Page 10 - Read Online

Page 4 of 18 Mueller et al. Microbiome Res Rep 2024;3:33 https://dx.doi.org/10.20517/mrr.2024.09

Additional data access
T
Additional Akkermansia genomes (n = 221), including that of Muc and excluding MAGs, were retrieved
from NCBI. A summary of each isolate can be found in Supplementary Table 1. The 16S rRNA-based
[39]
microbial community profiling results in the form of a phyloseq object and raw metagenomic reads for
validation of the proposed phylogroup prediction methods were obtained from the Pediatric Obesity
Microbiome and Metabolism Study (POMMS) . Metagenomic sequencing samples for the inflammatory
[40]
bowel disease (IBD) and non-small cell lung cancer (NSCLC) datasets were obtained from the Sequence
Read Archive using Bioproject accessions PRJNA398089, PRJNA400072, PRJEB42151, PRJEB42155,
PRJNA751792 and PRJNA782662 [41-44] .

Pangenomic analysis
A comparison of 234 A. muciniphila genomes was performed using the pangenomic workflow in Anvi’o
version 7.1 . Briefly, this workflow included construction of a pangenome from genome fasta files [46,47] ,
[45]
[48]
calculation of average nucleotide identity of assigned phylogroup , core gene cluster identification and
single-copy core protein sequence alignment , estimation of functional enrichment , and extraction of
[49]
[50]
full-length 16S rRNA sequences . An in-depth description of these methods and the parameters used may
[51]
be found at https://gitlab.oit.duke.edu/valdivia-lab/akkermansia-species-and-phylogroups.
Phenotypic characterization
Fatty acid methyl ester profiling of select strains was performed by EMSL Analytical, Inc. The API 20A
system (bioMérieux catalog 20300) was used to measure 21 different biochemical phenotypes, including
indole formation, urease and catalase activity, gelatin and esculin hydrolysis, and 16 acidification reactions.
These tests were performed according to the manufacturer’s protocol under anaerobic conditions for 48 h.

Analysis of community profiling data
The 16S rRNA-based microbial community profiling of the POMMS cohort was obtained as a phyloseq
[40]
object , which included the prevalence of identified amplicon sequence variants (ASVs). Identification of
Akkermansia species and phylogroups from the data was performed using Phyloseq (v1.32.0) in
[39]
R (v4.0.4) . Raw metagenomic sequencing data were processed using the bioBakery suite [53,54] , StrainR ,
[52]
[55]
and strain-level metagenomic estimation of growth rate (SMEG) . Both the 16S rRNA and metagenomics
[56]
analyses are detailed with examples at https://gitlab.oit.duke.edu/valdivia-lab/akkermansia-species-and-
phylogroups. In all cases, POMMS samples with Akkermansia of known phylogroup were used as controls.
Statistics and visualization
Figures were finalized using Inkscape . Metabolic pathway completeness was visualized using the package
[57]
[58]
pheatmap (v1.0.12) in R. Phylogenetic trees were visualized using the Interactive Tree of Life . GraphPad
[59]
Prism 10 (v10.1.0) was used to generate all remaining graphs, calculate the average ANI and 16S rRNA
sequence identity between phylogroups, and perform statistical analyses. Principal Component Analysis of
FAME results used the standardized method in GraphPad Prism. Mann-Whitney U tests were used to
compare relative abundances in datasets containing two disease groups, Kruskal-Wallis H tests were used to
compare relative abundances in datasets containing more than two disease groups, and Cox regression was
used for survival analyses.