Southern Ocean Reference Gene Catalogs

These data correspond to gene catalogs and matrices obtained from 218 metagenomes sampled during the Antarctic Circumpolar Expedition (ACE). The methodology and results obtained using these data are presented in the study entitled Water mass specific genes dominate the Southern Ocean microbiome, by Faure et al. (2025).

Here is a description of each catalog/dataset :

• ACE_Unigenes_catalog.fa.gz: Catalog of all unigenes obtained from the 218 ACE metagenomes, i.e. ORFs clustered at 95% similarity and 90% coverage thresholds. Includes ORFs detected by Prodigal in the eukaryote-dominated >3µm size fraction. In FASTA format.
• SO-RGC.fa.gz: Catalog of unigenes that included at least one ORF from the 0.2-3µm size fraction, i.e. limited to the bacterial free-living size fraction used to build Tara Oceans and Polar Circle Reference Gene Catalogs (OM-RGC). In FASTA format.
• Polar_Genes.fa.gz: Catalog of ACE unigenes that were only detected in polar samples from Tara Oceans and Polar Circle (see Faure et al. for detailed methods). In FASTA format.
• Annotation_Table_AGN_CDH_Tax_KEGG_EGG.tsv.gz: Annotation table of all ACE ORFs (tab-separated format, one line = one non-dereplicated ORF), including (in column order):
  • ORF ID
  • AGNOSTOS cluster ID
  • AGNOSTOS cluster representative ID
  • AGNOSTOS cluster size
  • AGNOSTOS cluster category
  • If AGNOSTOS singleton, singleton category
  • Unigene cluster representative ID
  • Predicted domain (Prokaryote / Eukaryote)
  • KEGG KO
  • EggNOG seed ortholog
  • EggNOG OGs
  • EggNOG narrow OG name
  • EggNOG narrow OG category
  • EggNOG best OG name
  • EggNOG best OG category
  • EggNOG best OG description
  • EggNOG preferred name
  • EggNOG CAZy
  • EggNOG BiGG Reaction
  • EggNOG PFAMs
• AGNOSTOS_CLSTRLVL_GENE_MAT_COV.tsv.gz: AGNOSTOS cluster-level matrix of coverage. Columns correspond to samples and lines to AGNOSTOS clusters, with ID matching those of the Annotation table. Cell values correspond to ORF-level DESeq2-normalized per-base pair coverage values, summed by AGNOSTOS cluster.
• AGNOSTOS_CLSTRLVL_GENE_MAT_DET_Maximum.tsv.gz: AGNOSTOS cluster-level matrix of Detection. Detection is calculated at ORF-level after mapping: proportion of the ORF covered at at least 1X. In this matrix, the maximum detection value observed for each AGNOSTOS cluster in each sample is given.
• EnrichmentPolar_CDHitLevel_EggNogDesc.tsv.gz: Functional enrichment test outputs for polar unigenes (i.e. genes present in Polar_Genes.fa.gz) versus the rest of ACE unigenes.
• Contig_GTDBlineage_kraken.tsv.gz: Taxonomic annotation of ACE contigs based on the Genome Taxonomy DataBase (GTDB) using Kraken2, please refer to the study for more details on the methodology.
• ACEsamples_CorrespondanceTable.xlsx: Table of correspondance linking the different IDs given to ACE samples accross different steps, including BioSamples and ENA run and experiment codes.
• RF_AGC_NZVuniquecut20_T60MAX_ATT_All.txt.gz: Random forest models outputs for the large size fraction. Presents for each AGC the model R-squared, mean squared error and predictors' importance.
• RF_AGC_NZVuniquecut20_T60MAX_FL_All.txt.gz: Random forest models outputs for the small size fraction. Presents for each AGC the model R-squared, mean squared error and predictors' importance.
• DMSP_Lyase.tar.gz: Folder containing the matrix of abundance of the DMSP lyase AGCs, the corresponding ORF-level annotation table and random forest outputs.
• contigs_virus_genomadannot.tsv.gz: Annotation of ACE viral contigs (geNomad outputs).
• contigs_mimiviridae.xlsx: More detailed annotation of mimivirus contigs (please refer to the study for more details on the methodology).

Please note that all CAG-related files are available on Figshare at DOI 10.6084/m9.figshare.29821949.