Page 26 - Read Online

Page 8 of 15 Ladeira et al. Microbiome Res Rep 2023;2:9 https://dx.doi.org/10.20517/mrr.2023.01

Figure 3. Variation of the gut microbiome between Bifidobacterium-based partitions. (A) Relative abundance of Bifidobacterium across
partitions; (B) species-based Shannon index for the gut microbiome; (C) differential analysis of the gut microbiome across partitions.
The top 30 most abundant species are depicted. Red indicates a higher species abundance in the first partition tested (significant when
FDR < 0.05, trend when FDR < 0.1). Species and partitions are ranked according to hierarchical clustering based on Euclidean distance,
with the exclusion of Bifidobacterium species from the graph.

hallii). The no_Bif partition was associated with a lower abundance of Streptococcus salivarius than the other
Bifidobacterium-based partitions.

Overall, the partitions enriched in both B. longum and B. adolescentis were associated with higher gut
microbiome diversity and abundance of butyrate producers and were more prevalent in healthy individuals
(i.e., health-associated Bifidobacterium communities).

A pangenomic analysis of Bifidobacterium reveals functions associated with gut ecology and health
Finally, we investigated whether intra-species functions were associated with Bifidobacterium partitions as a
surrogate for a more diverse gut microbiome and a higher prevalence of healthy subjects. We selected 2,263
MAGs constructed from an extensive dataset included in the cMD database. Therefore, we could pair
[16]
participants’ Bifidobacterium partitions with their Bifidobacterium MAG content. An analysis of 11,673
unique OGs functionally distinguished MAGs from different Bifidobacterium species on the basis of OGs
prevalence (chi-squared test, FDR < 0.05 within species) [Supplementary Table 2]. The most significant OGs
were that for asparagine synthase (COG0367), which was detected in 99% of the MAGs assigned to B.
adolescentis and 0.25% of those from other species. We confirmed known functional differences, relating,
for example, to glycoside hydrolases (GH) involved in the metabolism of host carbohydrates (mucin/milk),
such as GH 20, GH 29, GH 33, and GH 95, which were specific to B. bifidum and had prevalences ranging
from 90 to 97%, vs. 0.1% in other species. Similarly, we detected a high prevalence of alpha-L-
arabinofuranosidase (COG3534) for the MAGs of B. longum (99% vs. 0.7% in other species). For B.
pseudocatenulatum, we detected OGs assigned to the GH 43 family (xylosidase).