UniProt-Formatted, Reannotated Metagenome-Derived Protein FASTA Files from Human and Other Vertebrate Microbiome Catalogs

This repository provides reformatted protein FASTA files derived from CNGB and MGnify/MetaPilot metagenomic catalogs obtained from human- and other vertebrate-associated microbiomes. To support downstream processing by bioinformatics pipelines, and in particular metaproteomic peptide identification from tandem mass spectrometry (MS/MS) data, all sequence headers were systematically reformatted into a standardized, UniProt-like syntax that embeds the original taxonomic and functional annotations directly within the header. 1. CNGB Catalogs For the CNGB microbiome protein catalogs (https://db.cngb.org/microbiome), the human gut Integrated Gene Catalogs (11M and 9M) were merged into a single database. To remove sequence redundancy, exact duplicate protein sequences were collapsed at 100% amino acid sequence identity using SeqKit. KEGG functional annotations available alongside the catalog were incorporated by converting the multi-accession TWINS.KEGG.catalog matrix into a mapping of gene IDs to KEGG Orthology (KO) terms. The KO terms were then converted into descriptive protein names using a reference KO annotation repository (v. 2026-02-01). The accompanying taxonomic annotations were not incorporated as, in most cases, their taxonomic nomenclature is now outdated. The same annotation, deduplication, and formatting workflows were applied to all CNGB microbiome catalogs for which functional annotations were available, including the mouse, pig, and rat gut microbiomes. The organism source descriptor (OS=) was customized in the UniProt-like headers to reflect the corresponding metagenomic source. 2. MGnify Catalogs The MGnify/MetaPilot microbiome protein catalogs consist of MGnify-derived collections of representative species-level metagenome-assembled genomes (MAGs), originally distributed through the Resources directory of the MetaPilot software repository (https://github.com/cksakura/MetaPilot-Releases). These catalogs were processed to integrate the taxonomic classification of each source MAG directly into the FASTA sequence headers while preserving the original functional protein descriptions. This release includes the following nine host-associated catalogs: Human-associated: Gut (4,744 species), Oral (452 species), Skin (579 species), Vaginal (280 species) Other vertebrate-associated: Mouse gut (2,847 species), Pig gut (1,376 species), Cow rumen (2,729 species), Sheep rumen (2,172 species), Chicken gut (1,322 species). Header Format All FASTA headers strictly follow the standard UniProt-like header format: >db|UniqueIdentifier|EntryName ProteinName OS=OrganismName PE=4 This format allows software tools that natively support UniProt databases to correctly parse and extract both taxonomic and functional metadata during database searching.