What it is
Canonical name: NMDC ARKin. 1
Known alias: nmdc_arkin. 1
Stable external identifier: No stable external identifier is specified in the source. 1
NMDC ARKin is a dataset layer in the KBase Data Lakehouse inventory providing environmental abundance, multi-omics, taxonomy, and biosample records for cross-tenant biological analyses. 1 The NMDC Context Audit refines this description by identifying kbase.nmdc_arkin as an Arkin Lab derivative located in the kbase tenant, rather than a native resource in the nmdc tenant; the derivative adds embeddings and traits that do not exist upstream. 2
The audit also refines the provenance context: NMDC ARKin is one of 7 maintained resources represented by 20 nmdc-named database entries across three tenants and six provenance classes, alongside genuine NMDC resources, external re-hosts, an NMDC-derived MAG resource, and the NEON namesake collision. 2 Consequently, the nmdc label alone is not a reliable dataset boundary or authority indicator. 2
Key facts
The KBase Data Lakehouse inventory contains 16,640 NMDC biosamples and 75,119,498 metatranscriptomic abundance rows. 1 NMDC covers 11 biological topics and has a topic-distribution entropy of 2.61. 1 Its metabolomics, proteomics, and lipidomics layers contain 3.1M, 346K, and 1.4M records, respectively, and are described as largely untapped resources. 1
The atlas identifies NMDC as a cross-validation resource for UC4, environmental distributions of clinically relevant pathogens and associated biogeochemistry, and UC5, ENVO ontology completeness in NMDC biosamples. 1 The proposed UC4 bridge connects NMDC with PROTECT through 10 shared keys, while the proposed UC5 bridge connects NMDC with refdata through 9 shared keys; neither bridge had realized use at audit time. 1 Join-key presence demonstrates schema-level compatibility but not valid value-space overlap, and UC4 and UC5 require live-cluster execution. 1
The Context Audit supports the need for explicit provenance-aware discovery: nmdc_arkin, nmdc_mags, and nmdc_neon are located in the kbase tenant even though users searching only the nmdc tenant would not find them. 2 It further refines scale interpretation by distinguishing the genuine NMDC biosample universe of 16,640 samples from the co-hosted nmdc.ncbi_biosamples mirror containing 51,711,888 biosamples and 756,112,544 attribute rows. 2 These resources should therefore be selected by provenance, authority, scale, and currency rather than by name alone. 2
The community metabolic-ecology integration supports ARKin’s multi-omics role with a 220-sample community pathway matrix linked to 27,690 GTDB species, 80 pathways, 305M GapMind pathway records, and NMDC taxonomy and metabolomics. 3 Taxonomy-bridge coverage averaged 94.6%; 92% of samples mapped at least 85% of community abundance to GTDB pangenome species, and the analysis-ready merged matrix contained 174 samples. 3 This study refines the atlas’s general multi-omics opportunity by showing that 33 Freshwater samples lacked paired metabolomics and that the H1 pathway–metabolomics test therefore had 131 samples. 3
The same analysis supports NMDC as a resource for testing community metabolic ecology: 11 of 13 amino-acid pathways had negative pathway-completeness/metabolite correlations, with a significant sign test (p = 0.011), while leucine (r = −0.390, q = 0.022, n = 62) and arginine (r = −0.297, q = 0.049, n = 80) remained significant after Benjamini-Hochberg false-discovery-rate correction. 3 It also supports ecosystem-level analysis: Soil and Freshwater communities separated strongly in pathway-completeness PCA, with Soil versus Freshwater U = 3,674 and p < 0.0001, and 17 of 18 amino-acid pathways differing by ecosystem type after correction. 3 These results are genomic-potential measurements rather than evidence of pathway expression, and absent pH, temperature, and total organic carbon measurements prevented abiotic partial-correlation controls. 3
The ENIGMA Carbon Census supports NMDC ARKin’s role as an environmental-abundance resource: its atlas used 3825 taxonomy-bearing NMDC metagenomes, detected 83 of 86 implicated utilizer genera in 1719 metagenomes, and achieved 99% sample labeling through two independent ontology systems. 4 That analysis refines interpretation of NMDC abundance data: it measures organismal occurrence or abundance rather than compound catabolic activity because the environmental datasets did not measure the census compounds. 4
The eukaryotic-read analysis supports ARKin’s value for environmental metadata studies: native results from 2,759 ReadbasedAnalysis runs across 9 studies showed detectable eukaryotic reads in 77% of runs, with a median eukaryotic fraction of 2.7% and a mean of 13.3%; GOTTCHA2 was the usable estimator because the deployed Kraken2 and Centrifuge references were prokaryote-restricted. 5 Environment predicted eukaryotic fraction under random validation (R²=0.35) but failed out-of-study GroupKFold validation (R²=−0.30), refining cross-study interpretation by indicating that collection-wide environmental associations were largely confounded with study or batch. 5 Within the dominant NEON soil study, local vegetation and geography remained associated with eukaryotic fraction; a model using both achieved within-study five-fold R²=+0.17 ± 0.06, supporting batch-controlled rather than naive cross-study regression. 5
The Functional Dark Matter analysis supports this batch-controlled environmental interpretation: NMDC independently mapped 5 of 6 carrier genera to 47 taxon columns across 6,365 metagenomic samples and confirmed all 4 testable pre-registered abiotic predictions, including correlations with total nitrogen (ρ = +0.109, n = 1,231, FDR = 2.3e-4), ammonium nitrogen (ρ = +0.231, n = 1,230, FDR = 8.0e-16), pH (ρ = +0.157, n = 4,366, FDR = 7.4e-25), and dissolved oxygen (ρ = -0.298, n = 272, FDR = 1.5e-6). 6 It also supports NMDC’s value for trait validation: all 7 pre-registered trait-condition predictions reached FDR < 10⁻²¹, including nitrogen-source carriers with nitrogen fixation (ρ = 0.60) and nitrate denitrification (ρ = 0.52), and carbon-source carriers with aerobic chemoheterotrophy (ρ = 0.73) and fermentation (ρ = 0.59). 6 However, 441 of 449 exploratory tests reached FDR < 0.05, largely attributed to compositional coupling, so these associations require cautious interpretation and sample-label permutation testing. 6
The eukaryotic-read report refines the proposed cross-tenant bridging role by showing that workflow results must be joined to biosamples and studies through workflow_run_id, with large taxonomy tables aggregated before joining; 1,067 of 2,759 runs were pooled from multiple biosamples, so biosample-level joins could create pseudo-replication. 5
NMDC ARKin contributes to multi-omics-integration because its metabolomics, proteomics, lipidomics, metatranscriptomic, taxonomy, and biosample layers provide potential cross-tenant multi-omics linkages. 1 The community analysis supports this placement by linking community-weighted GapMind pathway completeness to metabolite intensity, while refining it through missing freshwater metabolomics, study imbalance, string-based metabolite matching, and the distinction between genomic potential and expression. 3 The eukaryotic-read analysis refines this opportunity by demonstrating that classifier reference databases are not interchangeable for contamination estimation and that sequencing and environmental metadata require study- or batch-controlled comparisons. 5 The Functional Dark Matter results further support integrating NMDC with fitness, pangenome, and environmental evidence while refining the interpretation: genus-level carrier correlations can validate pre-registered environmental predictions but cannot by themselves establish specific dark-gene functions. 6 The Context Audit supports retaining these derivative products while refining their interpretation: embeddings, traits, metabolomics, omics files, and MAG catalogs have different provenance and authority contexts and should be exposed with those relationships rather than treated as a single homogeneous NMDC resource. 2
It also contributes to environmental-resistome through the proposed NMDC–PROTECT bridge for examining environmental distributions of clinically relevant pathogens and associated biogeochemistry. 1
The atlas recommends NMDC ARKin for environmental abundance and multi-omics analyses, alongside enigma-coral for field samples and protect-genomedepot for pathogen genomes. 1
Source
See the full inventory and integration analysis in BERDL Data Atlas — Inventory, Topic Map, and Cross-Reference Synergies. 1
See the community pathway-completeness, metabolomics, and ecosystem analysis in Community Metabolic Ecology via NMDC × Pangenome Integration. 3
See the compound-to-environment census and NMDC environmental atlas in ENIGMA Carbon Census — Tiered Knowledge Census of 83 Enrichment Compounds. 4
See the eukaryotic-contamination metadata analysis in Metadata Correlates of Eukaryotic Contamination in NMDC Prokaryote-Targeted Metagenomes. 5
See the Functional Dark Matter integration of NMDC environmental validation with fitness and gene-prioritization evidence in Functional Dark Matter — Experimentally Prioritized Novel Genetic Systems. 6
See the NMDC resource provenance, tenant placement, scale, currency, and discovery audit in NMDC Context Audit. 2