An Expanded Gene Catalog of Mouse Gut Metagenomes

Abstract

High-quality and comprehensive reference gene catalogs are essential for metagenomic research. The rather low diversity of samples used to construct existing catalogs of the mouse gut metagenome limits the numbers of identified genes in existing catalogs. We therefore established an expanded catalog of genes in the mouse gut metagenome (EMGC) containing >5.8 million genes by integrating 88 newly sequenced samples, 86 mouse gut-related bacterial genomes, and 3 existing gene catalogs. EMGC increases the number of nonredundant genes by more than 1 million genes compared to the so-far most extensive catalog. More than 60% of the genes in EMGC were assigned to Bacteria, with 54.20% being assigned to a phylum and 35.33% to a genus, while 30.39% were annotated at the KEGG orthology level. Nine hundred two metagenomic species (MGS) assigned to 122 taxa are identified based on the EMGC. The EMGC-based analysis of samples from groups of mice originating from different animal providers, housing laboratories, and genetic strains substantiated that diet is a major contributor to differences in composition and functional potential of the gut microbiota irrespective of differences in environment and genetic background. We envisage that EMGC will serve as a valuable reference data set for future metagenomic studies in mice.IMPORTANCE We established an expanded gene catalog of the mouse gut metagenome not only to increase the sample size compared to that in existing catalogs but also to provide a more comprehensive reference data set of the mouse gut microbiome for bioinformatic analysis. The expanded gene catalog comprises more than 5.8 million unique genes, as well as a wide range of taxonomic and functional information. Particularly, the analysis of metagenomic species with the expanded gene catalog reveals a great novelty of mouse gut-inhabiting microbial species. We envisage that the expanded gene catalog of the mouse gut metagenome will serve as a valuable bioinformatic resource for future gut metagenomic studies in mice.

Keywords: diet; gene catalog; metagenomic species; mouse gut metagenome.

FIG 1

Construction of the EMGC. Metagenomic sequencing data of 88 mouse gut metagenomes were processed by the pipeline as displayed to generate nonredundant genes for PMGC. Unassembled strains of miBC (under BioProject PRJEB10572) were assembled and filtered by genome quality (completeness, >90%; contamination, <5%) of assembled genomes. Qualified genomes were used for gene prediction. CDSs from assembled genomes and downloaded genomes were gathered and clustered to MiCB. PMGC and MiCB along with 3 downloaded gene sets, FDGC, MGGC, and iMGMC, were merged to generate EMGC.

FIG 2

Performance of the EMGC. (A) Comparison of mapping rates between MGGC, iMGMC, and EMGC. ****, BH-adjusted P value of <0.0001 by Wilcox rank sum test. (B) Display of mapping rates among samples’ providers, including the Wallenberg Laboratory for Cardiovascular and Metabolic Research (CMR), the Jackson Laboratory in the United States (JUS), the laboratory animal center of Southern Medical University (SMU), the laboratory animal center of Sun Yat-Sen University (SYSU), and Taconic in Denmark (TDK) and in the United States (TUS). Dashed line represents a mapping rate of 80%. (C and D) Density curves for the mapping rates of fecal metagenomes and cecal metagenomes which were not included in gene catalog construction. Dashed lines represented the average values of the mapping rate of each gene catalog.

FIG 3

Description of new genes included in EMGC. (A) Two-dimensional (2D) density histogram showing the distribution of occurrences and mean relative abundances of new genes. (B) General display of taxonomic composition of new genes by Krona. (C) Frequency of functional pathways associated with the new genes. (D) Stacked histogram of KO coverage of functional pathways improved in EMGC compared to that in MGGC and iMGMC. Coverage is calculated as [(annotated KO numbers)/(total KO numbers)] × 100 in a given pathways.

FIG 4

Influence of diet on the composition of the microbiota. (A) PERMANOVA to estimate the influence of diet on the composition of gut metagenomes among all 7 sample groups. G1, C57BL/6 mice provided by Taconic in Denmark (TDK) and hosted in National Institute of Nutrition and Seafood Research of Norway (NIFES); G2, Sv129 mice provided by TDK and hosted in NIFES; G3, C57BL/6 mice provided by the Jackson Laboratory in the United States (JUS) and hosted by Pfizer-I; G4, C57BL/6 mice provided by Taconic in the United States (TUS) and hosted in Pfizer-I; G5, C57BL/6 mice provided by TDK and hosted in the University of Copenhagen (KU); G6, Sv129 mice provided by TDK and hosted in KU; G7, C57BL/6 mice provided by the laboratory animal center of Sun Yat-Sen University (SYSU). *, P < 0.05. Shannon index (**, BH-adjusted P < 0.01, Wilcox rank sum test) (B) and PCoA based on genus profile (C) for 7 groups fed the HF and LF diets. (D) Genera differently enriched (BH-adjusted P < 0.05, Wilcox rank sum test; relative abundance, >1e−5) in mice fed the HF and LF diets among all 7 groups. Genera in light red represent genera enriched in HF-diet-fed mice, while genera in light green represent genera enriched in LF-diet-fed mice.

FIG 5

Phylogenetic tree of the 902 MGSs and 830 high-quality iMGMC MAGs. MUMi distances for MGSs and MAGs were used to construct the phylogenetic tree using hierarchical clustering. MAGs are shown as red branches and MGSs as gray branches. The outer ring shows the relation between MGSs and MAGs. MGSs which have a MUMi value of >0.54 are marked as “Ann_MGS” in green blocks, otherwise, in purple blocks. MAGs are all in gray blocks. Colored blocks in the inner cycle indicate phyla assigned to MGSs and MAGs.