We produced protein cluster catalogues in order to establish the gene-level taxonomic overlap between the human and mouse microbiome. For gene-clustering, we concatenated 76,937,350 pre-clustered human predicted proteins for non-redundant, near-complete genomes of the UHGG with 45,598,646 mouse predicted protein-coding sequences from non-redundant, near-complete species of the MMGC, and performed protein clustering using the ‘linclust’ function from MMseqs285 v10-6d92c (-c 0.8 --cov-mode 1 --cluster-mode 2 --kmer-per-seq 80). Proteins were clustered at 100%, 90%, 80% and 50% sequence identity; clusters were considered shared if they contained genes from both human and mouse commensals. This archive includes two files for each sequence identity threshold: a representative protein sequence file (.fa) and a sequence membership file (.tsv).