mgnify

name	mgnify
description	Use EMBL-EBI MGnify for analysed microbiome studies and biome genome catalogues (including UHGG / human-gut and other biomes). Use this skill when: (1) Browsing or citing MGnify analyses, (2) Downloading UHGG or other MGnify Genomes catalogues, (3) Searching genes/MAGs against catalogues, (4) Building comparison sets for MAG novelty. Upstream: https://www.ebi.ac.uk/metagenomics/ and https://github.com/EBI-Metagenomics. For raw reads use sra-ena. For taxonomy releases use gtdb. For routing use public-databases.
license	MIT
category	utilities
tags	["database","MGnify","UHGG","genome-catalogue","EBI"]
upstream	https://github.com/EBI-Metagenomics

MGnify (incl. UHGG / biome catalogues)

Upstream: https://www.ebi.ac.uk/metagenomics/ · Docs: https://docs.mgnify.org/ · API: https://www.ebi.ac.uk/metagenomics/api/v1/ · Genomes FTP: https://ftp.ebi.ac.uk/pub/databases/metagenomics/mgnify_genomes/ · Org: EBI-Metagenomics

Citation

Richardson, L. et al. MGnify: the microbiome sequence data analysis resource in 2023. Nucleic Acids Res. 51, D753–D759 (2023). https://doi.org/10.1093/nar/gkac1080

Almeida, A. et al. A unified catalog of 204,938 reference genomes from the human gut microbiome. Nat. Biotechnol. 39, 105–114 (2021). https://doi.org/10.1038/s41587-020-0603-3

See also docs/references.md.

Analytical thinking

MGnify is not a replacement for your pipeline. It is:

1. An archive of analysed microbiome runs (taxonomy/function under MGnify’s pipeline versions)
2. A set of biome-specific genome catalogues (species representatives + protein catalogues), e.g. human-gut / UHGG

Use catalogues to contextualize your MAGs (novelty, species clusters, protein families). Do not treat “found in UHGG” as proof of presence in your sample without mapping/abundance evidence (coverm, profilers).

Catalogue QC/dereplication rules (CheckM thresholds, dRep ANI) are theirs — state both their criteria and yours (mag-qc) when comparing.

Catalogue family (thinking)

MGnify Genomes publishes multiple biome catalogues (human-gut / UHGG is the most cited; others cover e.g. oceans, various hosts/environments — check the live FTP/API list). Each catalogue version has its own:

• species-representative set + all-genome metadata
• protein catalogues at several AA identity cutoffs
• sometimes prebuilt search indexes (Kraken2, etc.)

Cross-catalogue rules:

Claim	Do
Novelty in human gut	Compare to current UHGG / human-gut catalogue version
Non-gut biome	Use the matching biome catalogue, not UHGG alone
Global taxonomy	Still run gtdbtk on your MAGs; catalogue labels are secondary
Presence in my sample	Map/profile (coverm, MetaPhlAn, …) — catalogue hit ≠ local detection

MGnify analyses of public studies use MGnify pipeline versions — their taxonomy/function tables are not interchangeable with your MetaPhlAn/HUMAnN run unless you intentionally reanalyse the same ENA runs.

Prerequisites

Need	Resource
Browse studies	MGnify web / API
Download MAG catalogue	FTP under mgnify_genomes/<biome>/<version>/
Gene fragment search	MGnify COBS gene search
Whole-MAG similarity	MGnify Sourmash MAG search (-k 31 --scaled 1000)

How to browse / API

# List human-gut catalogue metadata (example)
curl -s "https://www.ebi.ac.uk/metagenomics/api/v1/genome-catalogues/human-gut-v2-0-2" \
  | python -m json.tool | head

Prefer the live catalogue ID/version from the website or API — do not hardcode obsolete versions into Methods without checking.

How to download genome catalogues (FTP)

# Example layout (pin the version you actually use):
# https://ftp.ebi.ac.uk/pub/databases/metagenomics/mgnify_genomes/human-gut/<version>/

mkdir -p db/mgnify/human-gut && cd db/mgnify/human-gut
wget -c https://ftp.ebi.ac.uk/pub/databases/metagenomics/mgnify_genomes/human-gut/v2.0.2/genomes-all_metadata.tsv
# species_catalogue/, protein_catalogue/, optional kraken2_db_* when published

Typical contents: species representatives, all-genomes metadata, protein catalogues (100/95/90/50% AA), sometimes prebuilt Kraken2 indexes.

Key thinking for MAG studies

Task	Approach
Is my MAG a known gut species?	Sourmash/Mash vs UHGG reps; or GTDB + literature
Cross-biome claim	Search the matching MGnify biome catalogue, not only human-gut
Functional comparison	Use protein catalogues carefully — AAI cutoffs change interpretation
Cite correctly	Catalogue name + version + FTP path + download date

Decision tree

MGnify vs other public resources?
├─ Need raw FASTQ → sra-ena (MGnify often links ENA accessions)
├─ Need biome MAG / protein catalogue → mgnify
├─ Need global prokaryotic taxonomy release → gtdb
└─ Need marker/pathway indexes for profiling → biobakery-databases

Related skills

public-databases · sra-ena · gtdb · progenomes · cami · drep · coverm · gtdbtk · mag-qc · visualization