Gut microbiome in Kazakh adults, cross-sectional study

While population stratification has been extensively studied and quantified in sexually reproducing species, it has been only anecdotally described in others. Here, we conducted a large-scale survey of population stratification in prokaryotic species, sampled from their natural habitat, without isolation or cultivation biases. We surveyed the variation landscape of 71 prevalent microbial species in 2144 human fecal metagenomes, and show that in 44 of these, accounting for 72% of the total assigned microbial abundance, single nucleotide variation can be clustered and stratified into sub-populations (termed subspecies). Individuals are usually dominated by only one such subspecies (per species), as expected of closely related organisms from ecological theory. Geographical dispersal patterns differ between phyla, with Firmicutes subspecies being significantly more geographically restricted. To demonstrate the functional significance of the delineated subspecies, we performed a metagenomic analysis, independent of reference genomes, of the gene content that consistently differs between conspecific subspecies, allowing us to associate groups of genes to properties of the microbial community and to the host. For example, some Eubacterium rectale subspecies specifically harbor an accessory pro-inflammatory flagellum operon that is associated with lower gut community diversity, higher host BMI and higher blood fasting insulin levels. Taken together, subspecies are a natural taxonomic level in prokaryotes and facilitate the functional and ecological understanding of the human gut microbiome in conjunction with its host.

Classified as human gut / stool shotgun DNA metagenome of age unknown by sonnet_x2 with confidence 0.85. Assay shotgun_dna.

• host run.scientific_name · “Homo sapiens”
• body site study.title · “human gut microbiome”
• body site study.description · “2144 human fecal metagenomes”

2,241 runs over 283 samples · mean 0.70 Gbp per run (median 0.30 Gbp, total 1,569.84 Gbp; 2,241 runs with a base count) · layout PAIRED 1,122, SINGLE 1,119 · platform ILLUMINA 2,241 · models Illumina HiSeq 2000 (2,241) · 36% of runs with a Sandpiper profile · first public 2017-12-01

Abdill RJ first · Adamowicz EM · Blekhman R last · Huang Y first · Liu J · Tun HM · Stanton C · Chen T · El-Nezami H · Wei H · Wang M · Wu Q last · Orihara K first · Yahagi K · Saito Y · Watanabe Y · Sasai T · Hara T · Tsukuda N · Oki K · Fujimoto J · Matsuki T last · Chanda D first · De D last · Sáenz JS first · Yergaliyev T · Rios-Galicia B · Seifert J · Camarinha-Silva A last · Heidrich V first · Fackelmann G · Ricci L · Spadazzi R · Baldanzi G · Punčochář M · Catassi G · Marchi P · Modesto M · Piccinno G · Porcari S · Rondinella D · Asnicar F · Valles-Colomer M · Mattarelli P · Ianiro G · Segata N last

PRJEB17632.csv.gz (9.5 KB) · parquet (84.2 KB) — all 283 samples, every field with route and confidence · PRJEB17632_determinations.csv.gz (31.1 KB) — 3,372 evidence rows