Inputs & Parameters
===================

vConTACT3 has numerous options, but really only requires either a single nucleotide file in FASTA format, or proteins
and a gene-to-genome mapping if using pre-called genes. All other options have defaults, which will be used when not specified.

Inputs
------

**--nucleotide**
Path to a FASTA-formatted nucleotide file. Selecting this option will enable the gene-calling tool and disable `--proteins`.

.. note::
   Avoid using ``mmseq`` in the input filename (e.g. ``my-mmseqs-results.fna``), as the temporary file cleanup may
   inadvertently match temporary files. See :doc:`FAQ & Troubleshooting <troubleshooting>` for details.

**--proteins**
FASTA file of predicted proteins. Requires ``--gene2genome``, while ``--len-nucleotide`` is optional (see below).

**--gene2genome**
TSV or parquet file linking protein IDs to genome IDs. Required when using ``--proteins``. Expected columns:
``protein_id``, ``genome_id``, and optionally ``keywords`` (filled with ``None`` if absent).

**--len-nucleotide**
TSV or parquet file mapping genome IDs to nucleotide lengths. Only applicable when using ``--proteins``. When
``--nucleotide`` is provided, genome lengths are computed automatically from the sequences. Optional even in protein
mode. If omitted, ``Size (Kb)`` will be ``NaN`` in the output and the ``ANI`` export will be disabled. Accepts a
``length`` column in base pairs (converted to KB automatically) or a ``Size (Kb)`` column.

**--output**
Path to the output directory. Defaults to `vConTACT3_results/`.

Key Parameters
--------------

Though not necessary, these are the most frequently used parameters

**--threads**
Number of CPU cores to use. Defaults to all available cores.

**--max-iterations**
Iterations to use when resolving mixed-realm components/clusters. Increase to reduce chance of encountering. Default: 3

**--reduce-memory**
Reduce memory usage by downcasting arrays to `float16` (~50% savings).

**--distance-metric**
The distance metric used between genomes in the gene sharing network. Options: `SqRoot` (default), `VirClust`, `Shorter`, `Jaccard`.

**--breaks**
Splits large networks/graphs into smaller chunks during export.

**--db-path**
Path to a specific database version file or directory. Defaults to using the latest version.

**--db-domain**
Specify domain: `archaea`, `bacteria`, `prokaryotes`, or `eukaryotes`.

**---exports**
Specify which export types to generate (e.g., `graphml`, `cytoscape`, `profiles`). See :doc:`Exports <exports>` for details.

Advanced Parameters
-------------------

**--verbose**
Increase logging verbosity (INFO, WARN, ERROR, DEBUG).

**--keep-temp**
Preserve intermediate files (generally for MMSeqs2).


For a full list of command-line options, see the :doc:`CLI Reference <cli/run>`.