Page 16 - Read Online

Page 10 of 18 Mueller et al. Microbiome Res Rep 2024;3:33 https://dx.doi.org/10.20517/mrr.2024.09

Figure 3. The relative abundance of the three main Akkermansia species and A. muciniphila sub-phylogroup in two different patient
cohorts can correlate with disease outcomes. (A) In vivo growth rate is a minor contributor to the relative abundance of Akkermansia
species in stool samples. A simple linear regression was performed to determine if the relative abundance of Akkermansia in a
metagenomics sequencing sample is predictive of the estimated growth rate as determined by SMEG. Each regression line is colored by,
and represents, either A. muciniphila, A. massiliensis, or A. biwaensis; (B) Comparison of the relative abundances of A. muciniphila,
A. massiliensis, and A. biwaensis between children with obesity, defined as having a gender-specific BMI greater than the 95th percentile,
and control samples from the POMMS dataset; (C) Comparison of the relative abundances of A. muciniphila, A. massiliensis, and
A. biwaensis between healthy controls, CD, and UC samples from the combined IBD dataset; (D) Comparison of the relative abundances
of AmIa and AmIb A. muciniphila between children with obesity and control samples from the POMMS dataset; (E) Comparison of the
relative abundances of AmIa and AmIb A. muciniphila between healthy controls, CD, and UC samples from the combined IBD dataset.
SMEG: Strain-level metagenomic estimation of growth rate; BMI: body mass index; POMMS: Pediatric Obesity Microbiome and
Metabolism Study; UC: ulcerative colitis; CD: Crohn’s disease.

metagenomic samples with matching isolates, we identified 26 samples with measurable levels of
Akkermansia, as assessed by MetaPhlAn3, for which we could predict the prevalent phylogroup by a
combination of StrainPhlAn3, StrainR, and SMEG. For these samples, 24/26 (92%) of the isolated strains
matched the predicted phylogroup. Eight samples in the entire dataset were predicted to contain more than
one Akkermansia species, accounting for a range of 1.57%-33.50% and a median of 13.83% of Akkermansia
abundance of the additional minor species. Note that this method has only been validated to distinguish
between the A. muciniphila, A. massiliensis, and A. biwaensis species, as the validation dataset did not
include Akkermansia AmIII, AmV, or AmVI isolates.

We noted that full-length 16S rRNA gene sequences clustered by the new proposed Akkermansia species
nomenclature. We determined that this clustering is maintained if we restrict our analysis to the V3-V4
subregion, which is commonly used in 16S rRNA gene-based microbial community profiling
[Supplementary Figure 5]. Therefore, we looked at 16S rRNA V3-V4 sequences generated in the POMMS
study to determine if we can make species-level assignments in this cohort. The sequences of amplified
sequence variants (ASVs) annotated as genus Akkermansia were extracted and aligned to 16S rRNA gene
sequences from our reference strains, then assigned to new Akkermansia species, depending on with which