci. Rep. 2015;5:12693.

39. Ringel Y, Maharshak N, Ringel-Kulka T, et al. High throughput sequencing reveals distinct microbial populations within the mucosal and luminal niches in healthy individuals. Gut Microbes 2015;6:173-181.

40. Carroll IM, Ringel-Kulka T, Keku TO, et al. Molecular analysis of the luminal- and mucosal-associated intestinal microbiota in diarrhea-predominant irritable bowel syndrome. Am. J. Physiol. Gastrointest. Liver Physiol. 2011;301:G799-807.

41. Yasuda K, Oh K, Ren B, et al. Biogeography of the intestinal mucosal and lumenal microbiome in the rhesus macaque. Cell Host Microbe 2015;17:385-391.

42. Chaudhary P, Gaci N, Borrel G, et al. Molecular methods for studying methanogens of the human gastrointestinal tract: current status and future directions. Appl. Microbiol. Biotechnol. 2015;99:5801-5815.

43. Dridi B, Henry M, El Khéchine A, et al. High prevalence of Methanobrevibacter smithii and Methanosphaera stadtmanae detected in the human gut using an improved DNA detection protocol. PloS One 2009;4:e7063.

44. Sahakian AB, Jee S-R, Pimentel M. Methane and the gastrointestinal tract. Dig. Dis. Sci. 2010;55:2135-2143.

45. Kim G, Deepinder F, Morales W, et al. Methanobrevibacter smithii is the predominant methanogen in patients with constipation-predominant IBS and methane on breath. Dig. Dis. Sci. 2012;57:3213-3218.

46. Vandeputte D, Falony G, Vieira-Silva S, et al. Stool consistency is strongly associated with gut microbiota richness and composition, enterotypes and bacterial growth rates. Gut 2016;65:57-62.

47. Falony G, Joossens M, Vieira-Silva S, et al. Population-level analysis of gut microbiome variation. Science 2016;352:560-564.

48. Knights D, Costello EK, Knight R. Supervised classification of human microbiota. FEMS Microbiol. Rev. 2011;35:343-359.

49. Forslund K, Hildebrand F, Nielsen T, et al. Disentangling type 2 diabetes and metformin treatment signatures in the human gut microbiota. Nature 2015;528:262-266.

50. Tap J, Mondot S, Levenez F, et al. Towards the human intestinal microbiota phylogenetic core. Environ. Microbiol. 2009;11:2574-2584.

51. Qin J, Li R, Raes J, et al. A human gut microbial gene catalogue established by metagenomic sequencing. Nature 2010;464:59-65.

52. Spanogiannopoulos P, Bess EN, Carmody RN, et al. The microbial pharmacists within us: a metagenomic view of xenobiotic metabolism. Nat. Rev. Microbiol. 2016;14:273-287.

Figures legends

Figure 1. Taxonomical and diversity analysis of fecal and mucosal microbiota in IBS patients and healthy subjects. A. Phylum relative abundance in stool and biopsies (paired Wilcoxon test between stool and biopsy, p<0.05). B. Alpha diversity measured by OTU number (paired Wilcoxon test, p<0.05). C. Similarity between fecal and mucosal microbiota measured by RV coefficient. Observed RV coefficient is illustrated by the red dash line. The black histogram shows the distribution of RV simulated coefficient from 99 Monte Carlo permutations (permutation test p<0.05, the difference between observed and simulated indicates that the observed RV is higher than what would be expected by chance, hence significant). D. Scatter plot of paired stool-biopsy microbiota from the co-inertia analysis using JSD metrics. Biopsy and fecal sample from the same subject are linked by a black line. PCs represent the two first principal components from the co-inertia analysis. Length of black line represents the distance between stool and biopsy from the same subject.

Figure 2. Enterotypes richness and distribution in healthy subjects and IBS patients.

A. Abundance of the main contributors of each enterotype supports previously described enterotypes (Clostridiales, Bacteroides and Prevotella) B. Microbial richness of the three detected enterotypes: Clostridiales, Bacteroides and Prevotella enterotypes (Pairwise Wilcoxon test, p<0.05). C. Enterotypes distribution in stool of healthy subjects and IBS subtypes (Pearson's Chi-squared test, p<0.05). D. Enterotypes distribution according to severity group (Pearson's Chi-squared test, p<0.05).

Figure 3. Univariate comparison (A) and machine learning (B) based on IBS severity. A. For each group, 30 samples per group (mild, moderate and severe IBS, and healthy controls, respectively) were randomly taken and each OTU was tested using a Wilcoxon test. This procedure was repeated 1,000 times. The number of significant OTU for each comparison is reported with boxplot. Dash red line illustrates the random expectation defined as 95th percentile of random comparison. B. Area under curve (AUC) of 1,000 classification models based on 10-fold CV bootstrapped 100 times. AUC is reported for exploratory set (50.4% sensitivity at the 80% specificity level), mucosal sample (82.9% sensitivity with 80% specificity), and validation set (39.4% sensitivity with 80% specificity). Dash red line illustrates the random expectation defined as AUC of 0.50.

Figure 4: Taxonomic assessment of OTUs microbiota signature for IBS symptom severity. Axis represent the two first components from principal coordinate analysis based on phylogenetic distance between OTUs representative sequences A. 90 OTUs (blue dots) are selected out of 2,911 OTUs (red dots) by the machine learning procedure. B. OTUs microbiota signature for IBS severity colored by taxonomical assignation at family level C. OTUs microbiota signature for IBS severity colored and sized by their weight and absolute weight in the model respectively (green corresponding to positive association to health and red to positive association to IBS severity). D. OTUs prevalence enrichment in microbiota from healthy subjects as function of their weight in the model (Wilcoxon test, p<0.05). Green boxplot represents positive weight while red boxplot represent negative weight.

Figure 5. Clinical and microbial ecology parameters interactions with microbial signature for IBS severity. A co-inertia analysis was undertaken between microbial signature OTUs relative abundance for IBS severity and clinical parameters. A.

Scatter plot of two first clinical data PC loadings. Each clinical parameter is labeled in red. B. Scatter plot of two first PC loadings for microbial signature OTUs relative abundance. Color accounts for weight in the model for IBS severity. Negative weights indicate OTUs associated with severe IBS. Size account for OTUs prevalence enrichment in healthy microbiota compared to severe IBS microbiota. Positive enrichment means that an OTU was found more frequently in healthy subjects than in severe IBS patients. For C,D,E,F scatter plot of two first component of co-inertia analysis. Each dot represents a fecal microbiota sample. C. Size accounts for microbial richness and colors for IBS symptoms severity group D. Size accounts for exhaled CH4 and color for Methanobacteriales presence. E. Color accounts for IBS subtypes. F. Each dot represents IBS fecal microbiota sized with IBS-SSS and colored by enterotypes.

Figure 6: Method and result graphical summary of the study. An exploratory (n=149) and a validation set (n=46) of IBS patients and healthy subjects were included in the study. DNA was extracted from fecal and sigmoid biopsy samples in order to assess gut microbiota by 16S rRNA gene sequencing. Methanobacteriales were detected by quantitative PCR in fecal samples. Most of IBS patients from exploratory set were sampled twice. Subsequent analyses of 16S sequencing data included various approaches: 1) Ecological analysis included alpha, beta-diversity assessment and enterotypes detection. 2) Machine learning procedure was used to select gut microbiota OTUs and to train models based on IBS severity on the exploratory set. The resulting microbial signature allowed classifying mucosal samples of the exploratory set- and the stool samples of the validation set based on IBS severity. 3) The microbial signature for IBS severity was taxonomically characterized and evaluated against clinical parameters. The microbial signature for

IBS severity was associated with low microbial richness, low CH4 exhaled, Bacteroides enterotypes enriched and absence of Methanobacteriales.

Table 1: Clinical and demographical characteristics for IBS patients and healthy subjects

Study set 1 (exploratory) Study set 2 (validation)

Median (interquartile) Healthy (n=39) IBS (n=110) Healthy (n=17) IBS (n= 29) Healthy set 1 vs set 2 p.value IBS set 1 vs set 2 p.value

Age (years) 27 (24 - 31) 33 (27 - 43) 29 (27 - 38) 28 (24 - 40) NS NS

Gender (M/F) 13/26 41/69 6/11 5/24 NS NS

BMI 21.9 (19.8 - 24.3) 22.6 (20.9 - 24.7) 22.4 (21.0 -23.9) 22.2 (20.3 -25.0) NS NS

Exhaled H2 (ppm) 8.0 (0 - 12.0) 2.0 (0 - 12.3) 4.0 (1.0 - 8.0) 5.5 (0 - 21.5) NS NS

Exhaled CH4 (ppm) 3.0 (1.0 - 6.5) 3.0 (1.0 - 7.5) 2.0 (1.0 - 2.3) 2.5 (1.0 - 6.3) NS NS

HAD anxiety 5 (2 - 7) 7 (5 - 11) 4 (2 - 6) 11 (7 - 14) NS NS

HAD depression 2 (1 - 4) 4 (2 - 7) 2 (1 - 4) 6 (3 - 9) NS NS

IBS subtype (C/D/M/U) N/A 18/43/43/2 (4 NA) N/A 3/14/9/1 (2 NA) N/A NS

IBS-SSS N/A 266 (198 - 359) N/A 270 (212 - 347) N/A NS

Number of patients with

severe IBS (IBS- N/A 45 N/A 13 N/A NS

SSS>300)

Stool consistency (BSF) 4.0 (3.5 - 4.4) 4.0 (2.8 - 4.7) 3.6 (2.9 - 4.4) 4.5 (3.3 - 5.2) NS NS

Stool freq. (stool per day) 1.2 (1.0 - 1.6) 1.5 (1.0 - 2.1) 1.4 (1.1 - 2.0) 1.3 (1.1 - 1.6) NS NS

OATT (days) 1.1 (0.7 - 1.7) 1.1 (0.6 - 1.8) 1.1 (1.0 - 1.5) 1.3 (0.9 - 2.1) NS NS

BMI: body mass index; BSF: Bristol Stool Form; OATT: oro-anal transit time; HAD: hospital anxiety and depression scale; IBS-C: irritable bowel syndrome with constipation; IBS-D: irritable bowel syndrome with diarrhea; IBS-M: irritable bowel syndrome with mixed pattern; IBS-U: irritable bowel syndrome unsubtyped; IBS-SSS: irritable bowel syndrome symptoms severity score ; N/A: not applicable; NA: not available; Data are shown as median (interquartile range); Statistical

significance was determined by nonparametric Wilcoxon test and Benjamini Hochberg multiplicity correction. Gender and IBS subtypes balance were evaluated with Chi-squared Test. NS indicate p value >0.05.

■a c

CEPTED MANUSCRIPT

Sample Type

E-3 Biopsy Stool

Bacteroidetes

Firmicutes

Phylum

Proteobacteria Actinobacteria

observed RV

Sample Type

% Biopsy Stool

Biopsy Stool

0.00 0.25 0.50 0.75 Simulated RV

0.2 PC1 74.9 %

75% ■

£ 25%-

Enterotypes 1 2 3

• ••

••

tio 75%

t o r e t n E

tio 75%

t o r e t n E

Healthy

Healthy

-£= O

o Z 150-l r o

.rç o

2 E ^100-

Clostridiales Bacteroides Prevotella

Enterotypes associated bacterial taxa

Enterotypes

mild IBS

moderate IBS

severe IBS

validation stool set

biopsy set ■

exploratory stool set

Number of significant OTUs per test

Signature OTUs selection

3 0.0-o

•ij»

• • »

• •

» •

gut microbiota OTUs

not selected selected

—i-1—

0.0 0.1 PCo1

o O Q.

Signature OTUs weight

• •

• « • • • • weight in the model 1.0

• • 0.5 0.0 -0.5 ■ "10

0.0 0.1 PCo1

Signature OTUs taxonomy

• d* •

-0.1 -

•• •

-0.2 H

Family tax.

Bacteroidaceae Lachnospiraceae Porphyromonadaceae Prevotellaceae Ruminococcaceae Subdominant Unassigned

i i 0.0 0.1 PCo1

o 0% -

11 0) o

<D Ë O =

C > <o s: ro ^

a) a) a1

^ ™ -20%

Signature OTUs prevalence

Negative Positive

signature OTUs weight for IBS severity

Clinical metadata

T3 TO O

Severe -----

AnxiD?Press

Stool_dayM

odera^

ACCEPTED MANUSCRIPT

M 0.25 H (0 O

Gut microbial OTUs

Healthy

PC1 loading

IBS severity and Microbial richness

^ 2H %

CM £ 0

—i-1-

PC1 29.5 %

IBS subtypes

—i— 3

^ 2H %

PC1 29.5 %

IBS-SS group

• Severe

• Moderate

• Mild

• Healthy

Microbial richness

• 50

• 100 0 150

IBS subtypes

# Healthy

# IBS-C IBS-D IBS-M

-0.25-

% • « ; •

• # • # A

• . aNT«* r#

OTUs prevalence

enrichement

in Healthy microbiota

-30% • -20% • -10% • 0% • 10%

weight in the model

-0.25 0.00 0.25 PC1 loading

gut Archea and Exhaled CH4

2 2 C P

Mr& v#

Methanobacteriales

• undetected

• presence

CH4 (ppm)

• 0 10 0 20 0 30

A 40 50

PC1 29.5 %

—i— 6

IBS severity and Enterotypes

2 2 C P

Enterotypes

• 1: Clostridiales

• 2: Bacteroides

• 3: Prevotella

IBS-SS Score

• 100 • 200 0 300 0 400

PC1 29.5 %

EXPLORATORY COHORT N=149

Samples collect: Stool (2X) and biopsy

VALIDATION COHORT N=46

Sample collect: Stool (1X)

DNA EXTRACTION

MICROBIOTA ANALYSIS

16S rRNA gene sequencing qPCR Methanobacteriales

1. ECOLOGICAL ANALYSIS

2. MACHINE LEARNING

3. MICROBIAL SIGNATURE

ALPHA AND BETA DIVERSITY

- Number of OTUs

- Distance within and between groups

"ENTEROTYPING"

Dirichiet multinomial model

LASSO procedure with 10-fold cross-validation combined with 10 bootstraps sampling

Taxonómica! characterization of signature

Clinical and microbial ecological parameters

SEVERE IBS

HEALTHY

LOW MICROBIAL RICHNESS LOW CH4 EXHALED

BACTEROIDES ENTEROTYPE ENRICHED

HIGH MICROBIAL RICHNESS HIGH CH4 EXHALED

CLOSTRIDIALES AND PREVOTELLA ENTEROTYPES ENRICHED

Identification of a gut microbial signature linked to severity of irritable bowel syndrome

Julien Tap, Muriel Derrien, Hans Tornblom, Rémi Brazeilles, Stéphanie Cools-Portier, Joël Doré, Stine Storsrud, Boris Le Nevé, Lena Ohman, Magnus Simrén

Supplementary information for online-only publication Questionnaires

Patients with IBS enrolled in the study completed questionnaires in order to characterize their symptom severity and bowel habits. The IBS Severity Scoring System (IBS-SSS) was used to assess the severity of IBS symptoms'. This is a well-validated questionnaire that is based on five items; intensity and frequency of abdominal pain, severity of abdominal distension, bowel habits dissatisfaction, and interference with daily life. The maximum score is 500 and patients can be categorized as having mild (<175), moderate (175-300), or severe (>300) IBS symptoms. General anxiety and depression was evaluated by the Hospital Anxiety and Depression (HAD) scale2. This is a 14-item questionnaire used to measure the severity of anxiety and depression on two subscales with 7 items each. Each item is scored between 0 and 3, with higher scores indicating more severe symptoms and with a total score range per subscale of 0-21.

The patients reported all bowel movements in a daily diary during two weeks, based on the Bristol stool form (BSF) scale3. The Stool consistency was scored on a 7-point scale: type 1 (separate hard lumps like nuts, difficult to pass), type 2 (sausage shaped but lumpy), type 3 (like a sausage but with cracks on surface), type 4 (like a sausage or snake, smooth and soft), type 5 (soft blobs with clear-cut edges), type 6

(fluffy pieces with ragged edges, a mushy stool) and type 7 (watery, no solid pieces, entirely liquid). Based on this information, the stool frequency (average number of stools/day) and stool consistency (average stool consistency/day) could be calculated. 111 subjects (89 IBS and 22 healthy) completed a food diary for four days (three weekdays and one day during the weekend). The diary included details regarding cooking methods, ingredients, brands of foods (if appropriate), time points for meals and quantity consumed in grams or household measurements. Patients were given written instructions to enable accurate completion of the food record. The subjects were instructed to consume their usual diet. Different food items and beverages were entered in DIETIST XP version 3.1 (Kostdata.se, Stockholm, Sweden), which converts food items into nutrients and energy amounts. Composite foods (e.g., casseroles) were split into ingredients (food items). DIETIST XP software covers around 1600 foods and 52 nutrients. DIETIST XP is designed to estimate macronutrients and micronutrients and energy intake. From the food records, average daily intakes were calculated for energy, proportion of fat, carbohydrates, and protein, and FODMAPs. All nutrients in the software DIETIST XP are based on food composition data from the National Food Administration in Sweden, except for the FODMAPs that were calculated using a new Swedish database for content of lactose, fructose, galacto-oligosaccharides, fructans and polyols in foods used in Swedish diets (Liljebo et al. Manuscript in preparation).

Oro-anal transit time measurement

For the oro-anal transit time (OATT) measurement, the participants ingested 10 radiopaque rings every morning for 5 days. On day 6, they ingested 5 radiopaque rings at 8:00 a.m. and 5 radiopaque rings at 8:00 p.m. in order to better define participants with accelerated transit. On the morning of day 7 the radiopaque rings

still present in the bowel were counted upon arrival at the laboratory, using fluoroscopy (Exposcop 7000 Compact, Ziehm GmbH, Nuremberg, Germany). OATT expressed in days was calculated by dividing the number of retained radiopaque rings by the daily dose (i.e. 10)4. All medications with known effects on the GI tract (proton pump inhibitors, laxatives, antidiarrheals, opioid analgesics, prokinetics, spasmolytics, antidepressants) were discontinued at least 48 hours before intake of the first radiopaque rings.

Breath CH4 and H2 measurement

This test was performed after an overnight fast, (i.e. not after intake of any substrate), and after the subjects had received thorough instructions to avoid a diet rich in fibre and poorly absorbed carbohydrates the day before the test. The amount of exhaled H2 and CH4 was measured in parts per million (ppm) in end-expiratory breath samples collected in a system used for the sampling and storing of alveolar air (GaSampler system, QuinTron Instrument Company, Milwaukee, WI, USA) and analyzed immediately using a gas chromatograph (QuinTron Breath Tracker, QuinTron Instrument Company, Milwaukee, WI, USA).

Collected data and missing values

196 subjects were included in this study. Information about gender, Age and BMI were available for all 196 subjects, exhaled H2 and CH4 in the fasting state, and HAD in 185 subjects, Bristol stool forms data in 166 subjects and OATT in 181 subjects. IBS severity and subtypes were available in 133 out of 139 IBS patients. Methanogens qPCR detection was performed on 231 fecal samples out of 278.

Microbial DNA extraction from biospies

Once collected, biopsies were immediately placed in liquid nitrogen and stored at -80° C until further analysis. Mucosal adherent microbiota DNA was isolated using

adapted protocol from Godon and colleagues5 for low biomass samples.

Each biopsy sample was transferred to a tube containing 250pL Guanidine Thiocyanate, 40pL N-lauroyl sarcosine 10% and 500pL N-lauroyl sarcosine 5%, and vortexed before incubation at 70°C for 1 hour. 200m g of Glass beads (0.001 mm) were added, and each tube was mixed for 10 minutes with the Vibrobroyeur (Retsch 25/s). 15mg of PVPP (polyvinylpolypyrrolidone) was added to the tubes that were

centrifuged for 5 min (12700 rpm, 4°C). The resulting supernatant was then transferred to a 2 mL sterile tube. The pellet was washed with 500 pL TENP (50 mM Tris pH8, 20 mM EDTA pH8, 100 mM NaCl, 1% de PVPP) and centrifuged for 5 min (12700 rpm, 4°C) This washing procedure was repeated twice and the resulting supernatants were pooled.

Nucleic acids were precipitated by adding 1mL of isopropanol in each tube. Samples were stored overnight at 4°C. After centrifugation for 1 hour (12700 rpm, 4°C), pellets were dissolved in 450pL of phosphate buffer (Na2HPO4, pH 8, 0.1M) and 50pL of potassium acetate (5M Acetate, 3M Potassium) for 1h 30 min at 4°C and then centrifuged (30 min, 12700 rpm, 4°C). Supernatants were transferred in a sterile tube with 2pL of RNase (10mg/mL). Tubes were incubated 30 min at 37°C. 50pL of sodium acetate and 1mL of 100% ethanol has been added and tubes were mixed gently. After being stored overnight at -20°C, tubes were centrifuged 1 hour (12700 rpm, 4°C). The resulting pellet was washed with 1 m L of ethanol (70%) and centrifuged for 5 min (12700 rpm). The supernatant was discarded and the washing procedure was repeated once. Once dried, up to 100pL of TE (10mM Tris Cl pH8, 1mM EDTA pH8) was added to each tube to dissolve the purified DNA. DNA was stored at -20°C until further analysis.

Microbial composition assessment

Quality filtering was done using SDM software6. Reads were further filtered for minimal and maximal length, any ambiguous nucleotides, barcode and primer errors and homopolymeric nucleotide runs. The default criteria parameter adapted to 454 sequencing platform were provided by LotuS. High quality sequence criterion (read average quality=27, minimal sequence length=250, no ambiguous bases, maximum of homopolymer=8, no mismatch allowed in primers barcode and primers, windows

quality threshold average of 25 from quality window of 50 bases) was used to build OTU. High and mid quality sequences were mapped to count the occurrence of established OTUs in single sample. OTU clustering at 97% identity was done with UPARSE which embedded UCHIME as chimera reads filterer. Each representative OTU sequence was aligned and taxonomically assigned using Greengenes database34 (release version 13.8 August 2013) and RDP II database35 (release version 11).

OTU prevalence and phylogenetic assessment

The prevalence of each OTU from the gut microbiota signature for IBS severity was estimated in both healthy subjects and severe IBS patients, and was defined as the proportion of subjects for whom a specific OTU was detected. The prevalence of each OTU in severe IBS patients was then subtracted from the prevalence of the same OTU in healthy subjects. Hence, OTUs with a positive value were enriched in healthy subjects, and OTUs with a negative value were enriched in severe IBS patients. To assess phylogenetic relationship between OTUs, pairwise nucleotides sequence identity were computed between OTU representative sequences using SeqinR R package7. A principal coordinate analysis was then carried out to assess phylogenetic specificity of the signature.

Co-inertia analysis and RV coefficient

Co-inertia analysis (COIA) is an ordination method for coupling two (or more) sets of parameters (e.g. clinical parameters and microbiota OTUs proportion) by looking at their linear combinations. Thus, co-inertia analysis enables the simultaneous ordination of several tables. COIA is related to other multivariate analysis such as canonical correlation analysis. In the case of COIA, the co-inertia (the sum of square

of co-variance) between the two sets is maximized and decomposed. Hence, the co-inertia value is a global measure of the co-structure between the two datasets. Co-inertia is high when the two sets vary together and low when they vary independently8.

Depending on the dataset, COIA is coupled with Principal Component Analysis (PCA) or correspondence analysis. In this study, we used COIA for two types of dataset coupling:

1. Two independent Principal Coordinate Analysis (PCoA) were computed based on fecal and mucosal microbiota composition JSD distance matrix and then subjected to a COIA (Figure 1).

2. Principal Component Analysis were computed on microbiota OTUs signature for IBS severity and successively coupled with a PCA computed from clinical parameters (Figure 5), diet and medication intake.

The overall relatedness of the two datasets was measured by the RV coefficient8. The RV-coefficient is the coefficient of correlation between two tables (in this study e.g. between the two fecal and mucosal microbiota JSD distance matrices). A Monte Carlo test was used to test the robustness of the RV-coefficient.

Statistical modeling by machine learning

We used a custom pipeline in R (R version 3.10) in order to extract the most discriminative features from fecal microbiota OTU composition to distinguish patients with severe IBS from the mild or moderate IBS and heathy controls. Here, like Zeller and colleagues, we used the LASSO logistic regression classifier9 implemented in

LIBLINEAR , because it generates a parsimonious classification model, which selects only few features out of a potentially very large set.

Briefly, our pipeline were as follows (see , for more details):

1. Feature transformation: We applied a log-transformation and subsequently standardized features (by centering to mean 0 and dividing by each features' standard deviation to which we added the 10th percentile of standard deviations across all features).

2. Partitioning data for tenfold stratified cross-validation (we resampled dataset partitions ten times to obtain more stable accuracy estimates).

3. Fitting a LASSO model on the training data of each cross-validation fold: The LASSO hyperparameter was optimized for each model in a nested fivefold cross-validation on the training subset using the area under the precision-recall curve as model selection criterion and also enforcing at least five nonzero coefficients.

4. Application of the trained LASSO models to obtain the corresponding cross-validation test predictions. Due to the resampled cross-validation (and also in external validation), there were several test predictions for each test examples. To get a single prediction score per example, we averaged all test predictions (from ten or 100 models in cross-validation or external validation, respectively).

5. Model evaluation using AUROC analysis: tenfold cross-validation repeated 10 times, we obtained mean test prediction scores, which we subjected to model performance analysis (Figure 3).

6. Model interpretation and marker extraction: Features (bacterial OTUs) with potential association with IBS severity were extracted as nonzero coefficients from all 100 LASSO models (trained in ten times resampled tenfold cross-validation).

Supplementary figures and table legends

Table S1: Fecal and mucosal sampling for IBS patients and healthy subjects by

study cohorts. Each row represents the number of samples collected by sample type, and the number within brackets displays the number of subjects in each group. Some healthy subjects and IBS patients in the exploratory set provided a second fecal sample (noted as sample #2).

Table S2: Distribution of IBS subtypes in IBS symptom severity group in the exploratory and validation set. The following numbers correspond to the distribution of subjects based on IBS severity (based on cut-off values in IBS-SSS) and IBS subtype in the exploratory set with validation set numbers within parenthesis. Regarding IBS subtypes distribution, no significant differences were observed between severe IBS and other IBS severity groups, or between the exploratory and validation set. No difference was observed between subtypes distribution within IBS severity groups compared to random distribution. (NA=information not available)

Table S3: Summary of Microbial OTU associated with IBS severity. The 90

OTUs that were selected by machine learning approaches are taxonomically described as well as their corresponding weight average in models.

Table S4: Univariate association between clinical parameters and the two first principal components (PC) from the co-inertia analysis made with gut microbial signature for IBS severity. Spearman rho values are indicated only when the p-value was below 0.05.

Table S5: Dietary intake. Data are shown as median (interquartile range)

Table S6: Medications in IBS patients. For each medication, a Wilcoxon test was carried out to test the association with IBS severity group.

Table S7: Methods description used in IBS microbiota studies.

Figure S1: Clinical and demographical distribution characteristics for IBS patients and healthy subjects in the two study cohorts. A. Density probability is represented by a violin plot and interquartile range is represented by boxplot in white. No difference was observed between exploratory and validation set. As expected, HAD anxiety and Depression scores were higher in IBS patients than in healthy subjects (p<0.05). B. Gender distribution in healthy subjects and IBS patients (p>0.05).

Figure S2: Alpha and beta diversity in fecal and mucosal microbiota across IBS subtypes and severity in gut microbiota. A. Microbial richness. No differences were observed between groups. B. beta-diversity calculated with JSD metrics. No differences were observed between groups except for IBS subtype for which mucosal microbiota from mild IBS subtype harbored lower richness than mucosal microbiota from other subtypes. C. Beta-diversity calculated with JSD metrics. No difference was observed between groups.

Figure S3: Identification of enterotype clusters using dirichlet multinomial mixture A. Optimum number of clusters in the microbiota dataset with Laplace and B. Bayesian information criterion parameters. C. Proportion of enterotype in the exploratory set.

Figure S4: Variation of clinical data within enterotypes among healthy subjects and IBS patients. A. Quantitative variables are represented regarding health status and enterotype stratification. Red, blue and green accounted for Clostridiales, Bacteroides and Prevotella enriched enterotypes, respectively. OATT was faster in Prevotella- enterotypes compared to other enterotypes in healthy subjects and in IBS patients (p<0.05). No differences were observed for other parameters (p>0.05). B. Enterotypes proportion in males and females (p<0.05). C. Enterotypes proportion in healthy subjects and IBS patients (p<0.05).

Figure S5: Association of Methanobacteriales with microbiota data and clinical data. A. Exhaled CH4 as a function of Methanobacteriales presence (p<0.05) B. Proportion of samples in enterotypes when Methanobacteriales were detected or not (p< 0.05) C. Proportion of healthy subjects and IBS subtypes samples based on detection of Methanobacteriales (p<0.05) D. Proportion of healthy subjects and IBS severity samples based on detection of Methanobacteriales (p<0.05). E. Exhaled CH4 as a function of IBS subtypes. F. Exhaled CH4 as a function of IBS severity. Density probability is represented by a violin plot and interquartile range is represented by boxplot in white.

Supplementary References

1. Francis CY, Morris J, Whorwell PJ. The irritable bowel severity scoring system: a simple method of monitoring irritable bowel syndrome and its progress. Aliment. Pharmacol. Ther. 1997;11:395-402.

2. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr. Scand. 1983;67:361-370.

3. Heaton KW, O'Donnell LJ. An office guide to whole-gut transit time. Patients' recollection of their stool form. J. Clin. Gastroenterol. 1994;19:28-30.

4. Törnblom H, Van Oudenhove L, Sadik R, et al. Colonic transit time and IBS symptoms: what's the link? Am. J. Gastroenterol. 2012;107:754-760.

5. Godon JJ, Zumstein E, Dabert P, et al. Molecular microbial diversity of an anaerobic digestor as determined by small-subunit rDNA sequence analysis. Appl. Environ. Microbiol. 1997;63:2802-2813.

6. Hildebrand F, Tadeo R, Voigt A, et al. LotuS: an efficient and user-friendly OTU processing pipeline. Microbiome 2014;2:30.

7. Charif D, Lobry J. SeqinR 1.0-2: a contributed package to the R Project for statistical computing devoted to biological sequences retrieval and analysis. In: Bastolla U, Porto M, Roman HE, et al., eds. Structural Approaches to Sequence Evolution. Biological and Medical Physics, Biomedical Engineering. Springer Berlin Heidelberg; 2007:207-232. Available at: http://dx.doi.org/10.1007/978-3-540-35306-5_10.

8. Dray S, Chessel D, Thioulouse J. Co-inertia analysis and the linking of ecological data tables. Ecology 2003;84:3078-3089.

9. Tibshirani R. Regression shrinkage and selection via the lasso: a retrospective. J. R. Stat. Soc. Ser. B Stat. Methodol. 2011;73:273-282.

10. Fan R-E, Chang K-W, Hsieh C-J, et al. LIBLINEAR: A Library for Large Linear Classification. J. Mach. Learn. Res. 2008;9:1871-1874.

11. Zeller G, Tap J, Voigt AY, et al. Potential of fecal microbiota for early-stage detection of colorectal cancer. Mol. Syst. Biol. 2014;10:766.

Table S1: Fecal and mucosal sampling for IBS patients and healthy subjects by study cohorts.

Each row represents the number of samples collected by sample type, and the number within brackets displays the number of subjects within each group. Some healthy subjects and IBS patients in the exploratory set provided a second fecal sample (noted as sample #2).

Study set 1 (exploratory) Study set 2 (validation) Total

Number of samples Healthy (n=39) (sample #1) Healthy (n=39) (sample #2) IBS (n=110) (sample #1) IBS (n=110) (sample #2) Total Set 1 Healthy (n=17) (sample #1) IBS (n=29) (sample #1) Total Set 2 microbiota sample included in this study

Stool (n=) 39 1 110 82 232 17 29 46 278

Mucosal biopsy (n=) 20 0 39 0 59 0 0 0 59

Table S2: Distribution of IBS subtypes in IBS symptom severity group in the exploratory and validation set.

The following numbers correspond to the distribution of subjects based on IBS severity (based on cut-off values in IBS-SSS) and IBS subtype in the exploratory set with validation set numbers within parenthesis. Regarding IBS subtypes distribution, no significant differences were observed between severe IBS and other IBS severity groups, or between the exploratory and validation set. No difference was observed between subtypes distribution within IBS severity groups compared to random distribution. (NA=information not available)

IBS-C IBS-D IBS-M IBS-U NA Total number

Mild IBS 5 (0) 7 (2) 5 (0) 1 (1) 1 (1) 19 (4)

Moderate IBS 5 (1) 13 (6) 19 (5) 1 (0) 1 (0) 39 (12)

Severe IBS 8 (2) 18 (6) 18 (4) 0 (0) 1 (1) 45 (13)

NA 0 (0) 5 (0) 1 (0) 0 (0) 1 (0) 7 (0)

Total number 18 (3) 43(14) 43 (9) 2 (1) 4 (2)

Table S3: Summary of Microbial OTU associated with IBS severity. The 90 OTUs that were selected by machine learning approaches are described with their weight in IBS severity signature model and with their associated taxonomy from Phylum to Genus. Negative weights were associated with severity.

OTU Median model weight in IBS severity signature Phylum Class Order Family Genus

OTU. OTI 1 3729 COQ 1,337281878 1 ni70CQ7CC Firmicutes Clostridia Clostridiales Lachnospiraceae ? Pl^cfrirlii irrt VV/III

O 1 U_ OTU OTI 1 _689 2098 2730 1,01/208/00 0,81919211 Firmicutes Bacteroidetes Erysipelotrichia Bacteroidia Clostridia Erysipelotrichales Bacteroidales Clostridiales Erysipelotrichaceae Bacteroidaceae Clostridium xviii Bacteroides

OTU_ OTU OTI 1 4082 0, /40555025 0,654223259 Firmicutes Bacteroidetes Bacteroidia Bacteroidales Bacteroidales Lachnospiraceae Rikenellaceae Lachnospiracea incertae sedis Alistipes

OTU_ OTU OTI 1 4235 6198 Ziono 0,63043254 0,596464694 n C77C137Q Bacteroidetes ? Bacteroidia ? Clostridia ? Clostridiales Porphyromonadaceae ? Butyricimonas ? ?

OTU_ OTU OTI 1 4903 2975 "7Q/I 0,5//51329 0,576535071 Firmicutes Firmicutes Clostridia Clostridiales Bacteroidales Lachnospiraceae Ruminococcaceae ? Parabacteroides

OTU_ OTU OTI 1 294 1678 0,572314559 0,565777309 Bacteroidetes Firmicutes Firmicutes Bacteroidia Clostridia Clostridia Clostridiales Clostridiales Porphyromonadaceae Ruminococcaceae Clostridiaceae 1 Faecalibacterium

OTU_ OTU rtji I 1720 5837 1 3C 0,539020346 0,524539098 n A QC1 /1C"73Q Firmicutes Clostridia Clostridia Clostridiales Clostridiales Ruminococcaceae Sarcina Clostridium IV

OTU OTU OTI 1 136 1439 707C 0,496145733 0,493847393 Firmicutes Firmicutes ? Clostridia ? Clostridiales Ruminococcaceae Lachnospiraceae Oscillibacter Coprococcus ?

OTU_ OTU OTI 1 2925 2353 OQQQ 0,493365666 0,48233137 H A 1 Q Firmicutes Clostridia ? Clostridiales Bacteroidales ? Lachnospiraceae Lachnospiracea_incertae_sedis

OTU_ OTU OTI 1 3888 2687 "7/tO/l 0,456156418 0,442558628 Bacteroidetes Firmicutes Firmicutes Bacteroidia Clostridia Q -if ¡1 li Bacilli Clostridiales Lactobacillales Bacteroidaceae Lachnospiraceae Bacteroides Clostridium XlVb ?

OTU_ OTU OTI 1 2404 _6103 7777 0,43885846 0,435824277 H QQHQQ/l CQ1 Firmicutes Clostridia Clostridia Clostridiales Clostridiales ? Lachnospiraceae Clostridium XlVb Blautia

OTU_ OTU OTI 1 2722 2775 CI Q1 0,380834681 0,374763341 n 371QrïO Q7Q Firmicutes Firmicutes ? ? ? Lachnospiraceae ? ? ?

OTU_ OTU 5181 6007 0,371302879 0,370239933 Firmicutes Firmicutes Clostridia ? Clostridiales ? Ruminococcaceae Faecalibacterium

OTU 4447 0,3684192 Firmicutes Clostridia Clostridiales Lachnospiraceae Lachnospiracea_incertae_sedis

OTU _6000 0,354465289 Firmicutes Clostridia Clostridiales Lachnospiraceae ?

OTU 16 0,339590645 Firmicutes Clostridia Clostridiales Lachnospiraceae Clostridium XlVa

OTU 6165 0,331824155 Bacteroidetes Bacteroidia Bacteroidales Prevotellaceae Paraprevotella

OTU 6044 0,330370323 ? ? ? ? ?

OTU 419 0,305595011 Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides

OTU 5307 0,297713437 Actinobacteria Actinobacteria Bifidobacteriales Bifidobacteriaceae ?

OTU 5642 0,283858822 Bacteroidetes Bacteroidia Bacteroidales Bacteroidaceae Bacteroides

OTU 3247 0,282681243 Firmicutes Clostridia Clostridiales Lachnospiraceae Blautia

OTU 3714 0,279193278 Bacteroidetes Bacteroidia Bacteroidales Prevotellaceae ?

OTU 6012 5017 0,270631706 0,263565195 Bacteroidetes Bacteroidia Bacteroidales Lachnospiraceae Prevotellaceae ?

OTU 2703 150 0,252047835 0,251828113 Bacteroidetes ? ? Lachnospiraceae ? ? ? ? ?

OTU 2735 2454 -0,252509919 Firmicutes Bacilli Lactobacillales ? ?

OTU 1493 -0,259083798 -0,26143299 Firmicutes Firmicutes Clostridia Clostridia Clostridiales Clostridiales Ruminococcaceae ? ?

OTU 4664 3951 -0,263932811 -0,272024007 Firmicutes Bacteroidetes Clostridia Bacteroidia Clostridia Bacter Clostr ■oidales Lachnospiraceae Bacteroidaceae ? Bacteroides ?

OTU 5312 3985 -0,273754775 -0,284546483 Firmicutes Firmicutes Clostridia Clostridia Clostr Clostr diales Lachnospiraceae Clostridiaceae 1 Clostridium sensu stricto ?

OTU 5905 3669 -0,285672158 -0,288162056 Firmicutes Firmicutes Clostridia Clostr diales Lachnospiraceae ? ?

OTU 509 2228 -0,292558747 -0,303818478 Firmicutes Firmicutes Clostridia Clostridia Clostr Clostr diales diales Lachnospiraceae Lachnospiraceae ?

OTU 5812 2174 -0,318229455 -0,318523681 Firmicutes Firmicutes Clostridia Clostridia Clostridia Clostr Clostr Clostr diales diales ? Lachnospiraceae ? ? ?

OTU 3962 659 -0,318709825 -0,326260985 Firmicutes Firmicutes Clostridia Clostr diales Lachnospiraceae Lachnospiraceae ?

OTU OTU 1436 _3331 -0,330308014 -0,333858778 Bacteroidetes Bacteroidetes Bacteroidia Bacteroidia Bacteroidales Bacteroidales Bacteroidaceae ? Bacteroides ?

OTU 3058 5207 -0,334858638 -0,339855891 Firmicutes Bacteroidetes Clostridia Bacteroidia Bacteroidales Ruminococcaceae Porphyromonadaceae ? Parabacteroides

OTU 1142 2276 -0,348075159 -0,365971574 Firmicutes Bacteroidia Clostridia Bacteroidales Clostridiales Lachnospiraceae Roseburia

OTU 5616 -0,373203564 Firmicutes ? ? ? ?

OTU 3219 -0,378415864 Firmicutes Clostridia Clostridiales Lachnospiraceae Clostridium XlVa

OTU 4057 -0,388786328 Firmicutes Erysipelotrichia Erysipelotrichales Erysipelotrichaceae Turicibacter

OTU 37 -0,401468615 Firmicutes Clostridia Clostridiales Lachnospiraceae Lachnospiracea_incertae_sedis

OTU 5532 -0,408924645 Firmicutes Bacilli Lactobacillales Lactobacillaceae Lactobacillus

OTU 3094 -0,41916313 Firmicutes Clostridia Clostridiales Ruminococcaceae Ethanoligenens

OTU 153 -0,419607879 Firmicutes Clostridia Clostridiales Lachnospiraceae Roseburia

OTU 4634 -0,43718205 Firmicutes ? ? ? ?

OTU 4556 -0,453131831 Firmicutes Clostridia Clostridiales ? ?

OTU 2958 -0,465579013 Firmicutes Clostridia Clostridiales Ruminococcaceae ?

OTU 6049 4743 -0,466289504 -0,469490869 Bacteroidetes Bacteroidia Bacteroidia Bacteroidales Bacteroidales Bacteroidaceae Bacteroidaceae Bacteroides Bacteroides

OTU 5327 3211 -0,475131552 -0,521293782 Firmicutes Bacteroidetes Clostridia Bacteroidia Clostridia Clostridiales Bacteroidales Clostridiales Porphyromonadaceae Parabacteroides

OTU 3496 4070 -0,525534943 -0,540626283 Firmicutes Firmicutes Bacilli Lactobacillales Lachnospiraceae Lactobacillaceae Lactobacillus

OTU 978 424 -0,558817877 -0,574737029 Firmicutes ? ? ? ? ?

OTU 4481 3934 -0,578996037 -0,647352345 Bacteroidetes Bacteroidia Bacteroidia Bacteroidia Bacteroidales Bacteroidales Bacteroidales Bacteroidaceae Bacteroides

OTU 116 5071 -0,677742382 -0,717542886 Firmicutes Clostridia Clostridia Clostridiales Clostridiales Lachnospiraceae ?

OTU 2654 1822 -0,752515322 -0,798469741 Firmicutes Firmicutes Clostridia Clostridiales Lachnospiraceae Lachnospiraceae Roseburia

OTU 5215 1108 -0,849017242 -0,860978754 Bacteroidetes Bacteroidia Bacteroidia Bacteroidales Bacteroidales Bacteroidaceae Porphyromonadaceae Bacteroides Barnesiella

OTU 3706 1395 -0,97691772 -1,059585164 Firmicutes Bacteroidetes Clostridia Bacteroidia Clostridiales Bacteroidales Lachnospiraceae Bacteroidaceae Bacteroides

OTU 3527 -1,262778259 Firmicutes Clostridia Clostridiales ? ?

Table S4: Univariate association between clinical parameters and the two first principal components (PC) from the co-inertia analysis made with gut microbial signature for IBS severity.

Spearman rho values are indicated only when the p-value was below 0.05.

Clinical parameters Co-inertia PC1 (rho) Co-inertia PC2 (rho)

Age - -

BMI - -0,23

Exhaled H2 - -

Exhaled CH4 0,36 0,44

HAD anxiety -0,4 -

HAD depression -0,32 -

Stool consistency - -0,36

Stool frequency - -0,22

OATT - 0,54

Table S5: Dietary intake

Data are shown as median (interquartile range).

Health Group Energy intake (kcal) PROTEIN (%) FAT (%)_CARBOHYDRATE (%) Total FODMAPs (g)

Healthy control 2228(1860 - 2846) 14.65 (12 - 16.2) 35.15 (31.6 - 40.0) 45.15 (41.2 - 51.2) 17.7 (10.2 - 29.9)

Mild IBS 2188(2039 - 2435) 15.2 (13.6 - 16.4) 33.1 (30.5 - 35.3) 47.5 (44.8 - 51.4) 17.4 (9.3 - 20.6)

Moderate IBS 2295(1799 - 2482) 16.7 (15.3 - 18.7) 36.05 (30.8 - 42.8) 41.5 (35.2 - 47.3) 13.7 (8.9 - 20.7)

Severe IBS 1990(1477 - 2462) 16.1 (13.7 - 18.5) 37.4 (33.9 - 41.4) 45.7 (40.7 - 49.4) 13.3 (10.0 - 21.4)

Table S6: Medications in IBS patients

For each medication, a Wilcoxon test was carried out to test the association with IBS severity group.

Medication Mild IBS (n=21) Moderate IBS (n=48) Severe IBS(n=55)

Laxative / Bulking agent (p>0.05) 4 4 6

PPI / Acid suppression (p>0.05) 2 7 6

Antidiarrhoeals (p>0.05) 1 4 2

Antidepressants (p<0.05) 1 5 13

Table S7: Methods in IBS microbiota studies

Rajilic-Stojanavic1

Jeffery2

Maccaferri

Pozuelo

Zhernakova

This Study

Year of publication

Cohort description

Rome foundation (IBS) Site of investigation

IBS subtypes proportion % (C/D/M/U/A)

62 IBS (primary

care) and 46 healthy controls

Rome II Finland

29/40/31/0/0

37 IBS patients (secondary care) and 20 healthy controls

Rome II

Sweden

27/40/0/0/33

19 IBS patients and 34 healthy controls

Rome III Italy

21/53/26/0/0

236 IBS patients and 297 healthy controls

Rome II and III

Norway, Sweden, Denmark, Spain

17/44/4/11/22

113 IBS patients and 66 healthy controls

Rome III Spain

28/48/24/0/0

Population-based cohort (1,135 participants) and including 112 IBS patients

Unknown Netherland

Not Available

139 IBS patients (tertiary care) and 56 healthy controls

Rome III Sweden

15/41/37/2/0

IBS-C ■ IBS-D IBS-M m IBS-U BlBS-A

Stool collect and DNA extraction

Home freezer -20°C and -45°C, Wizard Genomic DNA purification kit

Home freezer -20°C then -80°C, Qiagen QIAamp DNA stool kit

Stored in anaerobic containers and frozen to -70°C, QIAamp DNA Stool Mini Kit (Qiagen)

Homogenization magnetic beads

Home freezer 4°C and -80°C, Mechanical lysis and guanidine/sarcosine based extraction

Home freezer then -80°C, AllPrep DNA/RNA Mini Kit Qiagen

RNA Later up to 3 weeks then -80°C, Mechanical lysis and phenolchloroform based extraction

Microbiota methods 16S region

Phylogenetic microarray (HITChip) - V1 and V6, qPCR Archaea

16S gene pyrosequencing, V4 region

Phylogenetic microarray

(HTFMicrobi.Array, whole 16S) and DGGE (V2-V3)

Probes/Array V3-V7

16S gene pyrosequencing, V4 region

Illumina MiSeq 16S sequencing (V4 region) and HiSeq whole metagenomic

16S gene pyrosequencing V5-V6 region

Bioinformatics/statistics approaches

Pearson correlation, Ward distance, RDA-MonteCarlo

Qiime, PCoA, CA, Univariate test, RDA- PCA and

Unifrac distance, MonteCarlo, Euclidean univariate

univariate test distance, ANOVA statistics

Qiime, univariate test, PCoA, Unifrac distance

MetaPhlan and HUManN2, Bray-Curtis distance, Multivariate Association with Linear Model, univariate test,

Permutational Multivariate Analysis of Variance

LOTUS with USEARCH v7, JSD and Bray-Curtis distance, Machine learning with LASSO, Bayesian Multimodel mixture, Co-inertia Analysis.

1. Rajilic-Stojanovic M, Biagi E, Heilig HG, Kajander K, Kekkonen RA, Tims S, de Vos WM. Global and deep molecular analysis of microbiota signatures in fecal samples from patients with irritable bowel syndrome. Gastroenterology. 2011 Nov;141(5):1792-801. doi: 10.1053/j.gastro.2011.07.043.

2. Jeffery IB, O'Toole PW, Öhman L, Claesson MJ, Deane J, Quigley EM, Simren M. An irritable bowel syndrome subtype defined by species-specific alterations in faecal microbiota. Gut. 2012 Jul;61(7):997-1006. doi: 10.1136/gutjnl-2011-301501.

3. Maccaferri S, Candela M, Turroni S, Centanni M, Severgnini M, Consolandi C, Cavina P, Brigidi P. IBS-associated phylogenetic unbalances of the intestinal microbiota are not reverted by probiotic supplementation. Gut Microbes. 2012 Sep-Oct;3(5):406-13.

4. Casen C, Veb0 HC, Sekelja M, Hegge FT, Karlsson MK, Ciemniejewska E, Dzankovic S, Fr0yland C, Nestestog R, Engstrand L, Munkholm P, Nielsen OH, Rogler G, Simren M, Öhman L, Vatn MH, Rudi K. Deviations in human gut microbiota: a novel diagnostic test for determining dysbiosis in patients with IBS or IBD. Aliment Pharmacol Ther. 2015 Jul;42(1):71-83. doi: 10.1111/apt.13236.

5. Pozuelo M, Panda S, Santiago A, Mendez S, Accarino A, Santos J, Guarner F, Azpiroz F, Manichanh C. Reduction of butyrate- and methane-producing microorganisms in patients with Irritable Bowel Syndrome. Sci Rep. 2015 Aug 4;5:12693. doi: 10.1038/srep12693.

6. Zhernakova A, Kurilshikov A, Bonder MJ, Tigchelaar EF, Schirmer M, Vatanen T, Mujagic Z, Vila AV, Falony G, Vieira-Silva S, Wang J, Imhann F, Brandsma E, Jankipersadsing SA, Joossens M, Cenit MC, Deelen P, Swertz MA; LifeLines cohort study, Weersma RK, Feskens EJ, Netea MG, Gevers D, Jonkers D, Franke L, Aulchenko YS, Huttenhower C, Raes J, Hofker MH, Xavier RJ, Wijmenga C, Fu J. Population-based metagenomics analysis reveals markers for gut microbiome composition and diversity. Science. 2016 Apr 29;352(6285):565-9. doi: 10.1126/science.aad3369.