International Journal of Molecular Sciences Article Intercontinental Gut Microbiome Variances in IBD Luis Mayorga 1,2,3,† , Gerard Serrano-Gómez 1,† , Zixuan Xie 1,2 , Natalia Borruel 1,3 and Chaysavanh Manichanh 1,2, * 1 Microbiome Lab, Vall d’Hebron Institut de Recerca (VHIR), Vall d’Hebron Hospital Universitari, Vall d’Hebron Barcelona Hospital Campus, Passeig Vall d’Hebron 119-129, 08035 Barcelona, Spain 2 Departament de Medicina, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallès, Spain 3 Crohn’s and Colitis Attention Unit, Digestive System Service, Hospital Vall d’Hebron, 08035 Barcelona, Spain * Correspondence: cmanicha@gmail.com; Tel.: +34-934-894-036 † These authors contributed equally to this work. Abstract: The development of biomarkers for inflammatory bowel disease (IBD) diagnosis would be relevant in a generalized context. However, intercontinental investigation on these microbial biomarkers remains scarce. We examined taxonomic microbiome variations in IBD using published DNA shotgun metagenomic data. For this purpose, we used sequenced data from our previous Spanish Crohn’s disease (CD) and ulcerative colitis (UC) cohort, downloaded sequence data from a Chinese CD cohort, and downloaded taxonomic and functional profiling tables from a USA CD and UC cohort. At the global level, geographical location and disease phenotype were the main explanatory covariates of microbiome variations. In healthy controls (HC) and UC, geography turned out to be the most important factor, while disease intestinal location was the most important one in CD. Disease severity correlated with lower alpha-diversity in UC but not in CD. Across geography, alpha-diversity was significantly different independently of health status, except for CD. Despite recruitment from different countries and with different disease severity scores, CD patients may harbor a very similar microbial taxonomic profile. Our study pointed out that geographic location, disease activity status, and other environmental factors are important contributing factors in Citation: Mayorga, L.; microbiota changes in IBD. We therefore strongly recommend taking these factors into consideration Serrano-Gómez, G.; Xie, Z.; Borruel, for future IBD studies to obtain globally valid and reproducible biomarkers. N.; Manichanh, C. Intercontinental Gut Microbiome Variances in IBD. Keywords: microbiome; IBD; geography Int. J. Mol. Sci. 2022, 23, 10868. https://doi.org/10.3390/ ijms231810868 Academic Editor: Alfredo Miccheli 1. Introduction Received: 29 August 2022 Inflammatory bowel diseases (IBDs), which include Crohn’s disease (CD) and ulcer- Accepted: 15 September 2022 ative colitis (UC), are multifactorial, chronic conditions of the gastrointestinal tract that Published: 17 September 2022 have been associated with gut microbiome alterations [1–9]. Reduced microbial community richness and diversity, as well as reduction of beneficial microbes and expansion of a variety Publisher’s Note: MDPI stays neutral of potentially harmful bacterial species, have been consistently reported in IBD, particularly with regard to jurisdictional claims in in CD patients. However, diet, drugs, ethnicity, geography, and a multitude of lifestyle or published maps and institutional affil- environmental variables may confound the interpretation and replication of microbiome iations. studies. This is particularly problematic with small cohort size studies in which these covari- ates may mask the disease effect. The importance of geography and related environmental exposure has been well-illustrated in migration studies, where a strong association between Copyright: © 2022 by the authors. microbiome diversity and migration was clearly demonstrated [10,11]. Vangay et al. [11] Licensee MDPI, Basel, Switzerland. performed 16S and deep shotgun metagenomic sequencing on stool samples collected This article is an open access article from individuals living in Thailand and the USA, including first- and second-generation distributed under the terms and immigrants before and after immigration. Intriguingly, USA immigration was found to be conditions of the Creative Commons associated with significant alterations in the gut microbiome, including loss of diversity, Attribution (CC BY) license (https:// bacterial strains, fiber degradation functions and a shift from Prevotella to Bacteroides strains. creativecommons.org/licenses/by/ Commonly, and quite consistently, the relative abundance of Enterococcus, Fusobac- 4.0/). terium, Streptococcus, Escherichia coli, and Ruminococcus gnavus has been found positively Int. J. Mol. Sci. 2022, 23, 10868. https://doi.org/10.3390/ijms231810868 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2022, 23, 10868 2 of 15 associated with IBD, while that of Faecalibacterium prausnitzii spp., and members of Rose- buria was negatively related to IBD [6,12–17]. However, some bacterial taxa identified as biomarkers for IBD were also found in other chronic diseases including type 2 diabetes (T2D). For instance, an enrichment in Christensenellaceae and Escherichia coli was linked to CD and T2D-associated dysbiosis [6,18,19], questioning the specificity of the available microbiome signatures for disease discrimination. Based on published datasets including prospective and cross-sectional population and patient cohorts [15,19–21], Metwaly et al. [15] reviewed the distribution of microbial profiles in IBD and T2D across different geographical regions. The authors identified geography as a major factor in the taxonomic composition variation of the gut microbiome. Another large longitudinal intercontinental study, using the 16S approach in 531 patients with IBD from Ireland and Canada, showed that geographic location was also the major determinant of microbiome variations, although most (90.3%) of the compositional variance remains unexplained [22]. Altogether, these studies emphasized the need to take into account this variable for the identification of specific disease biomarkers. Despite the fast- growing microbiome field, studies comparing intercontinental geographical differences in IBD gut microbiome remain scarce, probably due to the presence of confounders such as environmental and host-related factors and a lack of sample processing standards. In the present study, we aimed to investigate microbiome variations in IBD across countries using published DNA shotgun metagenomic data. Our results highlight the clinical, anthropometric, and demographic data that were commonly available and have allowed us to perform association analysis with microbiome variations to elucidate how these covariates may impact on taxonomic variations. We also emphasized the lack of standardized clinical data that has prevented us from integrating cohorts into this interconti- nental study or has limited the interpretation of the results. Our findings also demonstrated that, despite recruitment from different countries and with different disease severity scores, CD patients may harbor a very similar microbial taxonomic profile. 2. Results 2.1. Participants’ Characteristics and Fecal Samples To evaluate microbiome variation across countries, we compared our partially pub- lished Spanish IBD cohorts [23,24] with the USA [14] and Chinese [25] published cohorts. Characteristics of the cohorts are described in Table 1. For the Spanish and Chinese cohorts, we generated taxonomic and functional profiles from the shotgun metagenomic sequence reads; for the USA cohort, we downloaded these profiles from the inflammatory bowel disease multi-omics database (see Methods section). In total, 1470 fecal samples were included, 625 from patients with CD, 384 from patients with UC, and 461 from HC (or non-IBD) (Table 1). Table 1. Participants’ characteristics. United States Spain China Type of CD UC Non-IBD Total CD UC Non-IBD Total CD UC Non-IBD Total Cohort Number of 37 22 20 79 33 33 67 133 62 NA 52 114 subjects Number of 506 319 342 1167 57 65 67 189 62 NA 52 114 samples CD location (n) 109/45 18/2 L1/L1+L4 NA NA NA NA NA NA NA NA NA NA L2/L2+L4 58/23 0/0 L3/L3+L4 128/143 35/2 Gender (% 39.3 65.8 50.9 50.0 66.7 72.3 56.7 65.1 25.8 NA 5.8 5.8 female) Age (mean 22.5 ± 14.2 29.2 ± 19.5 27.5 ± 19.5 27.1 ± 17.7 35.8 ± 11.8 40.1 ± 10.5 43.1 ± 15.9 27.5±16.9 28.4 ± 8.1 NA 20.5 ± 7.8 20.5 ± 7.8 ± SD) BMI (mean ± 23.3 ± 7.5 23.0 ± 6.8 23.4 ± 7.0 23.3 ± 6.8 22.4 ± 4.0 23.6 ± 4.1 25.0 ± 4.3 23.2±6.7 18.6 ± 2.8 NA 21.4 ± 3.3 21.4 ± 3.3 SD) L1: terminal ileum, L2: colon, L3: ileocolon, L4: upper GIT (gastrointestinal tract). Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 3 of 16 Int. J. Mol. Sci. 2022, 23, 10868 3 of 15 2.2. Fecal Microbiome Variation Explained by Covariates in the Metadata 2.2. Fecal Microbiome Variation Explained by Covariates in the Metadata For the whole dataset (n = 1470), the most commonly reported variables from the For themetadata published whole dataset (n = country, included 1470), theage, most commonly disease reported activity index,variables body massfrom the index published metadata included country, age, disease activity index, body mass (BMI), and gender. To evaluate the impact of these covariates on microbiome variation,index (BMI), and we gender. appliedTotheevaluate the impact of PERMANOVA these covariates method (adonis on microbiome function) variation, to the we applied unweighted and the PERMANOVA method (adonis function) to the unweighted and weighted UniFrac distances between samples. While country and disease activity weighted UniFrac status distances contributedbetween samples. to most While country of the microbiome and disease variations, activity status the remaining contributed covariates to mosta also showed of the microbiome variations, significant impact (Figure 1). the remaining covariates also showed a significant impact (Figure 1). Figure 1. Covariates explaining fecal microbiome variance (PERMANOVA with unweighted (left Figure 1. Covariates explaining fecal microbiome variance (PERMANOVA with unweighted (left plot)/weighted plot)/weighted (right (right plot) plot) UniFrac UniFrac distances, distances, 999 999 permutations). permutations). AllAll covariates covariates have have aasignificant significant impact impact on microbiome variation (q < 0.05), except for BMI in HC (weighted UniFrac),gender on microbiome variation (q < 0.05), except for BMI in HC (weighted UniFrac), genderin inCD CD and andUCUC(weighted (weightedUniFrac). UniFrac).Whole Wholecohort cohort(n(n== 1408), 1408), HC HC (n (n == 461), 461), UC UC (n (n == 384), CD (n = 563). For the HC group, available covariates included country, age, gender, and BMI; for For the HC group, available covariates included country, age, gender, and BMI; for CD, disease localization, country, age, gender, BMI, and Harvey Bradshaw index (HBI); and CD, disease localization, country, age, gender, BMI, and Harvey Bradshaw index (HBI); for UC, country, age, gender, BMI, and colitis activity index (CAI) (Table 1, Supplementary and for UC, country, age, gender, BMI, and colitis activity index (CAI) (Table 1, Sup‐ Table S1). For HC and UC, country appeared to be the most significant explanatory variation plementary Table S1). For HC and UC, country appeared to be the most significant ex‐ factor, whereas for CD, disease localization (based on the Montreal classification system) planatory variation factor, whereas for CD, disease localization (based on the Montreal classification system) explained most of the variation (Figure 1). Both CAI (in UC) and HBI (in CD) indices had a significant effect on microbiome variation, however, CAI ap‐ Int. J. Mol. Sci. 2022, 23, 10868 peared to have a higher impact than HBI. Gender, BMI, and age presented a much lower 4 of 15 effect than country and disease. As all these covariates had a significant impact on mi‐ crobiome composition, they were taken into account for the differential abundance analysis described below. explained most of the variation (Figure 1). Both CAI (in UC) and HBI (in CD) indices had a significant effect on microbiome variation, however, CAI appeared to have a higher impact 2.3. Geographical and Disease‐Related Dysbiosis than HBI. Gender, BMI, and age presented a much lower effect than country and disease. As allInthese this study, dysbiosis covariates had aissignificant defined asimpact a deviation from a healthy on microbiome microbiome composition, theycom‐ were position taken intoataccount the alpha‐for and beta‐diversity the differential levels. To abundance assess alpha‐diversity, analysis described below.we calculated the Chao1 and Shannon indices [14,26,27]. The Chao1 richness estimator gives weight to the low‐abundant 2.3. Geographical and species as it only Dysbiosis Disease-Related accounts for singletons and doubletons, whereas the Shannon index accounts for both In this study, dysbiosis is defined the abundance and evenness as a deviation from aofhealthy the species present. The microbiome com- dysbiosis score is defined as the median weighted or unweighted UniFrac position at the alpha- and beta-diversity levels. To assess alpha-diversity, we calculated distance of any given the sample Chao1 andagainst Shannon a reference set built using indices [14,26,27]. The samples from HCs Chao1 richness [24]. estimator gives weight to One common diversity‐based observation that could be the low-abundant species as it only accounts for singletons and doubletons, drawn from the comparison whereas the of CD, UC, Shannon andaccounts index HC in the forthree both geographical the abundance cohorts was thatofCD and evenness thedisplayed a signifi‐ species present. The cantly lower dysbiosis scoremicrobial is defined diversity based on as the median Shannon weighted or and Chao1 indices unweighted UniFrac compared distance ofwith any HC and UC (Figure 2). UC was associated with lower richness given sample against a reference set built using samples from HCs [24]. compared with HC in both the Spanish One commonand thediversity-based USA cohorts but presented that observation lower evenness could compared be drawn with from the HC only comparison in the USA cohort, which suggested the contribution of dominant taxa of CD, UC, and HC in the three geographical cohorts was that CD displayed a significantly in UC dysbiosis in the USA cohort. lower microbial diversity based on Shannon and Chao1 indices compared with HC and CD exhibited UC (Figure 2). UClower richness (Chao1 was associated index) richness with lower compared with UC with compared in bothHCtheinSpanish both the and the USA Spanish and thecohorts, but lower USA cohorts butevenness presented (Shannon index) only lower evenness in the Spanish compared with HCcohort, only in which the USA again suggested cohort, the important which suggested contributionofof the contribution dominant dominant taxataxa to the in UC differences dysbiosis in the between USA cohort.CD and UC in the Spanish cohort (Figure 2A). Figure 2.2. Alpha-diversity across cohorts. Alpha‐diversity across cohorts.(A) (A)InInthe theSpanish Spanishcohort, cohort,Shannon Shannonand andChao1 Chao1indices indices showed aa significant significant decrease decrease in inboth bothmicrobial microbialrichness richnessand andevenness evennessininCD. CD.(B) (B)InInthe the USA USA co‐ cohort, hort, the the differences differences in Shannon in Shannon andindices and Chao1 Chao1 were indices were maintained maintained when comparing when comparing CD with HC, CD but with not HC, but when not whenCD comparing comparing with UC.CD(C)with The UC. (C) The Chinese Chinese cohort cohort also also showed showed decreased decreased richness andrichness evenness and evenness in CD compared with HC. in CD compared with HC. CD exhibited lower richness (Chao1 index) compared with UC in both the Spanish and the USA cohorts, but lower evenness (Shannon index) only in the Spanish cohort, which again suggested the important contribution of dominant taxa to the differences between CD and UC in the Spanish cohort (Figure 2A). Although we observed a greater severity of both CD and UC in the USA cohort compared with the Spanish cohort (Figure 3A), we could appreciate a lower diversity only in the USA UC cohort and not in the USA CD cohort (Figure 4B,C). Furthermore, Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 5 of 16 Int. J. Mol. Sci. 2022, 23, 10868 5 of 15 Although we observed a greater severity of both CD and UC in the USA cohort compared with the Spanish cohort (Figure 3A), we could appreciate a lower diversity only in the USA UC cohort and not in the USA CD cohort (Figure 4B,C). Furthermore, a a significant significant negative negative correlation correlation between between CAI CAI and Shannon and both both Shannon and Chao1 and Chao1 indices indices (Fig‐ (Figure ure 3B) 3B) waswas found found for for UC,UC, while while no no significant significant correlation correlation was was observed observed between between HBIHBI and CD and CD (Figure (Figure 3C). 3C). These These results resultsindicated indicatedthat thatdisease diseaseseverity severitywas wasassociated associatedwithwith diversity in diversity in UC, UC,but butnot notin inCD. CD. Figure 3. Figure Diseaseactivity 3. Disease activityscore scoreininCD CD(HBI, (HBI,Harvey‐Bradshaw Harvey-Bradshawindex) index) and and UCUC (CAI, (CAI, colitis colitis activity activity score). score). (A) (A)HBI HBIandandCAI CAIwere wereboth bothhigher higherininthe theUSA USAthan thanininthe theSpanish Spanishcohort cohort(Mann (Mann Whitney’s Whitney’s U U test). test). (B) (B)Correlation Correlationbetween betweenalpha‐diversity alpha-diversity(Shannon (Shannonand andChao1 Chao1 indices) and indices) andCAI CAI(Spearman (Spearman correlation test). rho = −0.019, p = 0.65 for the Shannon index and rho = −0.0097, p = 0.82 correlation test). rho = −0.019, p = 0.65 for the Shannon index and rho = −0.0097, p = 0.82 for the forChao1 the Chao1 index. (C) Correlation between alpha‐diversity (Shannon and Chao1 indices) and HBI index. (C) Correlation between alpha-diversity (Shannon and Chao1 indices) and HBI (Spearman (Spearman correlation test). rho = −021, p−=5 3 x 10‐5 for the Shannon index and rho = −0.26, p =−27 x correlation 10‐7 test). rho for the Chao1 = −021, p = 3 × 10 for the Shannon index and rho = −0.26, p = 2 × 10 for index. the Chao1 index. Remarkably, we observed that, across countries, HCs from the USA were associated Remarkably, we observed that, across countries, HCs from the USA were asso- with a lower diversity compared with those from Spain and China (Figure 4). This find‐ ciated with a lower diversity compared with those from Spain and China (Figure 4). ing was not explained by BMI, age or any other available covariates. It could be explained This finding was not explained by BMI, age or any other available covariates. It could by the lack of a mechanical lysis step during the DNA extraction procedure be explained by the lack of a mechanical lysis step during the DNA extraction proce- (https://ibdmdb.org/cb/document/Sample%20Handling%20Protocols/DNA_and_RNA_S dure (https://ibdmdb.org/cb/document/Sample%20Handling%20Protocols/DNA_and_ tool_Isolation_HMP2_protocol.pdf, accessed on 16 September 2022) or by environmental RNA_Stool_Isolation_HMP2_protocol.pdf, accessed on 14 September 2022) or by environ- factors such as diet, but this latter information was not available in the published mental factors such as diet, but this latter information was not available in the published metadata. Unexpectedly, diversity was not significantly different in CD between the metadata. Unexpectedly, diversity was not significantly different in CD between the Span- ish and USA cohort, whereas the Chinese cohort presented a lower diversity than the other two countries. Again, this difference could be related to missing covariates such as genetic as well as disease severity or diet in the Chinese metadata. Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 6 of 16 Spanish and USA cohort, whereas the Chinese cohort presented a lower diversity than Int. J. Mol. Sci. 2022, 23, 10868 6 of 15 the other two countries. Again, this difference could be related to missing covariates such as genetic as well as disease severity or diet in the Chinese metadata. Figure 4. Figure Alpha-diversity based 4. Alpha‐diversity based on on health health status status across acrosscountries. countries.Shannon Shannonand andChao1 Chao1indices indicesininHC (A), CD (B), and UC (C) (Mann-Whitney’s U test). HC (A), CD (B), and UC (C) (Mann‐Whitney’s U test). Beta-diversityanalysis Beta‐diversity analysisusing usingweighted weightedandandunweighted unweightedUniFrac UniFrac distances distances showed showed that, at the global microbial composition level, CD patients formed a that, at the global microbial composition level, CD patients formed a separate clusterseparate cluster from that of HC and UC. Using the PERMANOVA test, all comparisons from that of HC and UC. Using the PERMANOVA test, all comparisons were significant were significant (pval << 0.005): (pval 0.005):HC HCvs. vs.CDCDforforboth bothweighted weightedand andunweighted unweightedUniFrac; UniFrac; HCHCvs.vs. UCUC forfor unweighted and weighted UniFrac, CD vs. UC for both weighted unweighted and weighted UniFrac, CD vs. UC for both weighted and unweighted and unweighted UniFrac (Supplementary UniFrac Figure Figure (Supplementary S1), which is in agreement S1), which with with is in agreement previous findings previous [3,6].[3,6]. findings Beta- diversity was more Beta‐diversity affected was more by CDbylocalization affected than clinical CD localization activityactivity than clinical (Figure(Figure 1), however, 1), however, this observation could be explained by the fact that the majority of CD patients in this observation could be explained by the fact that the majority of CD patients were clinical were remission. in clinical remission. To evaluate the degree of dysbiosis in CD and UC, we calculated the dysbiosis score based on weighted and unweighted UniFrac dissimilarities to HC, as described by Lloyd- Price et al. [14]. Based on unweighted UniFrac dysbiosis scores, as expected, CD and UC were both significantly different from HC in Spain and the USA, and CD from HC in the Chinese cohort (Figure 5). However, based on the weighted UniFrac dysbiosis Int. J. Mol. Sci. 2022, 23, 10868 7 of 15 scores, the differences remained significant only in the Spanish and Chinese cohorts. This latter observation could be explained by a lower alpha-diversity of HC in the USA cohort compared with the other two cohorts. Moreover, CD was different from UC in Spain based on both weighted and unweighted UniFrac dysbiosis scores, however, in the USA cohort, CD differed from UC only based on the unweighted UniFrac dysbiosis scores. 8This Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW of 16latter observation indicated that alterations of the microbiome community in UC were more related to low-abundant taxa. Figure 5. Dysbiosis Figure scores 5. Dysbiosis andand scores correlation withwith correlation disease severity. disease (A) Dysbiosis severity. scores (A) Dysbiosis across scores cohorts across cohorts perper country (Wilcoxon test). (B) Correlation between dysbiosis score and disease activity country (Wilcoxon test). (B) Correlation between dysbiosis score and disease activity index index (Spearman (Spearmancorrelation test). correlation rhorho test). = 0.031, p = 0.47 = 0.031, for unweighted p = 0.47 for unweightedUniFrac and rho UniFrac and =rho 0.0056, p = 0.9 = 0.0056, p=for0.9 for weighted UniFrac of CD (HBI, Harvey Bradshaw index); rho = 0.26, p = 1.8 × 10−7 for−7unweighted weighted UniFrac of CD (HBI, Harvey Bradshaw index); rho = 0.26, p = 1.8 × 10 for unweighted UniFrac and rho = 0.12, p = 0.021 for weighted UniFrac of UC (CAI, colitis activity index). UniFrac and rho = 0.12, p = 0.021 for weighted UniFrac of UC (CAI, colitis activity index). 2.4. Taxonomic Profile Associated with Disease Phenotype and Severity The taxa were classified at the species level and filtered by achieving at least 0.1% of total abundance in at least 10% of the samples in order to remove outlier taxa with very low abundance and prevalence. Mixed‐effects linear models, as implemented by and CD localization were accounted as random effects. Samples from the Chinese cohort were excluded from these analyses due to the missing CD localization data. Results showed that the most significantly depleted species in CD compared with controls were Ruminococcus bromii (adj. pval = 0.03) and Ruminococcus bicirculans (adj. pval = 0.042), while Escherichia coli (adj. pval = 0.03) was the most enriched species in CD (Supplemen‐ Int. J. Mol. Sci. 2022, 23, 10868 8 of 15 tary Table S2, Figure 6, Supplementary Figure S2). However, UC (n = 384) compared to HC (n = 409), with age, gender, and BMI as fixed effects, while subject ID and country were handled as random effects, did not lead to significant differences at the species lev‐ el. To compare the degree of dysbiosis related to disease phenotype across countries, we calculated the bacterial To reveal ratio of median species CD dysbiosis associated withscores over relapse or that of HC.status remission This analysis, in CD andusing UC, unweighted UniFrac dysbiosis scores, revealed a higher dysbiosis level we split the CD and UC cohorts into two groups. The remission state consisted of 444 CD in the Chinese (ratio of 1.84) samples with followed by the HBI < 4 and 295Spanish (ratio with UC samples of 1.79) CAIand< 4the andUSAthe (ratio of state relapse 1.17) encom‐ cohorts (Figure 5A). This ratio, using weighted UniFrac dysbiosis scores, followed passed 119 CD samples with HBI ≥ 4 and 89 UC samples with CAI ≥ 4. In neither the the same trend CD (ratio of 1.329 for China; ratio of 1.315 for Spain; ratio of 0.957 for the USA). nor UC group did disease activity status associate with specific bacterial species usingThe latter result indicated a lack of difference between CD and HC in the USA cohort (pval > 0.05). MaAsLin2. 2.4. Taxonomic Profile Associated with Disease Phenotype and Severity 2.5. Taxonomic Differences across Countries and Health Status. The taxa were classified at the species level and filtered by achieving at least 0.1% of Country‐wise analysis between Spain and the USA showed that four species were total abundance in at least 10% of the samples in order to remove outlier taxa with very low differentially abundant in HC samples (n = 67 Spain, n = 342 USA) (Supplementary Table abundance and prevalence. Mixed-effects linear models, as implemented by MaAsLin2, S3, MaAsLin2 results for HC/country group), eight species in UC (n = 65 Spain, n = 319 were used to determine differences between CD, UC, and HC taxonomic profiles, as well USA) (Supplementary Table S4), and no species achieved significance in CD (n = 57 as between countries with the same disease phenotype. Spain, n = 506 USA) (using age, gender, and BMI as fixed effects while subject ID, HBI or We first compared the taxonomic profile of CD (n = 563) with that of HC (n = 409), CAI, and country as random effects using MaAsLin2). combining the Spanish and USA cohorts, taking into account covariates, such as gender, age The relative and body abundance mass index of Fusicatenibacter (BMI) included saccharivorans as fixed effects, was significantly while subject different ID, HBI, country, and in the HC group between the Spanish and USA cohorts, but did not achieve CD localization were accounted as random effects. Samples from the Chinese cohort were significance in the UCfrom excluded group, while these Agathobaculum analyses due to thebutyriproducens, Coprococcus missing CD localization data.catus, Blautia Results wexlerae, showed that Anaerostipes hadrus, and Eubacterium halii were significantly enriched in the most significantly depleted species in CD compared with controls were Ruminococcus the Spanish UC group, but not bromii (adj. pvalHC, whereas = 0.03) Coprococcus comes, and Ruminococcus Dorea bicirculans formicigenerans (adj. pval = 0.042),and Dorea while longicatena Escherichia coli were significantly enriched in the (adj. pval = 0.03) was the most enriched species in CD (Supplementary Table S2,6).Figure 6, both the Spanish HC and UC groups (Figure Interestingly, Supplementary we did Figure S2).not find anyUC However, species associated (n = 384) comparedwithtodisease HC (n severity for both = 409), with age, CD (when using HBI as a possible explanatory variable) and UC (when gender, and BMI as fixed effects, while subject ID and country were handled as random using CAI as a possible explanatory effects, did not lead tovariable) for differences significant both the Spanish and thelevel. at the species USA cohorts. Figure 6. Common differentially abundant (DA) species between Spanish and USA cohorts in the HC and UC groups. * DA in both HC and UC. ** DA only in HC. *** DA only in UC. To reveal bacterial species associated with relapse or remission status in CD and UC, we split the CD and UC cohorts into two groups. The remission state consisted of 444 CD samples with HBI < 4 and 295 UC samples with CAI < 4 and the relapse state encompassed 119 CD samples with HBI ≥ 4 and 89 UC samples with CAI ≥ 4. In neither the CD nor UC group did disease activity status associate with specific bacterial species using MaAsLin2. Int. J. Mol. Sci. 2022, 23, 10868 9 of 15 2.5. Taxonomic Differences across Countries and Health Status Country-wise analysis between Spain and the USA showed that four species were differentially abundant in HC samples (n = 67 Spain, n = 342 USA) (Supplementary Table S3, MaAsLin2 results for HC/country group), eight species in UC (n = 65 Spain, n = 319 USA) (Supplementary Table S4), and no species achieved significance in CD (n = 57 Spain, n = 506 USA) (using age, gender, and BMI as fixed effects while subject ID, HBI or CAI, and country as random effects using MaAsLin2). The relative abundance of Fusicatenibacter saccharivorans was significantly different in the HC group between the Spanish and USA cohorts, but did not achieve significance in the UC group, while Agathobaculum butyriproducens, Coprococcus catus, Blautia wexlerae, Anaerostipes hadrus, and Eubacterium halii were significantly enriched in the Spanish UC group, but not HC, whereas Coprococcus comes, Dorea formicigenerans and Dorea longicatena were significantly enriched in the both the Spanish HC and UC groups (Figure 6). Interestingly, we did not find any species associated with disease severity for both CD (when using HBI as a possible explanatory variable) and UC (when using CAI as a possible explanatory variable) for both the Spanish and the USA cohorts. 3. Discussion The relationship between microbiota and IBD pathogenesis is well established, but there is still limited information regarding the influence of geography and its related environmental impact on IBD microbiota composition. To our knowledge, this is the first study that compares gut microbiome variances in IBD patients across three continents using DNA shotgun metagenomics. Geographical differences observed may reflect differences attributed to host genetic, dietary habits, and lifestyle (living conditions, sanitation measure, urbanization level, stress level, antibiotics used in animal agriculture, etc.) [28]. The lower alpha-diversity described in both the healthy and UC USA cohorts could be attributed to diet. Indeed, according to the United States Department of Agriculture (USDA), consumers in the EU and the United States are very different, with less consumption of fruits and vegetables in the latter population, which could lead to a lower alpha-diversity (https://www.ers.usda.gov/ webdocs/outlooks/40408/30646_wrs0404f_002.pdf, accessed on 14 September 2022). Analyzing the influence of the metadata covariates on the microbiome, we observed that geographic location was a major determinant of microbiome variation in both IBD and HC, agreeing with previous studies [22,29,30]. In a large multicenter longitudinal study on early life, Stewart et al. found that environmental factors, including geographical location and household exposures, represented important covariates of the microbiome structure. More recently, in an intercontinental microbiome study, Clooney et al. [22] showed that the presence or absence of a CD diagnosis followed by geographic location had the greatest impact on the microbiome. Moreover, He et al. [29] characterized the gut microbiome of 7009 healthy individuals from 14 districts within one Chinese province and found that, among various covariates, host location presented the strongest associations with microbiome variations. In addition to geography, in this study, we have demonstrated that disease, age, BMI and gender also have a significant impact on microbiome composition, although to a lesser extent. With respect to dysbiosis, we have shown that disturbances in fecal microbiome composition were most pronounced in CD than in UC, which is in line with previous reports. Scanlan et al., described that the temporal stability of dominant species in CD was significantly lower than in HC [31]. Morgan et al., found that IBD population and ileal CD, in particular, were associated with a dysbiosis characterized by changes in Firmicutes and Proteobacteria phyla [4]. Lloyd-Price et al., followed 132 subjects for one year to generate integrated longitudinal molecular profiles of host and microbial activity during disease [14]. Samples from participants with CD or UC were overrepresented in the dysbiotic set, with 24.3% and 11.6% of their samples classified as dysbiotic, respectively. In addition, our Int. J. Mol. Sci. 2022, 23, 10868 10 of 15 dysbiosis findings in CD were observed regardless of geographic location and disease severity, which has not been reported elsewhere. In the Spanish cohort, we confirmed our previous findings [24] that CD was associ- ated with a significant reduction in diversity, based on both Shannon and Chao1 indices, compared with HC and UC, while UC did not have a lower diversity compared to HC, suggesting that the dysbiosis score might be higher in CD than in UC. In the USA cohort, based on the Shannon index, we observed differences in alpha-diversity only between UC and HC and between CD and HC, but not between CD and UC. These findings could be explained by the greater disease severity in the USA cohort than in the Spanish one. In HC, the USA cohort was associated with lower diversity compared with the other two cohorts. These findings were not explained by the available published metadata, suggesting that other unavailable factors, such as diet or ethnicity, could play a role in this observation. Based on disease phenotype and combining the Spanish and USA cohorts, Ruminococ- cus bromii and Ruminococcus bicirculans were the significantly most depleted microbial species in CD compared with HC, while Escherichia coli was the most enriched species. Only the latter is in agreement with other reports [32]. Interestingly, Fang et al. [33] who performed a strain level analysis of E. coli in CD, found that the strain identified by the metagenomic approach was similar to known pathogenic Adherent-Invasive E. coli (AEIC) strains. Furthermore, Ruminococcus bromii has been shown to have the ability to degrade dietary-resistant starches while Ruminococcus bicirculans has the capacity to utilize plant glucans. These two bacterial species, missing in CD, could be key players in the metabolism of a plant-based diet and maintenance of gut homeostasis. In UC, no significant differ- ences were observed at the species level. The fact that no species achieved significance in CD in our country-wise analysis suggested that the development of CD might lead to a convergence of this composition’s profile, and this finding could open the possibility of identifying a “core” CD microbiome as a universal biomarker. Neither in the CD nor the UC group did disease activity status associate with specific bacteria. This finding could be explained by the fact that other variables such as BMI and age had a greater impact on the microbiome variation than the disease severity scores. In this sense, we found reports with divergent results that suggested that clinical activity had an impact. In a large pediatric CD cohort, Gevers et al. [12] described an axis defined by an increased abundance of bacteria including Enterobacteriaceae, Pasteurellacaea, Veillonel- laceae, and Fusobacteriaceae, and decreased abundance of Erysipelotrichales, Bacteroidales, and Clostridiales that correlated strongly with disease status. Recently, Clooney et al., in a longitudinal study, demonstrated that the greatest changes in microbiota composition were linked to transitions across active and inactive phases of the disease, even though specific taxa were not associated with different activity states [22]. This was only possible when using intra-individual ratios of each taxon between two consecutive time points, but not when using taxa from single time points. Similarly, Lloyd-Price et al., showed that periods of disease activity were also marked by increased temporal variability, with characteristic taxonomic, functional, and biochemical shifts [14]. Globally, Coprococcus comes, Dorea formicigenerans, and Dorea longicatena were shared among countries but enriched in the Spanish cohort. These species were previously found to be linked to overweight/obesity in some reports [34,35] and to correlate with metabolic markers in others [36]. The use of shotgun metagenomic can be viewed as an advantage over 16S rRNA sequencing. Since 16S rRNA sequencing is a cost-effective method for taxonomic profiling, it has been widely used to investigate the association between gut microbiota and IBD [37]. However, this approach presents several pitfalls, as it is biased by the lack of truly universal primers for the PCR amplification and the variability of the rRNA operon copy number throughout the bacterial kingdom. Moreover, due to the short size of the 16S rRNA gene (about 300 bp usually sequenced), this approach leads to a taxonomic resolution mainly up to the genus level. Instead, the shotgun metagenomic approach, which consists of a random DNA sequencing from the complete content of a clinical sample, relies, in this Int. J. Mol. Sci. 2022, 23, 10868 11 of 15 study, on single copy marker gene databases for taxonomic profiling. This method avoids all the above-mentioned limitations and allows classification of known and unknown microorganisms at the species level. An important limitation of our study is that it analyzed fecal and not mucosal micro- biome. It is well established that feces are just a proxy for the intestinal mucosa, the actual site where interactions between bacteria and host immunity occur. However, applying shot- gun metagenomics on this type of sample is still challenging as it is considered low biomass sample and contains more human cells than microbial cells. Therefore, the sequencing depth should be much higher for mucosal than for fecal samples. Our study points out other limitations when attempting to perform intercontinental research on IBD-specific biomarkers. Among these limitations, we reported the lack of standardized clinical data and the unavailability of important covariates, including dietary data, which has been previously suggested as a potential risk factor for IBD [38]. In addition, the metadata lacked information regarding other valuable covariates such as ethnicity, disease behavior, location in UC, calprotectin, and endoscopic activity. Altogether, missing data decreases the margin of analysis of confounders responsible for microbiome variations. Indeed, a non-negligible number of patients had to be discarded from the analysis due to the lack of a standard clinical activity index throughout the cohorts. For instance, the Chinese cohort could not be used in the analysis of the association between taxonomic profile and patients’ characteristics as it lacked HBI and disease location using the Montreal classification sys- tems, as reported in the other studies. Moreover, we had to exclude an entire Belgium cohort as it did not use the same index for disease activity (Rutgeerts score instead of HBI). Another important point to consider was the lack of standardized DNA extraction methods, including the reporting of the use of a mechanical or chemical procedure across countries, which may limit interpretations concerning the observed geographical impact. Indeed, this extraction step in microbiome studies has been shown to be a possible contributor to the microbiome variation [39]. Therefore, micro bio-biomarkers identified based on the USA cohort, which did not perform additional mechanical lysis, should be taken with caution. Finally, sequence analysis methods should also be homogenized in particular in terms of taxonomic and functional database versions, as the interpretation of taxonomic and functional profiling may depend on the mapping rate. In conclusion, geographic location, disease activity status, and other environmental factors appear to influence microbiota changes in IBD. In the case of CD, we are inclined to believe that these factors have a lower impact on microbiome composition, even though dysbiosis was greater in CD than in UC. We strongly recommend taking into consideration all possible environmental and host factors, as well as the implementation of standardized extraction methods, when conducting microbiota studies in order to obtain globally valid and reproducible findings. Finally, as IBD is a multifactorial disorder, to acquire better in- sights into the ethology of IBD, future ideal study design should integrate mucosal shotgun metaomic techniques, host genetics and clinical phenotypes, lifestyle, and dietary habits. 4. Materials and Methods 4.1. Collection of Metagenomic Data Shotgun metagenomic sequencing data generated from 329 public human fecal sam- ples were downloaded from NCBI SRA [40] (https://www.ncbi.nlm.nih.gov/sra/, ac- cessed on 14 September 2022). The 329 public human metagenomic data were derived from two unique bioprojects, one of which was published in our previous studies (PR- JEB1220, 212 samples, Spanish cohort), and the other was from the PRJEB15371 bioproject (117 samples, Chinese cohort). The taxonomic and functional profiling of HMP (NIH Human Microbiome Project, https://doi.org/10.1038/s41586-019-1238-8, accessed on 14 September 2022), which includes 1,639 samples, were downloaded from the IBDMDB (The Inflammatory Bowel Disease Multi-Omics Database, https://ibdmdb.org/, accessed on 14 September 2022) [14]. The metadata for all the human metagenomic data can be found in Supplementary Table S1 and contains available information such as country, disease Int. J. Mol. Sci. 2022, 23, 10868 12 of 15 localization, gender, age, BMI, dysbiosis index, and severity scores. The taxonomic profiles from the HMP data were obtained using the MetaPhlAn 3.0 tool and provided relative abundances [41]. 4.2. Filtering of the Datasets All metagenomic datasets were filtered based on metadata information to perform a homogenized meta-analysis. We removed samples from (1) HC without body-mass indexes (BMI), age, or gender information; (2) from CD without disease localization defined by the Montreal classification method [42], Harvey-Bradshaw index (HBI) [43] (Harvey and Bradshaw 1980), gender, age or BMI information; and (3) from ulcerative colitis (UC) without Colitis Activity Index (CAI) [44], gender, age or BMI information. The taxonomic profiling tables were further filtered by criteria of species with less than 0.1% abundance [2] and samples with zero species abundance were removed. The Chinese cohort was included in the alpha-diversity, beta-diversity, and dysbiosis analyses, but was excluded from the analysis of multivariate association between covariates and microbiome features using MaAsLin2 because information on disease localization was not based on the Montreal classification method, in contrast to the other two cohorts, and HBI information was not available (Supplementary Table S5). 4.3. Upstream Sequence Analysis: Quality Control, Decontamination and Profiling For the Spanish and Chinese cohorts, the KneadData v0.7.4 pipeline (https://huttenhower. sph.harvard.edu/kneaddata, accessed on 14 September 2022) was used to trim the reads with lengths below 50% of the total input read length and to remove reads that mapped to the human genome. Taxonomic and functional profiling was performed by the HumanN3 pipeline [41] against the built-in database of Metaphlan3 and the UniRef90 database with default parameters. 4.4. Statistical Analyses Statistical analyses were performed using R (version 4.0.3). Alpha-diversity was assessed using Shannon [27] and Chao1 metrics [26] indices, while beta-diversity was estimated by calculating the weighted and unweighted UniFrac distances between sam- ples [45]. The dysbiosis score was calculated as previously defined [14,24], using weighted and unweighted UniFrac distances. Healthy controls from each country were set as the reference group for each cohort. Alpha-diversity and dysbiosis scores between groups were compared using the Mann-Whitney U test. The contribution of clinical, anthropometric, and demographic factors to the effect size was estimated using permutational analysis of variance (PERMANOVA), as implemented in the adonis2 function of the vegan R package (https://cran.r-project.org/web/packages/vegan/index.html, accessed on 14 September 2022) using weighted and unweighted UniFrac distance metrics. UniFrac distances were calculated as implemented in the phyloseq R package [46]. Differential abundance analysis of the taxonomic profiles was performed using the R version of the MaAsLin2 tool [47], which implements linear mixed-effects models that are useful for multivariable association discovery in population-scale microbiome studies. Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ijms231810868/s1. Author Contributions: Conceptualization, C.M.; methodology, L.M., G.S.-G. and Z.X.; software, G.S.-G. and Z.X.; validation G.S.-G. and Z.X.; formal analysis, L.M., G.S.-G. and Z.X.; investigation, C.M.; resources, L.M., G.S.-G. and Z.X.; data curation, G.S.-G. and Z.X.; writing—original draft preparation, C.M., L.M. and G.S.-G.; writing—review and editing, C.M., N.B., L.M., G.S.-G. and Z.X.; supervision, C.M.; project administration, C.M.; funding acquisition, C.M. All authors have read and agreed to the published version of the manuscript. Int. J. Mol. Sci. 2022, 23, 10868 13 of 15 Funding: Z.X. received a fellowship from the European Union’s Horizon 2020 research and innova- tion program under the Marie Sklodowska-Curie Action, Innovative Training Network: FunHoMic; grant number 812969. Chaysavanh Manichanh received grants from the Instituto de Salud Carlos III/FEDER (PI17/00614; PI20/00130). Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of the Hospital Vall d’Hebron (protocol code PR(AG)156/2017, approved the 8 May 2017). Informed Consent Statement: Written informed consent has been obtained from the patients to publish this paper. Data Availability Statement: All information regarding data used in this study can be found under “Collection of Metagenomic Data” in the Method section. Conflicts of Interest: No conflict of interest to declare. References 1. Bernstein, C.N.; Forbes, J.D. Gut Microbiome in Inflammatory Bowel Disease and Other Chronic Immune-Mediated Inflammatory Diseases. Inflamm. Intest. Dis. 2017, 2, 116–123. [CrossRef] [PubMed] 2. Franzosa, E.A.; Sirota-Madi, A.; Avila-Pacheco, J.; Fornelos, N.; Haiser, H.J.; Reinker, S.; Vatanen, T.; Hall, A.B.; Mallick, H.; McIver, L.J.; et al. Gut microbiome structure and metabolic activity in inflammatory bowel disease. Nat. Microbiol. 2019, 4, 293–305. [CrossRef] [PubMed] 3. Halfvarson, J.; Brislawn, C.J.; Lamendella, R.; Vázquez-Baeza, Y.; Walters, W.A.; Bramer, L.M.; D’Amato, M.; Bonfiglio, F.; McDonald, D.; Gonzalez, A.; et al. Dynamics of the human gut microbiome in inflammatory bowel disease. Nat. Microbiol. 2017, 2, 17004. [CrossRef] [PubMed] 4. Morgan, X.C.; Tickle, T.L.; Sokol, H.; Gevers, D.; Devaney, K.L.; Ward, D.V.; Reyes, J.A.; Shah, S.A.; LeLeiko, N.; Snapper, S.B.; et al. Dysfunction of the intestinal microbiome in inflammatory bowel disease and treatment. Genome Biol. 2012, 13, R79. [CrossRef] [PubMed] 5. Nishida, A.; Inoue, R.; Inatomi, O.; Bamba, S.; Naito, Y.; Andoh, A. Gut microbiota in the pathogenesis of inflammatory bowel disease. Clin. J. Gastroenterol. 2018, 11, 1–10. [CrossRef] [PubMed] 6. Pascal, V.; Pozuelo, M.; Borruel, N.; Casellas, F.; Campos, D.; Santiago, A.; Martinez, X.; Varela, E.; Sarrabayrouse, G.; Machiels, K.; et al. A microbial signature for Crohn’s disease. Gut 2017, 66, 813–822. [CrossRef] 7. Seksik, P.; Rigottier-Gois, L.; Gramet, G.; Sutren, M.; Pochart, P.; Marteau, P.; Jian, R.; Doré, J. Alterations of the dominant faecal bacterial groups in patients with Crohn’s disease of the colon. Gut 2003, 52, 237–242. [CrossRef] 8. Sheehan, D.; Shanahan, F. The Gut Microbiota in Inflammatory Bowel Disease. Gastroenterol. Clin. North Am. 2017, 46, 143–154. [CrossRef] 9. Zuo, T.; Kamm, M.A.; Colombel, J.-F.; Ng, S.C. Urbanization and the gut microbiota in health and inflammatory bowel disease. Nat. Rev. Gastroenterol. Hepatol. 2018, 15, 440–452. [CrossRef] 10. Copeland, J.K.; Chao, G.; Vanderhout, S.; Acton, E.; Wang, P.W.; Benchimol, E.I.; El-Sohemy, A.; Croitoru, K.; Gommerman, J.L.; Guttman, D.S.; et al. The Impact of Migration on the Gut Metagenome of South Asian Canadians. Gut Microbes 2021, 13, 1–29. [CrossRef] 11. Vangay, P.; Johnson, A.J.; Ward, T.L.; Al-Ghalith, G.A.; Shields-Cutler, R.R.; Hillmann, B.M.; Lucas, S.K.; Beura, L.K.; Thompson, E.A.; Till, L.M.; et al. US Immigration Westernizes the Human Gut Microbiome. Cell 2018, 175, 962–972.e10. [CrossRef] [PubMed] 12. Gevers, D.; Kugathasan, S.; Denson, L.A.; Vázquez-Baeza, Y.; Van Treuren, W.; Ren, B.; Schwager, E.; Knights, D.; Song, S.J.; Yassour, M.; et al. The Treatment-Naive Microbiome in New-Onset Crohn’s Disease. Cell Host Microbe 2014, 15, 382–392. [CrossRef] [PubMed] 13. Lepage, P.; Häsler, R.; Spehlmann, M.E.; Rehman, A.; Zvirbliene, A.; Begun, A.; Ott, S.; Kupcinskas, L.; Doré, J.; Raedler, A.; et al. Twin Study Indicates Loss of Interaction Between Microbiota and Mucosa of Patients with Ulcerative Colitis. Gastroenterology 2011, 141, 227–236. [CrossRef] 14. Lloyd-Price, J.; Arze, C.; Ananthakrishnan, A.N.; Schirmer, M.; Avila-Pacheco, J.; Poon, T.W.; Andrews, E.; Ajami, N.J.; Bonham, K.S.; Brislawn, C.J.; et al. Multi-omics of the gut microbial ecosystem in inflammatory bowel diseases. Nature 2019, 569, 655–662. [CrossRef] 15. Metwaly, A.; Reitmeier, S.; Haller, D. Microbiome risk profiles as biomarkers for inflammatory and metabolic disorders. Nat. Rev. Gastroenterol. Hepatol. 2022, 19, 383–397. [CrossRef] 16. Sokol, H.; Seksik, P.; Furet, J.P.; Firmesse, O.; Nion-Larmurier, I.; Beaugerie, L.; Cosnes, J.; Corthier, G.; Marteau, P.; Doré, J. Low counts of Faecalibacterium prausnitzii in colitis microbiota. Inflamm. Bowel Dis. 2009, 15, 1183–1189. [CrossRef] 17. Yilmaz, B.; Juillerat, P.; Øyås, O.; Ramon, C.; Bravo, F.D.; Franc, Y.; Fournier, N.; Michetti, P.; Mueller, C.; Geuking, M.; et al. Microbial network disturbances in relapsing refractory Crohn’s disease. Nat. Med. 2019, 25, 323–336. [CrossRef] [PubMed] 18. Li, Q.; Chang, Y.; Zhang, K.; Chen, H.; Tao, S.; Zhang, Z. Implication of the gut microbiome composition of type 2 diabetic patients from northern China. Sci. Rep. 2020, 10, 5450. [CrossRef] Int. J. Mol. Sci. 2022, 23, 10868 14 of 15 19. Reitmeier, S.; Kiessling, S.; Clavel, T.; List, M.; Almeida, E.L.; Ghosh, T.S.; Neuhaus, K.; Grallert, H.; Linseisen, J.; Skurk, T.; et al. Arrhythmic Gut Microbiome Signatures Predict Risk of Type 2 Diabetes. Cell Host Microbe 2020, 28, 258–272.e6. [CrossRef] 20. Langhorst, J.; Schöls, M.; Cinar, Z.; Eilert, R.; Kofink, K.; Paul, A.; Zempel, C.; Elsenbruch, S.; Lauche, R.; Ahmed, M.; et al. Comprehensive Lifestyle-Modification in Patients with Ulcerative Colitis–A Randomized Controlled Trial. J. Clin. Med. 2020, 9, 3087. [CrossRef] 21. Relling, I.; Akcay, G.; Fangmann, D.; Knappe, C.; Schulte, D.M.; Hartmann, K.; Müller, N.; Türk, K.; Dempfle, A.; Franke, A.; et al. Role of wnt5a in Metabolic Inflammation in Humans. J. Clin. Endocrinol. Metab. 2018, 103, 4253–4264. [CrossRef] [PubMed] 22. Clooney, A.G.; Eckenberger, J.; Laserna-Mendieta, E.; Sexton, K.A.; Bernstein, M.T.; Vagianos, K.; Sargent, M.; Ryan, F.; Moran, C.; Sheehan, D.; et al. Ranking microbiome variance in inflammatory bowel disease: A large longitudinal intercontinental study. Gut 2021, 70, 499–510. [CrossRef] 23. Nielsen, H.B.; Almeida, M.; Juncker, A.S.; Rasmussen, S.; Li, J.; Sunagawa, S.; Plichta, D.R.; Gautier, L.; Pedersen, A.G.; Le Chatelier, E.; et al. Identification and assembly of genomes and genetic elements in complex metagenomic samples without using reference genomes. Nat. Biotechnol. 2014, 32, 822–828. [CrossRef] [PubMed] 24. Serrano-Gómez, G.; Mayorga, L.; Oyarzun, I.; Roca, J.; Borruel, N.; Casellas, F.; Varela, E.; Pozuelo, M.; Machiels, K.; Guarner, F.; et al. Dysbiosis and relapse-related microbiome in inflammatory bowel disease: A shotgun metagenomic approach. Comput. Struct. Biotechnol. J. 2021, 19, 6481–6489. [CrossRef] [PubMed] 25. He, Q.; Gao, Y.; Jie, Z.; Yu, X.; Laursen, J.M.; Xiao, L.; Li, Y.; Li, L.; Zhang, F.; Feng, Q.; et al. Two distinct metacommunities characterize the gut microbiota in Crohn’s disease patients. Gigascience 2017, 6, 1–11. [CrossRef] 26. Chao, A. Nonparametric Estimation of the Number of Classes in a Population. Scand. J. Stat. 1984, 11, 265–270. 27. Shannon, C.E. A Mathematical Theory of Communication; Wiley Online Library: Hoboken, NJ, USA, 1948. 28. Martinez, I.; Stegen, J.C.; Maldonado-Gomez, M.X.; Eren, A.M.; Siba, P.M.; Greenhill, A.R.; Walter, J. The Gut Microbiota of Rural Papua New Guineans: Composition, Diversity Patterns, and Ecological Processes. Cell Rep. 2015, 11, 527–538. [CrossRef] 29. He, Y.; Wu, W.; Zheng, H.M.; Li, P.; McDonald, D.; Sheng, H.F.; Chen, M.X.; Chen, Z.H.; Ji, J.Y.; Zheng, Z.D.X.; et al. Regional variation limits applications of healthy gut microbiome reference ranges and disease models. Nat. Med. 2018, 24, 1532–1535. [CrossRef] 30. Stewart, C.J.; Ajami, N.J.; O’Brien, J.L.; Hutchinson, D.S.; Smith, D.P.; Wong, M.C.; Ross, M.C.; Lloyd, R.E.; Doddapaneni, H.; Metcalf, G.A.; et al. Temporal development of the gut microbiome in early childhood from the TEDDY study. Nature 2018, 562, 583–588. [CrossRef] 31. Scanlan, P.D.; Shanahan, F.; O’Mahony, C.; Marchesi, J.R. Culture-independent analyses of temporal variation of the dominant fecal microbiota and targeted bacterial subgroups in Crohn’s disease. J. Clin. Microbiol. 2006, 44, 3980–3988. [CrossRef] 32. Palmela, C.; Chevarin, C.; Xu, Z.; Torres, J.; Sevrin, G.; Hirten, R.; Barnich, N.; Ng, S.C.; Colombel, J.-F. Adherent-invasive Escherichia coli in inflammatory bowel disease. Gut 2018, 67, 574–587. [CrossRef] [PubMed] 33. Fang, X.; Monk, J.M.; Nurk, S.; Akseshina, M.; Zhu, Q.; Gemmell, C.; Gianetto-Hill, C.; Leung, N.; Szubin, R.; Sanders, J.; et al. Metagenomics-Based, Strain-Level Analysis of Escherichia coli from a Time-Series of Microbiome Samples from a Crohn’s Disease Patient. Front Microbiol. 2018, 9, 2559. [CrossRef] [PubMed] 34. Companys, J.; Gosalbes, M.; Pla-Pagà, L.; Calderón-Pérez, L.; Llauradó, E.; Pedret, A.; Valls, R.; Jiménez-Hernández, N.; Sandoval-Ramirez, B.; del Bas, J.; et al. Gut Microbiota Profile and Its Association with Clinical Variables and Dietary Intake in Overweight/Obese and Lean Subjects: A Cross-Sectional Study. Nutrients 2021, 13, 2032. [CrossRef] [PubMed] 35. Gomez-Arango, L.F.; Barrett, H.L.; McIntyre, H.D.; Callaway, L.K.; Morrison, M.; Nitert, M.D.; SPRING Trial Group. Connections Between the Gut Microbiome and Metabolic Hormones in Early Pregnancy in Overweight and Obese Women. Diabetes 2016, 65, 2214–2223. [CrossRef] [PubMed] 36. Brahe, L.K.; Le Chatelier, E.; Prifti, E.; Pons, N.; Kennedy, S.B.; Hansen, T.; Pedersen, O.; Astrup, A.; Ehrlich, S.; Larsen, L.H. Specific gut microbiota features and metabolic markers in postmenopausal women with obesity. Nutr. Diabetes 2015, 5, e159. [CrossRef] 37. Valles-Colomer, M.; Darzi, Y.; Vieira-Silva, S.; Falony, G.; Raes, J.; Joossens, M. Meta-omics in Inflammatory Bowel Disease Research: Applications, Challenges, and Guidelines. J. Crohn’s Colitis 2016, 10, 735–746. [CrossRef] 38. Lewis, J.D.; Abreu, M.T. Diet as a Trigger or Therapy for Inflammatory Bowel Diseases. Gastroenterology 2017, 152, 398–414.e6. [CrossRef] 39. Costea, P.I.; Zeller, G.; Sunagawa, S.; Pelletier, E.; Alberti, A.; Levenez, F.; Tramontano, M.; Driessen, M.; Hercog, R.; Jung, F.-E.; et al. Towards standards for human fecal sample processing in metagenomic studies. Nat. Biotechnol. 2017, 35, 1069–1076. [CrossRef] 40. Leinonen, R.; Sugawara, H.; Shumway, M. International Nucleotide Sequence Database C. The sequence read archive. Nucleic Acids Res. 2011, 39, D19–D21. [CrossRef] 41. Beghini, F.; McIver, L.J.; Blanco-Míguez, A.; Dubois, L.; Asnicar, F.; Maharjan, S.; Mailyan, A.; Manghi, P.; Scholz, M.; Thomas, A.M.; et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. eLife 2021, 10, e65088. [CrossRef] 42. Silverberg, M.S.; Satsangi, J.; Ahmad, T.; Arnott, I.D.R.; Bernstein, C.N.; Brant, S.R.; Caprilli, R.; Colombel, J.-F.; Gasche, C.; Geboes, K.; et al. Toward an Integrated Clinical, Molecular and Serological Classification of Inflammatory Bowel Disease: Report Int. J. Mol. Sci. 2022, 23, 10868 15 of 15 of a Working Party of the 2005 Montreal World Congress of Gastroenterology. Can. J. Gastroenterol. 2005, 19, 5A–36A. [CrossRef] [PubMed] 43. Harvey, R.F.; Bradshaw, J.M. A simple index of Crohn’s-disease activity. Lancet 1980, 1, 514. [CrossRef] 44. Walmsley, R.S.; Ayres, R.C.S.; Pounder, R.E.; Allan, R.N. A simple clinical colitis activity index. Gut 1998, 43, 29–32. [CrossRef] 45. Lozupone, C.A.; Hamady, M.; Kelley, S.T.; Knight, R. Quantitative and qualitative beta diversity measures lead to different insights into factors that structure microbial communities. Appl. Environ. Microbiol. 2007, 73, 1576–1585. [CrossRef] [PubMed] 46. McMurdie, P.J.; Holmes, S. Phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE 2013, 8, e61217. [CrossRef] 47. Mallick, H.; Rahnavard, A.; McIver, L.J.; Ma, S.; Zhang, Y.; Nguyen, L.H.; Tickle, T.L.; Weingart, G.; Ren, B.; Schwager, E.H.; et al. Multivariable association discovery in population-scale meta-omics studies. PLoS Comput. Biol. 2021, 17, e1009442. [CrossRef]