cdm_bakta_proteins

CTS (CDM Task Service) job wrapper for Bakta's protein-input mode, which annotates pre-existing protein sequences without re-predicting genes. Variant of cdm_bakta for the protein-input mode.

When to use this vs cdm_bakta

Tool	Input	What it does	Output locus tags
cdm_bakta	nucleotide assembly (.fna)	predicts genes + annotates	bakta-generated (e.g., AHLLAC_0001)
cdm_bakta_proteins (this)	protein FASTA (.faa)	annotates only, no gene-calling	preserved from input

Use this when the input proteins already have meaningful locus tags (NCBI RefSeq, prior annotation, pre-existing pipeline output, etc.) and you want bakta's annotation without losing those identifiers.

Container

Wraps the official oschwengers/bakta image (v1.12.0), invoking the bakta_proteins binary instead of bakta. Also overlays diamond v2.2.0 (same as cdm_bakta:0.1.3) to avoid the intermittent pseudogene-detection deadlock that affects the diamond v2.1.21 shipped in the conda bakta package. See oschwengers/bakta#424.

Published to ghcr.io/kbaseincubator/cdm_bakta_proteins.

Entrypoint: bakta_proteins. Append flags as arguments.

Reference data: Same Bakta DB as cdm_bakta. CTS mounts the bundle at /ref_data/db/. The container reads it via BAKTA_DB=/ref_data/db baked into the image, so callers do NOT need to pass --db on the command line.

Usage via CTS

job = tscli.submit_job(
    "ghcr.io/kbaseincubator/cdm_bakta_proteins:0.1.0@sha256:<digest>",
    input_files,                         # protein FASTA (.faa or .faa.gz)
    "cts/io/<user>/output/bakta_proteins/run1",
    cluster="kbase",
    declobber=True,
    output_mount_point="/out",
    args=[
        "--output", "/out",
        "--threads", "4",
        "--force",                # bakta refuses to overwrite the pre-created /out without this
        tscli.insert_files(),
    ],
    num_containers=len(input_files),
    cpus=4,
    memory="16GB",
    runtime="PT1H",
)

Output

Per input protein FASTA, bakta_proteins produces:

• <prefix>.tsv: per-locus annotation table (Sequence Id, Type, Start, Stop, Strand, Locus Tag, Gene, Product, DbXrefs). Imported to Delta Lake.
• <prefix>.gff3, .gbff, .embl: standardized sequence formats
• <prefix>.faa, .hypotheticals.faa: annotated proteins
• <prefix>.json: full machine-readable annotation data
• <prefix>.txt: summary
• <prefix>.log: bakta run log