Database composition¶

Database version 1.0.2 is archived at Zenodo. It uses SILVA 138.2 and PR2 5.1.1 in the curated profile. The img profile contains the same references plus IMG 16S rRNA gene and 18S rRNA gene sequences from eukcensus 2025.

Profiles¶

Profile	Source records	Unique sequences	16S rRNA gene index	18S rRNA gene index	Download size
`curated`	683,597	609,298	416,021	193,282	345.4 MiB
`img`	1,680,777	1,396,949	1,067,340	329,614	840.9 MiB

Five exact sequences belong to both rRNA gene indexes. Each profile stores an exact sequence once and records both marker memberships.

IMG-derived records keep the conservative member assignment separate from the centroid used to classify the sequence cluster. The database stores centroid name, calibrated taxonomy, and taxonomy source as evidence fields.

Source releases¶

Source	Release	Content	`curated` records	`img` records
SILVA	138.2 NR99	16S rRNA gene references and bacterial/archaeal taxonomy	451,555	451,555
PR2	5.1.1	18S rRNA gene and organellar 16S rRNA gene references; eukaryotic taxonomy	232,042	232,042
IMG	eukcensus 2025	Additional 16S rRNA gene and 18S rRNA gene records	0	997,180

Exact duplicate sequences are represented once in a profile. Of the 997,180 IMG source records, 787,651 add a unique sequence to the img profile.

Taxonomy¶

SILVA supplies bacterial and archaeal taxonomy. PR2 supplies eukaryotic taxonomy. Exact sequences with equally preferred assignments from different domains retain both alternatives and are reported as ambiguous.

Preferred domain assignment	`curated` sequences	`img` sequences
Bacteria	388,861	857,491
Archaea	20,182	43,767
Eukaryota	198,605	288,908
Unclassified	0	205,133
Ambiguous	1,650	1,650

The target ranks are phylum for Bacteria and Archaea and PR2 supergroup for Eukaryota. In database version 1.0.2, IMG sequences with a preferred domain assignment stop at domain; 205,133 sequences remain unclassified at domain. The target-rank resolution analysis includes domain-only sequences in the relevant domain denominator and excludes unclassified and ambiguous assignments. Eukaryota_X is retained as PR2's unassigned supergroup placeholder.

Figure 1. Sequences resolved to bacterial or archaeal phylum and eukaryotic PR2 supergroup. Percentages are calculated within each domain.

PR2 compartment suffixes (:apic, :chro, :chrom, :mito, :nucl, and :plas) are removed from taxonomy ranks during import. Compartment is stored in a separate field.

Figure 2. Reported lineage distributions in the SILVA and PR2 reference component. Percentages use sequences resolved to the reported rank as the denominator. Less abundant bacterial and archaeal lineages are combined as “Other named.” The eukaryotic panel shows the seven most abundant named supergroups plus Eukaryota_X, PR2's unassigned supergroup placeholder; complete counts are available below.

IMG sampling coordinates¶

The IMG metadata retains the IMG Taxon OID and latitude/longitude when both coordinates are available. Project names, contact details, comments, and other source metadata fields are excluded.

Coordinate measure	Count
IMG Taxon OIDs	32,521
IMG Taxon OIDs with coordinates	30,349
Distinct coordinate pairs	4,856
Source records linked to coordinates	931,810 of 997,180
Coordinate pairs with 16S rRNA gene and 18S rRNA gene records	3,445
Coordinate pairs with only 16S rRNA gene records	1,336
Coordinate pairs with only 18S rRNA gene records	75

Figure 3. Available IMG sampling coordinates. Point diameter scales logarithmically with the number of taxa at an exact coordinate. Coordinates describe the profile metadata; repeated taxa or studies at one coordinate do not measure environmental abundance or sampling effort. Select the map to open the interactive Plotly version.

Data and provenance¶

The executed analysis is in notebooks/database_composition.ipynb.