MAG / MAG-WGS Support
Implementation Report

In the main branch, project_type was parsed from metadata but used only to set complete=True for gnm. DDBJ submission requirements specific to MAG (Metagenome-Assembled Genome) and MAG-WGS — ENV division, MAG keywords, isolate qualifier, metagenome_source, environmental_sample — were not supported.

This implementation adds the following new features relative to the main branch.

config_util.py changes

genome.py changes

MAG/MAG-WGS submissions go to DDBJ's ENV division. The COMMON entry in the .ann file must include a dedicated block of division and keyword rows.

The internal data_file_division field (used in GenBank/EMBL output) also switches from BCT to ENV via genome.py add_source_information().

The main branch had no isolate support anywhere. Since MAG submissions use isolate instead of strain as the specimen identifier, it was added throughout the pipeline.

Propagation path

SOURCE line in source feature

In the main branch, MetadataField had no concept of project_type, and EX_SOURCE rendering used the simple condition field.value or field.mss_required. This implementation replaces that with two new TSV columns and four new methods.

New and updated rows in metadata_definition.tsv

New methods on MetadataField

def is_required_for(self, project_type: str) -> bool:
    return project_type in self.required_for

def is_excluded_for(self, project_type: str) -> bool:
    return project_type in self.excluded_for

def effective_required(self, project_type: str) -> bool:
    if project_type in self.excluded_for: return False
    if project_type in self.required_for:  return True
    return self.mss_required  # project_type-independent default

def should_render(self, project_type: str) -> bool:
    if project_type in self.excluded_for: return False  # excluded even if value is set
    if self.value:                          return True   # value present → always render
    return self.effective_required(project_type)         # render placeholder if required

render_ex_source() changes

get_field_definition() — new module-level function

Used by genome.py to look up TSV field definitions when building source feature qualifiers. Parses the TSV once and caches the result.

# module level
_FIELD_DEF_CACHE: "dict | None" = None

def get_field_definition(name: str):
    global _FIELD_DEF_CACHE
    if _FIELD_DEF_CACHE is None:
        _FIELD_DEF_CACHE = {}
        # parse TSV once, populate cache
        ...
    return _FIELD_DEF_CACHE.get(name)

In the main branch, ff_definition always used @@[strain]@@. For MAG/MAG-WGS, where strain is excluded, it now switches to @@[isolate]@@.

The main branch always used the biosample_strain pattern for output file names. Since MAG does not use strain, the identifier is now chosen based on is_mag.

In the main branch, the boolean render path emitted a valueless qualifier only when the value was explicitly set (anything other than "NO"). With an empty value it returned [], making it impossible to emit environmental_sample as a placeholder for MAG submissions.

Test categories