Prevotella copri (abbreviated P. copri in the source reports) is a human gut bacterial species whose relative abundance names one of four reference gut microbiome ecotypes (K = 4) — community states used as patient strata — derived from stool metagenomes: E2, the Prevotella copri enterotype, holds 920 samples, with the defining species reported as mean abundance per ecotype (P. copri 28 %, F. prausnitzii 6 %) 1.
Health association
The central discoveries digest tabulates the same stratum as “E2 — Prevotella copri enterotype | 920 | P. copri 28 % | 16.9 % of HC, ~0 % disease (non-Western healthy)”, i.e. 16.9 % of healthy controls (HC) and approximately none of the disease samples fall in this ecotype 2. The project report restates this reading directly, describing E2 as the P. copri enterotype, “almost entirely non-Western healthy”, and pairs it with E0 (diverse commensal) at 66.8 % of healthy controls as the two healthy-cohort ecotypes 1. These are prevalence counts over a reference clustering of cohort samples, so the healthy skew of E2 is an observation about the composition of the clustered cohorts and not, in the reported analysis, a test of whether P. copri carriage is protective 2,1.
Position among the reference ecotypes
Defining-species abundances (means per ecotype) and diagnosis patterns of the four consensus ecotypes, with E2 the only Prevotella-defined stratum 1:
| Ecotype | Samples | Defining species (mean abundance per ecotype) | Diagnosis pattern 1 | | --- | --- | --- | --- | | E0 — Diverse commensal | 3,604 | F. prausnitzii 6.8 %, R. bromii 4.5 %, B. uniformis 4.6 %, P. vulgatus 4.4 % | 66.8 % of HC | | E1 — Bacteroides2 transitional | 2,601 | P. vulgatus 9.8 %, B. uniformis 7.2 %, Phocaeicola dorei 3.5 % | 48 % CD, 58 % UC, 100 % T1D, 97 % T2D, 67 % nonIBD | | E2 — Prevotella copri enterotype | 920 | P. copri 28 %, F. prausnitzii 6 % | 16.9 % HC, ~0 % disease (non-Western healthy) | | E3 — Severe Bacteroides-expanded | 1,364 | P. vulgatus 14.2 %, B. fragilis 3.6 % | 50 % CD, 40 % UC, 67 % IBD acute, 38 % CDI, donor 2708 |
Absence in the UC Davis validation cohort
All 26 Kuehl_WGS samples (23 unique patients) were projected onto the K = 4 reference through a species-name synonymy layer, which normalized 262 unique kaiju-classified species to 97 canonical species in the training feature space; the UC Davis cohort distributed as E0 — diverse commensal: 7 samples (27 %), E1 — Bacteroides2 transitional: 11 samples (42 %), E2 — Prevotella copri enterotype: 0 samples, and E3 — severe Bacteroides-expanded: 8 samples (31 %) 1.
That E2 drew zero samples is consistent with the characterization of the P. copri enterotype as almost entirely non-Western healthy 1, but the observation rests on a single clinical cohort of 26 samples from 23 patients and on a feature space reduced from 262 classified species to 97 canonical ones, so it is best read as an expectation-matching null for this one cohort rather than as a test of P. copri geography 1.
The caveat the report itself documents on this projection is a classifier mismatch, not a Prevotella-specific one: the Kuehl samples were classified with kaiju while the reference embedding was trained on metaphlan3 profiles, and the two ecotype methods differ in robustness to that mismatch — LDA, the latent-factor topic model fitted on species pseudo-counts (sklearn LatentDirichletAllocation, scored by training perplexity) 2, is robust, handling the 54 % of Kuehl feature rows outside the training feature space by treating absence as not-detected, whereas GMM (Gaussian mixture model) on CLR (centered log-ratio) + PCA is fragile, the same sparsity forcing all 26 Kuehl samples into a single Gaussian (E3) at confidence > 0.97, which the report calls an artifact, not biology; LDA is therefore the primary Kuehl projection call and GMM advisory 1. Any further explanation specific to Prevotella — for example that name normalization selectively erased Prevotella signal — is an untested hypothesis that the reported projection does not evaluate 1.
Use as a proxy marker
P. copri is one of the species named for a proposed targeted qPCR (quantitative PCR) ecotype panel of 4–6 species — F. prausnitzii, P. copri, P. vulgatus, B. fragilis, mediterraneibacter gnavus (M. gnavus) — intended to assign ecotype without full metagenomic sequencing; the report presents it as a panel design, with no assay validation reported 1.