Microbial composition at the phylum level
The microbiota's relative abundance at the phylum level reveals some interesting patterns across different regions. Generally, among African indigenous populations, Bacteroidetes dominates at 57%, followed closely by Firmicutes at 36%, irrespective of geographical locations and subsistence patterns. However, in Central Africa, there is a noticeable presence of Proteobacteria at 11%, in addition to Bacteroidetes and Firmicutes.
In Asia, the predominant phylum is Firmicutes at 50%, with Bacteroidetes following at 32%, regardless of specific subsistence practices. Nevertheless, in West Asia, Actinobacteria takes the lead at 24%, followed by Proteobacteria at 11%, after Firmicutes. In South Asia, Proteobacteria holds a significant share at 14%, while in Southeast Asia, Actinobacteria takes the lead at 10%, followed by Proteobacteria at 5%, both after Firmicutes and Bacteroidetes.
Contrastingly, the North American indigenous population displays a more diverse composition at the phylum level. Firmicutes holds the highest abundance at 44%, succeeded by Bacteroidetes at 35%, Actinobacteria at 9%, Proteobacteria at 7%, and Verrucomicrobia at 5% (refer to Figure 1 and Figure 2). This diversity sets the North American gut microbiota apart from other global regions.
Microbial composition at the genus level
The genus-level distribution of the microbiota among indigenous populations is depicted in Figure 3. Prevotella stands out as the predominant genus in 17 out of 21 indigenous populations. In African indigenous communities, irrespective of geographical and subsistence-specific variations, Prevotella emerges as the most abundant bacterial genus, averaging over 55%. Notably, there's a consistent prevalence of Prevotella at 66% in the south and west, while Central Africa exhibits a different composition. The genus Bacteroides is found at 15% in South Africa, 3% in East Africa, and 6% in Central Africa, whereas Bifidobacterium is exclusively present in Central African populations. Oscillibacter is present at 6% in the south and west, dropping to less than 1% in central African regions. Sucinivibrio is at 8% in central Africa and less than 1% in both South and West Africa. Spirochaetaceae surpasses 6% in the Central region, contrasting with less than 1% in the South and 1% in the West. Roseburia is present at 1% in the South and West, increasing to 3% in the Central region. Conversely, Fecalibacterium maintains a consistent composition across the entire geographic region at 7%.
The core gut microbiota in South African populations comprises Prevotella (66%), Bacteroides (15%), Fecalibacterium (7%), Roseburia (1%), Oscillibacter (6%), Sutterela (1%), Alistipes (1%). In East Africa, Prevotella dominates at 72%, accompanied by Bacteroides (3%), Fecalibacterium (6%), Roseburia (1%), Spirochaetaceae (1%), Oscillibacter (6%), Sucinivibrio (1%). Central Africa's core microbiota includes Prevotella (50%), Bacteroides (6%), Fecalibacterium (8%), Sucinivibrio (8%), Roseburia (3%), Spirochaetaceae (6%), Paraprevotella (1%), Alistipes (1%), Coprococcus (1%), Ruminobacter (2%), Spirochaeta (6%).
Conversely, in South and Southeast Asian populations, the genus Prevotella dominates, constituting 34% of the microbiota regardless of subsistence patterns. In contrast, West Asia exhibits a significant prevalence of Enterococcus, accounting for a substantial 33%. Similar to Africa, there is considerable variation in bacterial composition across different geographic regions in Asia, but this diversity is even more pronounced compared to Africa.
In Asia, the core microbial abundance shows more similarities between Southeast and South Asian regions compared to West Asia. Bifidobacterium is present at 6% in Southeast Asia, 2% in the South, and notably higher at 22% in West Asia. In Southeast and Western Asia, there are similar abundances for the genera Blautia (3%) and Collinsella (6%), while in South Asia, they are less than 1%. On the other hand, Southeast Asia features genera such as Coprococus (5%), Dorea (2%), Succinivibrio (5%), and Hemphlus (3%), which differ from South and West Asia where they are less than 1%.
Among the core microbiota, a consistent abundance pattern is observed in South and Southeast Asia, including Prevotella (34%), Fecalibacterum (8%), Bacteroides (7%), Roseburia (7%), Streptococcus (5%), Ruminoccus (2%), Lactobacillus (1%), and Megamonas (1%). In South Asia, the core gut bacterial genera, covering 85% of the total bacterial abundance, include Bacteroides (10%), Bifidobacterium (2%), Fecalibacterum (9%), Hemphlus (3%), Lactobacillus (1%), Prevotella (37%), Roseburia (11%), Megamonus (2%), Septococcus (6%), Ruminoccus (1%), Succinivibrio (2%), Klebsiella (4%). Meanwhile, the Southeast Asian population exhibits core genera such as Bacteroides (7%), Bifidobacterium (7%), Blautia (3%), Collinsella (5%), Coprococus (5%), Dorea (2%), Fecalibacterum (8%), Lactobacillus (1%), Prevotella (30%), Roseburia (5%), Septococcus (5%), Succinivibrio (5%), Ruminoccus (2%). West Asia displays a core microbiota composition with Bifidobacterium (22%), Blautia (3%), Collinsella (9%), Enterococcus (33%), Lactobacillus (1%), Pediococus (2%), Septococcus (14%), Halomonas (4%), Shewanella (1%).
The gut bacterial composition in the North American indigenous population is notably unique. The core gut bacteria within this population include Lactobacillus at 1%, Halomonas at 1%, Ruminococcus at 2%, Streptococcus at 2%, Bifidobacterium at 3%, Alistipes at 3%, Roseburia at 4%, Blautia at 5%, Gardnerella at 7%, Faecalibacterium at 7%, Verrucomicrobiaceae at 8%, Prevotella at 8%, and Bacteroides at 35%.
Alpha diversity across different lifestyles
In terms of alpha diversity, the gut microbiota exhibits the highest diversity among agricultural populations, followed by hunter-gatherer, pastoralist, and agropastoralist populations. There are significant differences in bacterial diversity among populations with distinct subsistence-specific lifestyles. Notably, hunter-gatherer population’s show pronounced difference from agricultural populations compared to pastoralist populations, with agropastoralist populations falling in between, as depicted in Figure 4 in terms of species richness and evenness. Conversely, the Simpson diversity index reveals similar diversity differences between hunter-gatherer and agricultural populations. It is noteworthy that the dissimilarity in species compositions is distinct between hunter-gatherer and agropastoralist populations than between hunter-gatherer and pastoralist populations, as illustrated in Figure 5.
Beta diversity across different geography and subsistence-patterns
Regarding beta diversity, the Bray-Curtis Similarity statistics illustrate the relatedness among populations based on the genus-level abundance of the gut microbiota (Figure 6). The Bedouin of West Asia (Saudi Arabia), Temuan of Southeast Asia (Malaysia), and Inuit of North America (Canada) exhibit a higher level of dissimilarity compared to other populations. The Bray-Curtis Cluster analysis further depicts that indigenous populations within the same geographical territory show less distinctness. Specifically, populations from the same country tend to form close clusters, suggesting a similar pattern of beta diversity at the genus level when compared to others (Figure 6B).
In the context of beta diversity, the Principal Coordinates Analysis (PCoA) using Bray-Curtis dissimilarity at the genus level reveals notable differences in the abundance of gut microbiota across countries, continents, and lifestyles (subsistence patterns), explaining nearly 60% of the total variation (Coordinate 1 = 43.98%, Coordinate 2 = 17.88%) (Figures 7A, 7B, and 8). These distinctions were further supported by the PREMANOVA test results, indicating significant variations between countries (p = 0.0001, R2 = 0.2791, F = 6.026), continents (p = 0.001, R2 = 0.6151, F = 5.627), and lifestyle variations (p = 0.1897, R2 = 0.1827, F = 1.305). The genus-level abundance of gut bacterial taxa is significantly diverse across geographical regions (p < 0.05).
Figure 7B illustrates that Asian populations exhibit greater diversity compared to African populations. In (Figure 7A), both Asian and African countries form distinct clusters. Notably, the Bedouin of West Asia stands out as more distinct from other Asian countries. Additionally, the Inuit of North America holds a separate position in Figure 7. In contrast to Asian populations, African populations constitute a separate cluster and show significant diversity among themselves (p = 0.003), regardless of their specific countries (Figure 7B). However, within Africa, individual countries form separate clusters (Figure 7A).
On the other hand, beyond geographical differences, the genus-level abundance of gut bacterial taxa does not significantly differ across subsistence-specific variations in terms of beta diversity. Subsistence clusters overlap (Figure 8), providing limited insight into lifestyle-specific differences in gut microbiota. However, pairwise PREMANOVA tests indicate a significant difference in genus-level abundance between agropastoral and agricultural groups (p = 0.0087).
In Figure 9, the ribbons depict the movement of genera between different groups. Genera specific to a particular group flow ribbons towards the value '0,' while those in transition are shared with other groups based on their abundance percentage. The alluvial plot, visualized with the Similarity Percentage (SIMPER) of taxa abundance, identifies bacterial genera contributing significantly to the observed differences in Figure 9A between continents, Figure 9B among subsistence patterns, and Figure 9C among countries. The analysis considers an average dissimilarity of at least 1%, each genus contributing at least 1%, and a cumulative contribution of each bacterial genus of at least 90% for each plot. Among continents, a total of 19 bacteria (3 families and 16 genera) are identified with substantial contributions, being the primary contributors to the observed differences between groups. Specifically, Prevotella (p < 0.0001) for Africa and North America, as well as Bacteroides (p = 0.012, p = 0.049, p < 0.0001) for Africa, Asia, and North America, significantly contribute to the diverse composition of these groups (p values are Bonferroni corrected) (Figure 9A).
In terms of subsistence-specific group differences, a total of 17 microbiota (3 families and 14 genera) are identified as primary contributors to the observed distinctions between groups. Notably, bacterial genera such as Prevotella (p = 0.040, p = 0.052) for agropastoralist and agricultural groups, Enterococcus (p < 0.0001) for pastoral group, and Porphyromonadaceae family (p = 0.017) for the agricultural population significantly contribute to the diversity within these groups (p values are Bonferroni corrected) (Figure 9B).
In terms of country-wise differential abundance, 18 bacteria (3 families and 15 genera) play a crucial role in creating group distinctions, and the flow of bacteria between groups is depicted in Figure 9C. Among these bacterial contributors, Prevotella is a significant genus that notably contributes to group differences (p = 0.0008) for Botswana, (p < 0.0001) for Tanzania, Saudi Arabia, and Canada. Additionally, Bacteroides (p = 0.022, p < 0.0001) for Saudi Arabia and Canada, Porphyromonadaceae family (p < 0.0001), Erysipelotrichaceae family (p < 0.0001), and Coprococcus genus (p = 0.004) for Malaysia, Bifidobacterium, Enterococcus, Streptococcus, Collinsella (p < 0.0001) for Saudi Arabia, Succinivibrio (p = 0.0006) and Spirochaetaceae family (p < 0.0001) for Cameroon, and Roseburia (p = 0.002) for Bangladesh significantly contribute to group differences (p values are Bonferroni corrected). Across geography and subsistence-specific variations (Figure 9A, 9B, 9C), Prevotella stands out as the most influential bacterial genus, consistently contributing to significant group differences. In addition to Prevotella, the genera Bacteroides, Enterococcus, and the family Porphyromonadaceae are also notable contributors to creating group differences in both geographical and lifestyle practices.
For additional p values not mentioned in the main text, refer to Supplementary Table.