A biosynthetic gene cluster (BGC) catalog — a genome-derived inventory of contiguous gene sets predicted to make a secondary metabolite — can attach a named chemical mechanism to a disease-associated species when that species’ clusters match MIBiG, a curated reference of experimentally characterized clusters; where they do not match, MIBiG-themed analysis simply does not capture the species’ mechanism 1. In the Metagenome-Prioritized Phage Cocktails for Crohn’s Disease and IBD project the resulting asymmetry is stark: the Elmassry 2025 BGC catalog (10,060 BGCs across 6,221 species-annotated entries) provides direct genomic-content evidence that escherichia-coli alone among the 6 actionable Tier-A core species carries the canonical adherent-invasive E. coli (AIEC) virulence biosynthetic gene-cluster repertoire — 19 Yersiniabactin BGCs, 16 Enterobactin BGCs, 19 siderophore-class BGCs, 8 Colibactin BGCs, 15 Microcin B17 BGCs 1.
The mechanistic reading of those E. coli clusters is borrowed rather than measured in this project: Yersiniabactin is described as the iron-capture siderophore AIEC LF82 uses to survive in macrophage phagolysosomes (Prudent 2021, Dalmasso 2021, Dogan 2014), Colibactin as the pks+ genotoxin implicated in AIEC-associated colorectal carcinogenesis (Veziant 2016), and Enterobactin as another AIEC siderophore active in mucosal-associated IBD settings 1. What the catalog itself demonstrates is cluster carriage — an iron-capture and genotoxin capacity — which refines iron acquisition and pathobiont expansion by separating genomic capacity from the pathogenic consequence that the cited external literature supplies 1.
The same borrowing appears for a Tier-A candidate outside the six actionable core species: the NB05 streptococcus salivarius Salivaricin MIBiG hits are lantibiotic bacteriocins whose proimmune activity regulates the oral microbiome (Barbour 2023), a functional description again supplied by cited external work rather than by measurement here, and one the report ties to the rigor-controlled E1 Tier-A rather than to the actionable core set 1.
What the catalog does not explain
The other 5 Tier-A core species — mediterraneibacter gnavus, eggerthella lenta, enterocloster bolteae, hungatella hathewayi, flavonifractor plautii — sit in MIBiG dark matter: they carry significant non-MIBiG-annotated BGC content (notably E. lenta 41 BGCs, M. gnavus 58 BGCs of which 31 are lanthipeptides), and their CB-ORFs (candidate-BGC open reading frames quantified at the metagenomic read level) are Crohn’s-disease-versus-healthy-control (CD-vs-HC) RPKM-enriched (reads per kilobase per million mapped reads), yet their CD-association mechanisms are not currently captured by MIBiG-themed BGC analysis 1. This is the reusable limit the project exposes: for these five species an absent annotated cluster is an absent annotation, not an absent mechanism, which extends the arguments of structural-annotation-gap and organism-dark-compounds into disease-target selection 1.
One of the five has an explicit non-BGC alternative. E. lenta’s CB-ORF CD-DOWN pattern (54 % CD-down vs 0 % CD-up) is reported as mechanistically consistent with Koppel et al. 2018 (PMID 29760174), which established that Eggerthella species’ CD-association is driven by drug-metabolism activity (cardiac glycoside reductase Cgr2 / arginine-2,3-aminomutase) rather than secondary-metabolite biosynthesis — i.e., E. lenta’s CD-association is read as functional via a non-BGC mechanism 1. The direction (CD-down, not CD-up) is measured in this corpus, whereas the attribution to drug metabolism rests on consistency with that external study rather than on a test performed here, so it stands as a hypothesis about E. lenta’s mechanism 1.
Tensions
The read-level direction reported for the five dark-matter species is not internally uniform: the same report states that their CB-ORFs are CD-vs-HC RPKM-enriched at the read level, and that E. lenta’s CB-ORF pattern is CD-DOWN (54 % CD-down vs 0 % CD-up) 1. Both statements come from the single project and are recorded here as they stand rather than averaged: the group-level enrichment statement cannot be applied to E. lenta without contradicting its own per-species direction, and any use of CB-ORF enrichment to rank these five as targets must therefore be checked species by species 1.
Signaling-lipid families as a second mechanism class
A second route away from virulence-cluster reasoning is chemical rather than taxonomic. The “ebf” and “ecf” BGC families produce immunoactive fatty acid amide signaling lipids (similar in structure to host endocannabinoids) and were originally implicated by Elmassry et al. as a CD-defining functional signature 1. Crucially for target selection, the report states that the ebf/ecf signature is a sample-level CD biomarker with no per-species attribution available — suitable for treatment-response monitoring in clinical follow-up but not for cocktail target selection — so this mechanism class connects to metabolite-class-signatures-of-gut-inflammation without nominating any organism 1.
The ebf/ecf line is also where the project re-tests its own catalog source. Elmassry et al. 2025 (Cell Host & Microbe 33(2):243–256, PMID 39837311) is identified as the BGC catalog source and as the origin of the ebf/ecf RPKM CD-vs-HC comparison, which the report states was replicated independently in NB08a §4; no effect size appears in that bibliography entry 1. The replication is nonetheless quantified elsewhere in the report, which gives ebf/ecf BGC families as CD-up across 4 cohorts at p<1e-31 (NB08a §11 Test 4), with meta z = 11.71 (ebf) and z = 11.97 (ecf) on a cohort architecture of 1,349 samples × HMP2-IBDMDB / MetaHIT / LLDEEP / PRISM — a strongly supported sample-level result 1.
Figure
Figure NB08a_bgc_pathobiont_enrichment.png is described as a “3-panel: Tier-A core BGC repertoire (Grouped Class stacked bar) + theme enrichment Fisher OR (iron+genotoxin+NRPS-PKS supported) + ebf/ecf RPKM CD-vs-HC by cohort (boxplot)”, i.e. the Fisher exact-test odds ratio for each BGC theme alongside cohort-stratified ebf/ecf abundance 1.
Open Directions
- Re-run the Tier-A mechanism assignment with a cluster-class-agnostic method (e.g. clustering the 41 E. lenta and 58 M. gnavus BGCs against each other and against non-gut genomes) to ask whether the MIBiG dark matter is shared novel chemistry or species-private, closing the gap that read-level CB-ORF enrichment currently leaves mechanistically unlabelled 1.
- Test the lanthipeptide route for M. gnavus directly, pairing presence of its 31 lanthipeptide BGCs with per-cohort CB-ORF RPKM rather than with MIBiG theme membership, which would say whether that block is CD-associated chemistry or incidental genome content 1.
- Pair the E. lenta CB-ORF direction (54 % CD-down vs 0 % CD-up) with read-level quantification of its drug-metabolism loci (Cgr2, arginine-2,3-aminomutase) in the same cohorts, turning a literature-consistency argument into a within-corpus test of whether its CD-association is non-biosynthetic 1.
- Attempt per-species attribution of the ebf/ecf signature — assigning the CD-up reads to contributing genomes rather than to samples — since the absence of that attribution is what currently bars the project’s largest-effect BGC signal (p<1e-31 across 4 cohorts) from informing cocktail target selection 1.
- Record, for each Tier-A target, whether its mechanism is measured in this data or imported from external literature, since Yersiniabactin, Colibactin, Enterobactin and Salivaricin function all rest on cited external work while only cluster carriage and read-level direction are measured here 1.