petasearch

name	petasearch
description	Petasearch — efficient parallelized peta-scale protein database search (Steinegger lab). Use when: (1) Searching enormous protein collections beyond routine MMseqs2 setups, (2) Preindexing large target lists for repeated queries, (3) Selective Smith–Waterman after coarse filters. Upstream: https://github.com/steineggerlab/petasearch. Route via homology-search; compare mmseqs2 / erast.
license	MIT
category	analysis-tools
tags	["Petasearch","protein-search","peta-scale","homology"]
upstream	https://github.com/steineggerlab/petasearch
stage	function

Petasearch

Upstream: steineggerlab/petasearch

Citation

See upstream README / releases for the current preferred citation; software at https://github.com/steineggerlab/petasearch. Cross-link Steinegger-lab search stack in docs/references.md.

Analytical thinking

Petasearch targets very large protein target sets with create-DB / preindex / easy workflows and optional selective SW alignment (block-aligner). Use when warehouse-scale protein search is the bottleneck; for standard MAG catalogues prefer mmseqs2, for embedding retrieval erast.

Requires Rust toolchain to build from source — follow upstream install notes. Verify exact binary names/easy-petasearch flags against the installed tag (README examples evolve).

How to run

# follow upstream: build → create DBs → preindex → search
# easy workflow pattern (confirm on installed version):
#   … easy-petasearch query.fasta targetlist alignments.m8 tmp

Decision tree

Protein search scale?
├─ Routine / NR catalogues → mmseqs2
├─ Peta-scale / custom huge targets → petasearch
├─ Vector embeddings → erast
└─ Structure → foldseek

Related skills

homology-search · mmseqs2 · erast · foldseek · lexicmap · tool-selection