The taxonomic analysis by WGS at Genus level
Next, we analysed the frequency of most abundant genera in 12 fecal samples at four different altitudes (Figure 4). The relative abundance of around 20 genera was more than 1%. At H1 time point genus Bacteroides was the most abundant (46%), followed by Prevotella (19%), Parabacteroides and Alistipes (7%), Eubacterium (5%), Faecalibacterium (3%) Ruminococcus (2%), Blautia, Coprobacillus, Clostridium, Escherichia, Bifidobacterium, and unclassified (1%) (Figure 4 A). Interestingly, after ascending to 4420m (H2) and staying for 6 months, the richest genus was Prevotella (63%) with significant decrease in Bacteroides to 2%, followed by Eubacterium, Faecalibacterium 3%, and Parabacteroides 1%. There was a marginal increase in Coprococcus contribution to 4%. Surprisingly even after getting adapted to stressful environment at H2, on descending to H3, the scenario changed markedly where Bacteroides reversed and increased significantly to 31%, though still lower than Prevotella which was reduced to 43%. Incidently, there was sudden appearance of Fusobacterium (5%), Roseburia (4%) and Escherichia (2%). The pattern was similar to H1. Interestingly, again the phylogeny structure was changed when they ascended to 5800m (H4) and stayed for four months. Bacteroides significantly reduced to 9%, Prevotella significantly increased to 65%, Eubacterium (3%), Faecalibacterium (2%), whereas Parabacteroides, Alistipes, Coprococcus, Escherichia, and Coprobacillus (1%) started reappearing. The pattern looked similar to H2. When analyzing the overlap of different genera at four different altitudes, it mostly belonged to Bacteroides and Prevotella, where Prevotella covered the 63% of the total genera at H2 and H4 heights and Bacteroides at H1 and H3. Although 3500m (H3) is defined as high altitude, interestingly, the pattern of the metagenomic profile obtained at H3 was quite similar to the base line profile (H1= 210m).
Further confirmation of the differences in the composition of fecal microbiota of sojourners at different heights was achieved by LEfse using the Greengene software. FDR correction of the p value was to determine statistical significance (p < 0.05/number of tests) and considered p < 0.05 to be statistically significant for reproducibility (Figure 4D). Thus the genus abundance clearly indicated the significant change in the microbiota mainly with abundance of Bacteroides (p= 0.05) and Prevotella (p= 0.05). LEfse analysis also showed concordant result with Wilcoxon rank test (Supplementary FigureS4 with LDA score) indicating altitude or hypoxia, could be the most important contributing factor to the diversity.
The WGS analysis of three subjects revealed the presence of 20 genera. Amongst these, the relative abundance of genus Bacteroides, Prevotella, Bifidobacterium and Fecalibacteria were significantly altered (with p <0.05 Kruskal Wallis H test) (Figure 4 A). We reproduced and correlated the significant presence of 20 genera in another 16 subjects of the same cohort at H1 and H2 heights (p<0.05 Wilcoxon rank test) with odds ratio of 95% (Figure 4B), by 16s rRNA analysis. The second method reconfirmed the changes in the abundance of genus Bacteroides, Prevotella, Faecalibacterium & Bifidobacterium as in WGS (Figure4 C). Interestingly compared to WGS, the 16s rRNA also had higher prevalence detection rate for other genus like Ruminococcus and Akkermansia at H2 height.