Eukaryotic taxonomy/completeness metadata sources (GTDB/CheckM/GTDB-Tk equivalents)

Description of feature

magmap's species_preference logic (and general genome metadata reporting) depends on GTDB metadata (taxonomy + completeness/contamination for remote genomes), GTDB-Tk (taxonomy classification for local genomes), and CheckM/CheckM2 (completeness/contamination for local genomes) -- all prokaryote-specific. If magmap gains eukaryotic transcriptome support (#239), the equivalent metadata sources for eukaryotes need identifying, since GTDB itself doesn't cover eukaryotes at all.

What's available (checked directly, not assumed)

Taxonomy: MMETSP entries are mostly pre-identified, cultured strains -- taxonomy comes as metadata from culture collection records, not something needing a GTDB-Tk-style algorithmic classifier for the reference side. Where a curated eukaryotic taxonomy backbone would matter is classifying an unidentified sequence (e.g. a user's own uncultured eukaryotic MAG/transcriptome) -- PR2 (Protist Ribosomal Reference database) is the standard for marine microbial eukaryotes; MarFERReT itself uses PR2 taxonomy IDs alongside NCBI taxonomy.

Completeness/contamination (the CheckM analog): EukCC is explicitly built as CheckM's eukaryotic equivalent -- same completeness/contamination framing, works on eukaryotic genome/MAG assemblies, and reportedly outperforms BUSCO specifically on contamination estimation. Separately, and specific to the MMETSP path: the 2019 re-assembly paper already published per-sample BUSCO completeness metrics as supplementary data (MMETSP_all_evaluation_matrix.csv) -- potentially reusable directly as ready-made quality metadata for MMETSP entries without re-running anything.

Open questions

• Does EukCC work sensibly on transcriptome assemblies (not just genome assemblies), given MMETSP entries are transcripts not genomes? CheckM's own genome-focus suggests EukCC may share that assumption -- worth checking before assuming direct applicability to transcript-level references.
• Would species_preference-equivalent logic for eukaryotic entries reuse the existing local/completeness/gtdb preference modes, or need its own scheme given the different metadata shape (MMETSP's own BUSCO/Transrate scores vs. CheckM/GTDB-style completeness+contamination)?

Related: #239 (MMETSP support, which this metadata question feeds into), #240 (eukaryotic gene calling).

Start by reviewing the MMETSP support proposal in issue #239 and the cited EukCC, PR2, and MMETSP_all_evaluation_matrix.csv sources. Check whether EukCC applies to transcriptome assemblies, then compare available taxonomy and quality metadata with the existing species_preference modes. Done means documenting suitable eukaryotic sources and recommending how their metadata should be represented.