Commands

The supported command-line workflows and their important options. The installed --help output is the exact option reference for your version.

giftag build

giftag build [-d DIR] [--gifter-db PATH|URL] [--sources kofam,ncbifam,pfam,dbcan]

Builds the local profile database from gifter’s marker list. With no --gifter-db, giftag reads the database at gifter’s current GitHub main; supply a local SQLite database or marker TSV to pin the marker catalogue. The default build includes all four supported sources.

Option Effect
-d, --db Database directory; default is $GIFTAG_DB, then ~/.local/share/giftag.
--gifter-db Gifter SQLite database or marker TSV, local path or URL.
--sources Comma-separated sources to build. Omitted sources are marked source_skipped in markers.tsv.
--kofam-release, --ncbifam-release, --dbcan-release Choose source releases. KOfam accepts current or an archive date.
--kofam-dir, --ncbifam-dir, --pfam-dir, --dbcan-dir Use local copies of the source files.
-t, --threads Parallel downloads; default 8.
--force Replace an existing built database at the destination.

The command downloads profiles from their publishers and redistributes none. Inspect source terms and coverage before choosing sources.

giftag annotate

giftag annotate -i FASTA... -o DIR [-d DIR] [-t THREADS]
giftag -i FASTA... -o DIR  # shorthand

Searches nucleotide or protein FASTA files, plain or gzipped, and writes gifter-ready marker rows. -i accepts multiple files or a directory. Input types are detected per file unless --input-type is given.

Option Effect
-i, --input One or more FASTA paths, or directories containing FASTA files.
-o, --outdir Directory for output tables, run manifest and predicted proteins.
-d, --db Built giftag database; uses the same default as build.
-t, --threads Search CPU threads; 0 uses all available threads.
--input-type Set auto, nucleotide or protein instead of per-file detection.
--mode Set auto, single or meta for nucleotide gene calling. In auto, single-genome training is used from 100 kb upward.
--gate-subfamilies Search dbCAN-sub clusters only on proteins with a matching family domain. Faster, but stricter than the default run_dbcan-like search.

For --gate-subfamilies, the validation comparison found 144 of 157 retained subfamily calls relative to the ungated search. Keep the default when reproducing run_dbcan-like evidence matters more than speed.

Inspect a build

giftag info
giftag info -d /path/to/giftag-db
giftag version

info reports source releases, marker coverage and terms of use. version reports the installed giftag package version.

Logs

Global logging options go before the command:

giftag --log-file run.log annotate -i genomes/ -o annotations/
giftag --quiet annotate -i genomes/ -o annotations/
giftag --verbose build

--log-file writes UTC-stamped plain-text progress. --quiet keeps warnings and errors; --verbose shows debug messages. Progress bars become plain log lines when output is piped or sent to a job log.