Raw report: acinetobacter_adp1_explorer Feeds into: circularity-in-metabolic-model-validation, condition-space-dimensionality, condition-specific-fitness, core-gene-annotation-paradox, cross-condition-metabolic-comparability, cross-tenant-data-bridging, data-landscape-ownership-and-coverage-bias, gene-essentiality, metabolic-model-gapfilling, multi-omics-integration, pangenome-conservation-fitness-decoupling, pangenome-integration, two-speed-bacterial-genome
Overview
This exploration analyzes a user-provided SQLite database containing 15 tables, 461,522 total rows, and 135 MB of data for Acinetobacter baylyi ADP1 and 13 related genomes. The database integrates genome features, TnSeq (transposon-insertion sequencing) essentiality, flux balance analysis (FBA, constraint-based prediction of metabolic fluxes), mutant growth fitness, proteomics, pangenome classification, functional annotations, metabolic reactions, and growth phenotypes, and connects these data to multiple KBase Data Lakehouse collections. 1
Key Findings
Multi-omics database structure
The database’s 15 tables hold 461,522 total rows and 135 MB of data for ADP1 and 13 related genomes; its central genome_features table contains 5,852 genes and 51 annotation columns spanning six data modalities: TnSeq essentiality with 58% coverage, FBA metabolic flux with 15% coverage, mutant growth fitness on 8 carbon sources with 39% coverage, proteomics across 7 strains with 41% coverage, pangenome classification with 54% coverage, and functional annotations through COG (Clusters of Orthologous Groups), KO, Pfam, and UniRef with 34-55% coverage. No single gene has data across all six modalities, although pairwise overlaps are substantial, particularly among essentiality, pangenome, and proteomics. 1
The database documents 7 engineered ADP1 strains, including wild-type ADP1 and 6 derivatives with aromatic amino acid pathway modifications involving ΔaroF and ΔaroG or dgoA variants; proteomics data are available for all 7 strains. Protein abundance was measured for 2,383 genes, and cross-strain correlation was high, indicating targeted rather than global effects of the engineered modifications. 1
The database contains 15 interconnected tables with 27 relationship pairs defined by shared columns. Its core data flow runs from genome to genome_features (annotations, essentiality, proteomics, growth), then to pan_genome_features (pangenome context), genome_reactions (metabolic model), gene_phenotypes (phenotype associations), and growth_phenotypes_detailed (FBA predictions). 1
KBase Data Lakehouse connectivity
Four of five tested connection types matched KBase Data Lakehouse at greater than 90%: 13 of 13 genome IDs matched the pangenome, 1,210 of 1,330 reactions (91%) matched biochemistry, 230 of 230 compounds matched biochemistry, and 4,891 of 4,891 cluster IDs matched the pangenome through mapping. ADP1 had 0 matches among 1 Fitness Browser organism query and was absent from the Fitness Browser. 1
All 13 KBase Data Lakehouse-format genomes belong to s__Acinetobacter_baylyi and the clade s__Acinetobacter_baylyi--RS_GCF_000368685.1. This KBase Data Lakehouse pangenome contains 3,207 core and 1,684 accessory gene clusters. The 120 unmatched reactions, representing 9% of the 1,330 reactions, may be custom or draft reactions not yet present in ModelSEED. ADP1 is absent from the Fitness Browser, so the report describes this database’s mutant growth data as a resource not available elsewhere in the KBase Data Lakehouse. 1
Pangenome cluster-ID bridge
ADP1 uses mmseqs2-style cluster IDs such as NHSXFYEX_mmseqsCluster_NNNN, whereas KBase Data Lakehouse uses centroid gene IDs such as NC_005966.1_1024; the two naming systems have 0% direct string match. A bridge through the KBase Data Lakehouse’s gene_genecluster_junction table links KBase Data Lakehouse cluster IDs to member gene IDs, which match the feature_id column in ADP1’s pan_genome_features table and thereby expose the ADP1-style cluster_id. 1
All 4,891 KBase Data Lakehouse clusters mapped successfully to 4,081 unique ADP1 clusters, yielding a 100% gene-level match across 43,754 genes. The generated data/cluster_id_mapping.csv file enables KBase Data Lakehouse pangenome annotations, including eggNOG and functional predictions, to be joined to ADP1 genes. 1
FBA and TnSeq essentiality
Of 866 genes with both FBA flux predictions and TnSeq essentiality calls, 639, or 73.8%, were concordant and 227 were discordant; the report describes this discordant 26% as genes where the metabolic model and experimental data disagree. The discordant genes are candidates for metabolic-model refinement or may reflect regulatory effects not represented by FBA. 1
FBA vs TnSeq concordance. From acinetobacter_adp1_explorer.
Essentiality was condition-specific: 499 genes were essential on minimal media compared with 346 on LB, consistent with an additional biosynthetic burden under minimal-media growth. 1
FBA flux classes changed between rich and minimal media for 177 of 866 genes, or 20%, indicating condition-dependent metabolic rewiring among the genes with flux data. 1
Condition-specific mutant fitness
Mutant growth fitness across 8 carbon sources had a mean pairwise correlation of 0.44. Urea fitness was nearly uncorrelated with quinate, at r = 0.11, and was weakly correlated with all other conditions, at r = 0.12-0.28, suggesting that urea catabolism involves a largely independent set of genes. Butanediol-acetate and butanediol-lactate showed the strongest correlations, at r = 0.58 and r = 0.53, respectively, consistent with shared central carbon metabolism. 1
Carbon source fitness correlation. From acinetobacter_adp1_explorer.
Because ADP1 is absent from the Fitness Browser, its mutant growth fitness measurements for 8 carbon sources constitute a resource not otherwise available in the KBase Data Lakehouse. 1
Essentiality, annotation, and pangenome status
Essential genes were more annotation-rich than dispensable genes: 33% of essential genes had COG assignments compared with 5% of dispensable genes, while 92% of essential genes had KEGG KO assignments compared with 53% of dispensable genes. Approximately 8% of essential genes lacked KO assignments and were identified as potential novel essential functions. 1
Essential genes were also more likely to belong to the core pangenome, consistent with the association between conservation and essentiality. 1
Metabolic reaction conservation and gapfilling
Of 1,330 unique metabolic reactions, 1,248, or 94%, were shared across all 14 genomes and classified as core; 62 were variable, occurring in 2-13 genomes, and 20 were genome-unique. Gapfilling accounted for 7.7% of reactions on average, with 243 missing functions cataloged. 1
Of 121,519 growth phenotype predictions across 14 genomes, 105,376, or 87%, required at least one gapfilled reaction. Prediction accuracy was therefore tightly coupled to gapfilling quality in this analysis, and false negatives had higher mean gap counts than correct predictions. 1
KBase Data Lakehouse-linked and novel data resources
The project connected the ADP1 database to the kbase_ke_pangenome collection, which supplied pangenome context for A. baylyi (13 genomes, 4,891 clusters); the kbase_msd_biochemistry collection, which supplied ModelSEED reaction and compound name resolution; and the kescience_fitnessbrowser organism table, which was checked for ADP1 and did not contain it. It also identified a 37-table Acinetobacter genome browser in phagefoundry_acinetobacter_genome_browser without deeply querying it. The user-provided database contributes mutant growth fitness on 8 carbon sources, proteomics across 7 engineered strains, TnSeq essentiality on minimal and LB media, and FBA predictions with gapfilling metadata. 1
Figures
essentiality_overview.png— TnSeq essentiality on minimal vs LB media with cross-tabulation;essentiality_distribution.png— essentiality category bar charts;essentiality_vs_pangenome.png— core/accessory fraction by essentiality class;fba_tnseq_concordance.png— FBA vs TnSeq essentiality agreement matrix;fba_flux_class_transition.png— rich vs minimal media flux class heatmap;metabolic_flux_classes.png— FBA flux class distributions;mutant_growth_fitness.png— fitness distributions and defect fractions by carbon source;mutant_growth_by_condition.png— mutant growth boxplots;growth_condition_correlation.png— pairwise carbon-source fitness correlation matrix;proteomics_cross_strain.png— protein abundance distributions and strain correlation;proteomics_wt_vs_engineered.png— wild-type vs most divergent strain scatter plot;data_coverage_by_modality.png— gene count per data modality;annotation_by_essentiality.png— annotation coverage of essential vs dispensable genes;reaction_gapfilling.png— reactions by evidence type per genome;reaction_conservation.png— core/variable/unique reaction breakdown;growth_phenotype_predictions.png— growth prediction classes and simulated vs observed scatter;gapfilling_impact.png— gapfill count distribution and mean by prediction class;ontology_distribution.png— ontology terms by source. 1
Caveats and Limitations
No gene has measurements across all 6 data modalities, and FBA flux data cover only 15% of genes; consequently, the FBA-TnSeq concordance analysis is limited to 866 genes. 1
The pangenome cluster mapping is indirect and passes through 3 tables. Although it is 100% complete at the reported gene level, the indirection could introduce edge cases where clusters split or merge between the KBase Data Lakehouse and ADP1 pangenome analyses. 1
The 87% dependence of growth phenotype predictions on gapfilled reactions limits interpretation of those predictions, and the 243 missing functions represent gaps in genomic evidence that affect prediction reliability. 1
The database covers only A. baylyi, so cross-species comparisons require comparable databases for other organisms. 1
The observed FBA-TnSeq discordance, urea-specific fitness pattern, core-metabolism conservation, and gapfilling dependence are primarily results from this ADP1-centered dataset. The report presents testing whether the 227 FBA-TnSeq-discordant genes are enriched for specific pathways or regulatory functions, to guide metabolic-model refinement, cross-species fitness comparison, PhageFoundry cross-referencing, urea-specific gene identification, and pangenome-informed gapfill confidence assessment as future analyses rather than completed findings. 1
Slots Into
- multi-omics-integration — the database’s six data modalities, incomplete modality overlap, and cross-strain proteomics provide an integrated multi-omics case study.
- gene-essentiality — the FBA-TnSeq concordance, condition-specific essentiality, annotation enrichment, and essentiality-pangenome relationship provide evidence for comparing gene-essentiality predictors.
- condition-specific-fitness — carbon-source fitness correlations, especially the weak relationship of urea to other conditions, provide condition-specific mutant-growth evidence.
- pangenome-integration — the 100% KBase Data Lakehouse-to-ADP1 cluster bridge and core/accessory comparisons provide a cross-dataset pangenome integration example.
- metabolic-model-gapfilling — reaction conservation, 243 missing functions, FBA flux transitions, and the 87% gapfilling dependence of growth predictions constrain metabolic-model interpretation.
- cross-tenant-data-bridging — genome, reaction, compound and cluster-ID match rates against Data Lakehouse collections, and the 0% direct cluster-ID match resolved through a junction-table bridge.
- data-landscape-ownership-and-coverage-bias — because ADP1 is absent from the Fitness Browser, this database is the lakehouse’s only source of condition-specific fitness data for ADP1.
- condition-space-dimensionality — moderate mean pairwise carbon-source fitness correlation, with urea as a weakly correlated outlier.
- core-gene-annotation-paradox — essential genes are more annotation-rich, yet a fraction lack KO assignments.
- pangenome-conservation-fitness-decoupling — essential genes are more likely to be core pangenome members.
- two-speed-bacterial-genome — most metabolic reactions are core across the 14 genomes, with few variable or genome-unique reactions.
- circularity-in-metabolic-model-validation — most growth predictions depend on gapfilled reactions, and false negatives carry more gaps.
- cross-condition-metabolic-comparability — FBA flux-class changes between rich and minimal media. 1