Table 2.
Description of methods and results of microbial diversity and richness assessment.
| Study ID (Author, Year) | Comparison Groups 1 | Alpha Diversity Method 2 | Alpha Diversity and Richness 1,2 | Beta Diversity Method | Beta Diversity |
|---|---|---|---|---|---|
| Andoh 2016 [17] | O vs. L | Shannon index | Significantly lower in O (2.40 ± 0.17) vs. L (2.80 ± 0.14) p < 0.01 |
PCA 3 | PCA 3 at phylum level showed different distribution of O and L peoples |
| Beaumont 2016 [37] | High vs. Low BMI | Shannon index | Significantly lower in high vs. low BMI (p = 0.0001) | — | — |
| Borges 2018 [38] | — | — | — | — | — |
| Borgo 2018 [18] | O vs. NW | Shannon index, observed species and Faith’s PD | Significantly lower α-diversity (PD) in O vs. n (p < 0.01). Shannon index and observed species are not reported by BMI group |
Weighted and unweighted UniFrac 4 metrics and PCoA 5 Bray–Curtis distances | No separation was obtained between O and NW subjects (p > 0.05). Significant separation in LAM 7 samples between NW and O was observed |
| Chavez-Carbajal 2019 [19] | O vs. OMS vs. NW | Shannon index, Simpson index, Chao1, observed species. | Shannon index: no significant difference between groups O: 6.61 ± 0.36 OMS: 6.56 ± 0.38 NW: 6.32 ± 0.46 O vs. NW (p = 0.17) OMS vs. NW (p = 0.09) Simpson index: no significant difference between groups O: 0.97 ± 0.01 OMS: 0.97 ± 0.01 NW: 0.97 ± 0.02 O vs. NW (p = 0.28) OMS vs. NW (p = 0.52) Chao1 index: significantly higher in O vs. NW O: 787.1 ± 137.8 OMS: 769.4 ± 101.7 NW: 583.5 ± 87.8 OMS vs. NW (p = 0.003) O vs. NW (p = 0.002) |
Unweighted UniFrac 4 analysis, PCoA 5 | For the unweighted, PCoA 4 analysis clearly grouped the O and OMS separating them from the NW (ANOSIM 8; p = 0.01). Weighted analysis showed a similar result (ANOSIM 8, p = 0.01) |
| Chen 2016 [20] | — | Shannon index | NR by BMI groups | unweighted and weighted UniFrac 4 distances | NR by BMI groups |
| Davis 2016 [40] | O vs. OW vs. NW | Simpson’s Index of Diversity, Chao1, Shannon index Effective Number of Species | Shannon index Effective Number of Species: No significant differences reported O: 228.2 ± 134.1 OW: 218.1 ± 134.2 NW: 179.9 ± 103.1 p = not reported Chao1 and Simpson’s Index NR |
— | — |
| Davis 2020 [39] | High FMI vs. Low FMI | Shannon index, Fishers index, Chao 1, Observed species | The alpha diversity and richness indices were lower in the high versus low FMI 1 group: Shannon index: no significant differences (data not shown) Fisher index: no significant differences (MD −6.2, 95%CI −12.7, 0.4; p = 0.065). Chao1 index: significantly lower in high vs. low FMI (MD −46.1, 95%CI −90.2, −2.0; p = 0.040); Observed species: significantly lower in high vs. low FMI (MD −46.1, 95%CI −86.5, −5.7; p = 0.026) |
— | — |
| De la Cuesta-Zuluaga 2018 [41,42] | O vs. OW vs. L | Shannon index and number of observed OTUs | Shannon index: significant differences (lower in O/OW vs. L) O: 3.1 ± 0.7 OW: 3.0 ± 0.7 L: 3.2 ± 0.7 p = 0.04 # observed OTUs: significant differences (lower in O/OW vs. L) O: 142.5 ± 36.4 OW: 138.6 ± 35.6 L: 153.5 ±38.8 p = 0.002 |
Weighted and unweighted UniFrac 4 matrices (PERMANOVA 6) | NR by BMI groups |
| Fei 2019 [43] | O vs. OW vs. L | Shannon index, Chao1 diversity, observed OTUs | NR by BMI groups | Beta diversity (PERMANOVA 6) | NR by BMI groups |
| Finucane 2014 [44] | O vs. L | Shannon index, observed OTUs | Shannon index: No differences in O vs. L Richness (total number of OTUs): No difference in O vs. L |
— | — |
| Gallè 2020 [45] | O/OW vs. NW/UW | Shannon index | No significant differences in O/OW (2.5 ± 0.2) vs. NW/UW (2.5 ± 0.2) p = 0.77 |
PCoA 5 using the METAGEN assist platform | ANOSIM 8 test yielded no significant dissimilarity for the BMI groups (R = −0.011, p = 0.5) |
| Gao 2018 [46] | O vs. NW | Shannon, Simpson, Number of observed OTUs | Shannon index: Significantly lower in O vs. NW (p < 0.01) Simpson index: Significantly lower in O vs. NW (p < 0.001) # observed OTUs: No significant difference in O vs. NW |
PCoA 5 of samples by weighted and unweighted UniFrac 4 distance | Fecal microbial communities of the four BMI groups were not distinct from each other, indicating low among-group dissimilarities |
| Harakeh 2020 [47] | O vs. OW vs. NW vs. UW | Shannon index | No difference between UW, NW, OW and O individuals | — | — |
| Kaplan 2019 [48] | NW vs. OW vs. O (class I, II, III) | Shannon index | Significantly lower in O Class III vs. NW: Beta (95% CI) NW Ref. OW −0.01 (−0.08, 0.10) O Class I −0.08 (−0.18, 0.01) O Class II −0.09 (−0.21, 0.03) O Class III −0.19 (−0.35, −0.03) |
Bray-Curtis distances | NR by BMI groups |
| Kasai 2015 [21] | O vs. L | Shannon index | Significantly higher in O vs. L p < 0.05 |
PCA 3 | L subjects formed a cluster distinct from O subjects |
| Loftfield 2020 [49] | O vs. OW vs. NW | Shannon index, Faith phylogenetic diversity index (PD), and number of observed sequence variants | Shannon index, PD and number of observed sequence variants: No significant differences across groups. | Bray–Curtis and unweighted UniFrac 4; PCoA 5 | Being O compared with normal BMI at age 46 was statistically significantly associated with Bray–Curtis, unweighted Uni-Frac, and weighted UniFrac distances (all p values ≤ 0.001); whereas OW BMI and BMI history were not statistically significantly associated with the beta diversity matrices |
| Oduaran 2020 [34] | O vs. L | Shannon index, Chao1 | Shannon diversity in Bushbuckridge: No significant differences in O (4.56 ± 0.39) vs. L (4.49 ± 0.53 (4.56 ± 0.41 after exclusion of an outlier)) p = 0.85 Chao 1: Significantly higher in O vs. L (p = 0.001) Shannon diversity in Soweto: No significant differences in O (4.30 ± 0.56) vs. L (4.49 ± 0.34) p = 0.45 Chao 1: No significant differences in O vs. L (p = 0.33) |
Bray-Curtis distances, PcoA 4,5 | Beta diversity measurements showed statistically significant differences between the lean and obese groups in Bushbuckridge with calculated Bray-Curtis distances using the permutational analysis of variance (PERMANOVA 6) test (p = 0.02 for Bushbuckridge and p = 0.84 for Soweto) |
| Org 2017 [50] | — | Pielou’s index (evenness), and Fisher’s alpha (diversity) | NR by BMI groups | Bray–Curtis distance | NR by BMI groups |
| Osborne 2020 [51] | Tertile 1: 12.9–19.1 kg/m2; Tertile 2: 19.1–23.4 kg/m2; Tertile 3: 23.4–38.9 kg/m2 |
Shannon index, Chao1, number of observed OTUs | Shannon index: Significant decrease across BMI tertiles tertile 1: 4.6 ± 0.5 tertile 2: 4.5 ± 0.5 tertile 3: 4.4 ± 0.5 p < 0.01 Chao1: Significant decrease across BMI tertiles tertile 1: 10,848 ± 3916 tertile 2: 9761 ± 3006 tertile 3: 9162 ± 3590 p = 0.02 # observed OTUs: Non-significant decrease across BMI tertiles tertile 1: 3613 ± 1462 tertile 2: 3160 ± 999 tertile 3: 3093 ± 1317 p = 0.07 |
Unweighted UniFrac 4, weighted UniFrac 4, and Bray-Curtis distances | NR by BMI groups |
| Ozato 2019 [32] | High VFA vs. Low VFA | Shannon index | Significantly lower in High vs. Low VFA in men (p = 0.053) Non-significantly higher in High vs. Low VFA in women (p >0.05) |
— | — |
| Patil 2012 [52] | O vs. NW | Shannon index and Simpson index | Shannon index: No significant differences reported O: 2.89 ± 0.56 NW: 2.84 ± 0.75 p = not reported Simpson index: No significant differences reported O: 0.11 ± 0.08 NW: 0.14 ± 0.18 p = not reported |
UniFrac 4 analysis | Library cluster analysis clearly demonstrates clustering of lean and normal libraries except L3 (which has an unusually high Bacteroides genus counts). Interestingly, libraries O1 and O2 cluster in the normal/lean clade |
| Peters 2018 [33] | O vs. NW | Shannon index, Richness, and Evenness | Shannon index: Non-significantly lower in O vs. NW (beta = −0.11, p = 0.11, pHolm = 0.22) Evenness: Non-significantly lower in O vs. NW (beta = −0.01, p = 0.22, pHolm = 0.44). Richness (i.e., number of OTUs): Significantly lower in O vs. NW (beta = −9.87, p = 0.04, pHolm = 0.08); Significantly lower richness in O vs. NW in women (p = 0.03), but not in men (p = 0.47) |
Weighted UniFrac4 distance, PCoA 5 | Partial constrained analysis of PCoA 4 of the weighted UniFrac distance revealed separation of obese from both healthy-weight and OW participants on the main axis, with OW separated from healthy-weight participants on the secondary axis, although PCoA 4 did not reveal clustering by BMI category. In PERMANOVA 6 analysis of the weighted UniFrac distance, BMI category was not associated globally with overall microbiome composition (p = 0.14). In pairwise comparisons, overall microbiome composition differed between O and HW participants (p = 0.04, pHolm = 0.07), while OW and HW participants did not differ significantly (p = 0.64, pHolm = 0.64) |
| Rahat-Rozenbloom 2014 [53] | OW vs. L | Shannon index | No significant difference in OW (4.66) vs. L (4.92) p = 0.18 |
Weighted UniFrac 4 distances | PCoA 5 plots failed to reveal any difference in between the L and OW groups (data not shown) |
| Salah 2019 [54] | O vs. OD vs. NW | Shannon index, Number of OTUs | Shannon index: Significantly lower in O vs. NW (p < 0.01) Significantly higher in OD vs. NW (p < 0.05)- Number of OTUs: NR |
PCoA 5 unweighted and weighted UniFrac 4 distance matrix | PCoA 5 plot based on unweighted Uni-Frac was built and showed significant BMI and diabetes-dependent clustering of samples (PERMANOVA 6; p = 0.001) |
| Thingholm 2019 [55] | O vs. L | PD calculated using the phylogenetic tree built on the aligned OTU sequences | PD significantly lower in O vs. L (p = 3.20310−11 by robust regression). | Function betadisper from the R package vegan with default settings to evaluate dispersion between groups | Composition (beta-diversity) of taxonomic and functional profiles (adonis q < 0.1), and taxonomic evaluation of dispersion (genera, betadisper q < 0.1) significantly lower in O vs. L, although not for functional features (betadispersion q > 0.1) |
| Verdam 2013 [56] | O vs. NO | Simpson’s reciprocal index of diversity (1/D) | Significantly lower in O (128.7 ± 33.2) vs. NO (174.6 ± 37.3) p = 0.002 |
— | — |
| Vieira-Silva 2020 [57] | — | Observed richness was calculated using phyloseq | NR by BMI groups | PCoA 5 using Bray–Curtis dissimilarity with Hellinger transformation | NR by BMI groups |
| Whisner 2018 [58] | — | PD metrics calculated by QIIME via Faith’s PD | NR by BMI groups | PCoA 5 using weighted and unweighted UniFrac 4 distances | NR by BMI groups |
| Wilkins 2019 [59] | — | — | — | Weighted UniFrac 4 beta-diversity | NR by BMI groups |
| Yasir 2015 [35] | O vs. NW | Shannon Index Chao Index Number of OTUs |
Shannon Index, Chao Index, and Number of OTUs reported at OTU cutoffs of 3, 6 and 9 distance units France O significantly lower diversity and richness than NW at all the OTU cutoffs (p < 0.05). Saudi Arabia No significant difference in diversity and richness between O and NW at all the OTU cutoffs |
PCoA 5 calculated in QIIME by choosing Bray–Curtis distance methods at the genus level | PCoA 5 showed that O and NW individuals clustered independently. NW individuals from France and Saudi Arabia clustered together, but O Saudis clustered independently from obese French |
| Yun 2017 [60] | O vs. OW vs. NW | PD metrics calculated by QIIME 9 | Significantly lower diversity (PD) in O vs. NW (p < 0.01) and OW vs. NW (p < 0.01) | PCoA5 of weighted UniFrac 4 | Weighted UniFrac 4 PCoA 5 identified significant differences between groups (ANOSIM 8; R = 0.020, p = 0.001) |
1 L: Lean; NO: non-obese; NW: Normal weight; OW: Overweight; O: Obese; OD: obese diabetic; OMS: Obese and metabolic syndrome; UW: underweight; VFA: visceral fat area; FMI: fat mass index; 2 PD: phylogenetic distance; OTUs: operational taxonomic units; 3 PCA: Principal component analysis; 4 UniFrac: unique fraction metric; 5 PCoA: Principal Coordinates analysis; 6 PERMANOVA: permutational multivariate analysis of variance 7 LAM: lumen-associated microbiota; 8 ANOSIM: Analysis of Similarities; 9 QIIME: Quantitative Insights Into Microbial Ecology. NR: not reported.