Download PhaRBP
Complete database releases, updated quarterly. Each release is a frozen snapshot: files never change after publication, so cite the release you used.
Release dev-full
latest 2026-10-15Contents and row counts
CDSs | 154064 |
domain_categories | 11 |
domain_interpro_clusters | 1277 |
domains | 4541 |
ec | 0 |
evidence_types | 8 |
experimental_validation | 9306 |
gene_ontologies | 20131 |
genome_features | 4017394 |
genomes | 40244 |
hosts | 1596 |
link_domain_interpro_cluster | 1300 |
link_ec_protein | 0 |
link_phage_hosts | 21249 |
link_protein_genome | 211520 |
link_protein_sequence | 199682 |
link_sequence_domains | 169425 |
literature | 35377 |
novel_domain_annotations | 2068 |
phages | 25784 |
protein_evidence | 154684 |
protein_sequences | 82193 |
proteins | 199681 |
sequence_symmetry | 81652 |
spike_segmentation | 24803 |
structure | 83555 |
structure files (.pdb) | 80609 |
Getting started
The tables archive is enough for most analyses: start with
proteins_flat.tsv (one row per protein) and sequences.fasta.
Every table is also included individually and as a MySQL dump. Structures are named
<sequence_id>.pdb, where sequence_id is the SHA-256 of the
amino-acid sequence. Prophage-derived proteins are not included, matching the website.
# always-latest download from a script
curl -s http://pharbp.bionetic.org/api/releases/latest/ \
| python -c "import json,sys; [print(f['url']) for f in json.load(sys.stdin)['files']]"
# check integrity
sha256sum -c <<< "<sha256> pharbp_<version>_tables.zip"
# load the MySQL dump
mysql -e "CREATE DATABASE pharbp CHARACTER SET utf8mb4"
zcat pharbp_<version>_mysql.sql.gz | mysql pharbpFor filtered subsets (by host, domain, evidence, …) use Export or the API.