Condition-space dimensionality describes how many independent axes are needed to represent gene-dependent growth effects across tested substrates. In the Acinetobacter baylyi ADP1 deletion collection, the evidence supports a multidimensional phenotype space in which general growth sensitivity is distinct from substrate-specific metabolic requirements, while broad condition-dependent variation coexists with a small discrete pathway-specific module. 1 The same evidence motivates, but does not establish, the hypothesis that perturbation modality may influence whether experiments reveal continuous condition-dependent effects or discrete phenotypic modules. 1 The respiratory-chain analysis supports this multidimensional interpretation by showing that carbon sources can select qualitatively different respiratory configurations, not merely different magnitudes of total reducing-equivalent production. 2
The pathway-capability analysis supports this distinction at a broader comparative scale: among 161 organism–pathway combinations, complete pathways were separated into Active Dependency and Latent Capability classes according to whether pathway genes showed fitness importance under the tested conditions. 3 All 66 aggregate Latent Capability pairs became fitness-important under at least one condition type, supporting the interpretation that latent requirements occupy condition-specific dimensions rather than representing permanent dispensability; however, the median-based condition-specific threshold can cause reclassification by construction, so this evidence requires independent calibration. 3
The new 48-organism analysis supports this capability–dependency distinction with a larger set of 1,695 complete pathway–organism pairs: 267 (15.8%) were latent, 547 (32.3%) intermediate, and 881 (51.9%) active under thresholds based on pathway fitness t-scores and essential-gene fractions. 4 It refines the earlier condition-space interpretation by showing that the distinction is especially common for carbon-source utilization pathways, for which 217 of 892 (24.3%) were latent, compared with 48 of 735 (6.5%) amino acid biosynthesis pathways. 4 The latent fraction remained non-trivial across the tested threshold combinations, ranging from 4.7% to 21.1%, although the large intermediate class makes the exact fraction calibration-sensitive. 4
Key evidence
The analysis used a complete growth matrix of 2,034 genes measured across 8 carbon sources, with growth ratios defined as mutant/wild-type values. 1 Growth ratios below 1.0 indicated growth defects, whereas values above 1.0 indicated no defect or a slight growth advantage from experimental normalization. 1 Mean growth ratios ranged from 0.41 on urea to 1.36 on quinate, a 3.3-fold range. 1
Principal component analysis (PCA), a method that summarizes correlated variation into orthogonal components, found that 5 components captured 82% of the variance in the 2,034×8 matrix. 1 PC1 explained 36.7% of the variance and represented general growth sensitivity because all conditions loaded positively on it. 1 PC2 explained 12.7% of the variance and isolated urea, which had a loading of +0.75, separating nitrogen metabolism from carbon metabolism. 1 No single component captured more than 37% of the variance, supporting a multidimensional interpretation rather than a single dominant sensitivity axis. 1
Pairwise Pearson correlations were moderate at best: acetate–butanediol had the highest correlation at r = 0.58, while the median across all 28 condition pairs was r = 0.25. 1 These correlations support approximately 5 independent dimensions of phenotypic information rather than a simple two-group model contrasting demanding and robust conditions. 1 A separate analysis of 478 TnSeq-dispensable genes across eight carbon sources found that 333 genes (70%) showed condition-specific growth defects, 10 genes (2%) showed defects across all eight conditions, and 135 genes (28%) showed no defect on any condition; mean pairwise defect correlation was 0.38, ranging from -0.03 to 1.0. 5 That report’s Methods list its eight carbon sources as glucose, acetate, asparagine, butanediol, glucarate, lactate, pyruvate, and succinate, so its panel includes pyruvate and succinate where the deletion-phenotype panel uses urea and quinate; the two panels are therefore not matched. 5,1 This supports the multidimensional interpretation while refining its scope: condition dependence is also evident in the subset used for combined TnSeq, FBA, and mutant-growth analysis, but the two analyses differ in gene coverage and in their condition panels.
The ADP1 explorer analysis reports a different summary statistic for mutant growth fitness across 8 carbon sources: mean pairwise correlation was r = 0.44, with butanediol–acetate (r = 0.58) and butanediol–lactate (r = 0.53) the strongest pairs, consistent with shared central carbon metabolism. 6 Urea fitness was nearly uncorrelated with quinate (r = 0.11) and weakly correlated with all other conditions (r = 0.12-0.28); the report interprets this as suggesting that urea catabolism involves a largely independent set of genes, which is an interpretation rather than a direct gene-set test. 6 This supports the PCA result in which PC2 isolates urea, but the explorer’s mean of r = 0.44 is higher than the deletion-phenotype median of r = 0.25; because the reports use different summary statistics and do not state matched gene sets, the difference is recorded under Tensions rather than reconciled. 6,1
The broader capability analysis refines this interpretation by showing that condition-space dependence is not limited to one deletion matrix: nitrogen limitation, stress, and carbon limitation were the most frequent condition types that exposed latent pathway requirements. 3 Thus, a pathway can be genomically complete yet phenotypically silent under standard conditions and important along specific environmental axes. 3 The new analysis supports this conclusion: carbon utilization pathways were much more often latent than amino acid biosynthesis pathways, indicating that the type of environmental axis helps determine whether encoded capacity is expressed as measured dependency. 4
General sensitivity and substrate-specific requirements
The 8 carbon sources formed demanding, moderate, and robust tiers, with urea, acetate, and butanediol in the demanding tier; asparagine and lactate in the moderate tier; and glucarate, glucose, and quinate in the robust tier. 1 Mean growth ratios ranged from 0.41 on urea to 1.36 on quinate, and the fraction of genes with severe defects at growth ratio < 0.5 ranged from 97.9% on urea to 1.6% on quinate; the report describes this tier structure as consistent across multiple thresholds. 1 This tier structure supports a broad sensitivity axis but does not make condition profiles interchangeable: the moderate correlations and PCA results indicate additional substrate-linked axes. 1
A cross-project digest refines the general-versus-specific distinction with a different dataset: the top 18 genes significant across all 21 Fitness Browser metabolite conditions are amino acid biosynthesis enzymes, including homoserine O-acetyltransferase (methionine), ATP phosphoribosyltransferase (histidine), dihydroxy-acid dehydratase (branched-chain), isopropylmalate dehydrogenase (leucine), shikimate dehydrogenase (aromatic), and imidazoleglycerol-phosphate dehydratase (histidine), described as essential for growth on any minimal medium, whereas the ~370 genes significant in only 1-2 conditions are described as the truly substrate-specific ones. 7 This supports separating a shared minimal-medium requirement axis from substrate-specific axes; the digest flags the distinction as critical for interpreting Web of Microbes–Fitness Browser integration, and it is an interpretive classification rather than a formal dimensionality analysis. 7
A total of 625 genes, or 31% of the complete matrix, had a condition-specificity score ≥ 1.0, indicating that their growth importance was concentrated on one carbon source. 1 Quinate selected pcaC, pcaG, pcaH, pcaB, quiA, quiB, pqqC, and pqqD for protocatechuate/quinate degradation and PQQ biosynthesis, while urea selected ureA, ureB, ureC, ureD, ureE, ureF, and ureG for the urease complex. 1 Other substrates likewise selected pathway-linked genes, including fatty acid β-oxidation and the glyoxylate shunt on acetate, the Entner-Doudoroff pathway and PQQ-dependent glucose dehydrogenase on glucose, and respiratory and lactate-associated genes on lactate. 1
The respiratory-chain study supports this substrate-specific interpretation: among ADP1’s 62 respiratory-chain genes across 8 subsystems, quinate specifically required Complex I, acetate required Complex I, cytochrome bo3, ACIAD3522, and additional components, lactate specifically required cytochrome bo3, glucose had no specifically required respiratory component, and urea was generally demanding across the reported respiratory profile. 2 This refines the broad demanding-versus-robust tiering by showing that even substrates with different overall sensitivity can impose distinct pathway configurations. The report attributes these differences to carbon-source-dependent NADH flux and capacity constraints rather than a simple quantitative change in total reducing-equivalent yield. 2
The 478-gene analysis supports this substrate-specific interpretation but found that condition-specific growth does not necessarily correspond to condition-specific FBA behavior. 5 Condition-specific FBA flux had weak, mixed correlations with measured growth: glucose ρ = -0.021 (p = 0.677; n = 387), acetate ρ = -0.153 (p = 0.004; n = 352), asparagine ρ = -0.257 (p < 0.001; n = 286), butanediol ρ = -0.145 (p = 0.092; n = 137), glucarate ρ = +0.246 (p = 0.005; n = 127), and lactate ρ = -0.160 (p = 0.065; n = 135). 5 The positive glucarate correlation is opposite the expected direction and refines the interpretation of substrate-linked axes by indicating that some axes may be poorly represented by model constraints rather than absent from the phenotype.
The respiratory analysis supports this model–experiment qualification: FBA, or flux-balance analysis, predicted zero flux through NDH-2 and ACIAD3522 on all standard media because growth optimization routes NADH through ATP-favorable Complex I. 2 This refines the interpretation of weak or discordant condition-specific FBA by showing that optimization can omit alternative, suboptimal respiratory pathways whose capacity constraints may nevertheless affect measured growth. The proposed quinate mechanism is that β-ketoadipate-pathway cleavage produces succinyl-CoA and acetyl-CoA simultaneously, creating a concentrated TCA-cycle NADH burst that may exceed NDH-2 reoxidation capacity; this is a biochemical interpretation based on theoretical stoichiometry, not measured flux distributions. 2
The quinate-specific gene set contained 51 genes at specificity > 0.5 and z < -1. 1 It included core aromatic degradation genes and NADH-ubiquinone oxidoreductase subunits from respiratory Complex I, suggesting the hypothesis that aromatic catabolism creates distinctive electron-transport-chain demands. 1 PQQ biosynthesis genes were condition-specific for both quinate and glucose, consistent with PQQ-dependent dehydrogenases catalyzing the first step of both pathways. 1 The 24-gene quinate module is therefore a particularly concentrated subset of the larger population of 625 condition-specific genes. 1
The respiratory analysis supports the hypothesis that aromatic metabolism can produce experimentally observed condition-specific requirements, which FBA did not predict, and refines it with a possible flux-capacity mechanism: Complex I had growth ratios of 0.37 on quinate, 1.44 on glucose, and 0.49 on acetate, while ACIAD3522 had a growth ratio of 0.013 on acetate and was dispensable on quinate and glucose. 2 However, ACIAD3522 may not be a respiratory NADH dehydrogenase in the strict sense and could have another metabolic function. 2 The related model–experiment gap remains: 9 of 11 aromatic-degradation genes were FBA-discordant, with odds ratio (OR) = 9.70 and Benjamini-Hochberg false discovery rate (FDR)-adjusted q = 0.012, while directional enrichment for FBA under-prediction was OR = 12.0, q = 0.004. 5 This refines the proposed respiratory-demand interpretation by suggesting that missing aromatic substrates or mismatched environmental assumptions may contribute to apparent pathway-specific requirements. 5
The comparative capability analysis supports the separation between general sensitivity and substrate-specific requirements beyond ADP1. Among 48 organisms and 1,695 complete pathway–organism pairs, 267 (15.8%) were classified as latent, 547 (32.3%) intermediate, and 881 (51.9%) active; pathway category strongly predicted class (χ²=163.6, df=4, p=2.5×10⁻³⁴). 4 Carbon-source utilization pathways were 3.7× more likely to be latent than amino acid biosynthesis pathways, using the reported 24.3% and 6.5% fractions. 4 This refines the ADP1 result: environmental specificity is not only a property of individual substrates in one organism, but can also appear as category-dependent realization of complete pathways across organisms. 4
Continuous variation, discrete modules, and perturbation modality
Hierarchical clustering produced an optimal K = 3 but a low silhouette score of 0.24. 1 The two large modules contained 1,160 and 850 genes and represented broadly sensitive versus broadly tolerant profiles, rather than sharply distinct functional categories. 1 No specific functional enrichment survived false discovery rate (FDR) correction, where FDR controls the expected proportion of false positives among declared findings. 1
These results support a continuous condition-dependent essentiality gradient rather than a landscape composed mainly of discrete functional classes. 1 They refine gene-essentiality by showing that gene importance can vary along a continuum even when genes are classified as dispensable or essential. 1 The two large clusters represent broad sensitivity and tolerance, but their weak separation and lack of surviving functional enrichment do not establish sharply discrete biological categories. 1
The triple-essentiality analysis supports this continuum interpretation from an independent measurement design: in that report’s expanded knockout-validation ROC analysis, which is separate from its 478-gene TnSeq-dispensable subset, continuous inverted fitness predicted knockout essentiality better than binary essentiality fraction, with AUC = 0.700 in rich media and AUC = 0.725 in minimal media, whereas essentiality fraction had AUC = 0.344 and AUC = 0.403, respectively. 5 AUC means Area Under the ROC Curve, a threshold-independent classification measure. This refines condition-specific-fitness by showing that continuous fitness can retain information that is lost when condition-dependent effects are reduced to binary labels.
A small module of 24 genes retained a sharply pathway-specific phenotype within this broader gradient. 1 These genes had extreme quinate-specific defects, with a mean quinate z-score of -7.28 and near-zero scores on the other conditions. 1 The genes belonged to aromatic degradation pathways, making this the only discrete phenotypic module identified in the dataset. 1 This discrete structure is most detectable when a pathway is specifically required by one tested substrate, rather than when genes have broadly shared effects across conditions. 1
The continuous-versus-modular distinction supports condition-specific-fitness while distinguishing graded condition dependence from genuinely discrete pathway-level structure. 1 It also refines gene-essentiality, because the 625 condition-specific genes demonstrate why binary essential/non-essential labels do not capture the full phenotype architecture. 1 The approximately 5 dimensions can therefore coexist with a low-dimensional pathway-specific exception. 1
The ADP1 result supports the interpretation that single-gene deletions can expose a continuous gradient of condition-dependent fitness, with a small pathway-specific module superimposed on that gradient. 1 The report contrasts this pattern with discrete phenotypic modules reported for chemical-genetic profiles in escherichia-coli, but does not establish whether the difference is caused by perturbation modality, organismal metabolism, experimental design, or some combination of these factors. 1 Accordingly, genetic deletions and chemical perturbations may expose different phenotype architectures because they alter biological systems through different perturbational routes. 1 An alternative hypothesis is that ADP1 metabolism is more interconnected than the system represented by the compared E. coli chemical-genetic profiles, causing single-gene deletions to produce a smoother landscape. 1
The new comparative result supports the existence of graded and condition-dependent realization at a different resolution: complete pathways were not uniformly active or latent, but distributed among active, intermediate, and latent classes. 4 It therefore refines the continuous-versus-modular distinction by showing that thresholded pathway classes summarize a continuum of measured dependence and should not be interpreted as discrete ecological states without independent validation. 4
Scope and interpretation
The condition-specific genes mapped to expected substrate-linked pathways, including aromatic degradation for quinate, the urease complex for urea, fatty acid β-oxidation and the glyoxylate shunt for acetate, and the Entner-Doudoroff pathway for glucose. 1 The quinate-specific set contained 51 genes with specificity > 0.5 and z < -1, including core aromatic degradation genes and respiratory Complex I subunits. 1
The analysis used 8 carbon sources, and the report notes that the approximately 5 independent dimensions may increase when additional conditions are measured. 1 Growth ratios were single-timepoint measurements with unknown technical noise, so some condition-specificity scores may reflect measurement error as well as biology. 1 The complete matrix excluded 499 essential genes and 316 genes with incomplete data, biasing the analysis toward dispensable genes with successful deletion mutants. 1
The aromatic catabolism network analysis supports this caveat from the co-fitness side: it estimates that the 8-condition growth matrix provides only ~5 independent dimensions, and it notes that its co-fitness analysis uses only 8 conditions (8-dimensional growth vectors), limiting the resolution of gene–gene correlations. 8 Limited condition-space dimensionality therefore constrains not only phenotype-axis inference but also downstream co-fitness module and network assignment. 8
The new triple-essentiality analysis is likewise restricted in part by viable-mutant coverage: its original growth analysis concerns 478 TnSeq-dispensable genes and cannot be interpreted as a whole-genome test of FBA lethality prediction. 5 This refines the scope of the dimensionality claim: the observed condition structure is robustly detectable in a dispensable-gene subset, but the relationship between condition-space dimensionality and lethal genes remains incompletely sampled.
The respiratory evidence adds analogous limits: NDH-2 has no growth data, the stoichiometric analysis uses theoretical rather than measured fluxes, the cross-species comparison retained only 4 organisms lacking NDH-2, and annotation-based searches may miss orthologs. 2 The proposed wiring model therefore remains a condition-specific ADP1 hypothesis, not a general cross-species rule. 2
The pathway-capability study extends the scope from one organism’s carbon-source matrix to 2,810 GTDB species with at least 10 genomes and 80 pathways. 3 Variable pathway count correlated with pangenome openness at partial Spearman rho=0.530, p=2.83e-203 after controlling for genome count, and metabolic ecotype count correlated with openness at partial rho=0.322, p=8.0e-07. 3 These results support the interpretation that conditionally expressed metabolic requirements can form broader within-species dimensions associated with genome fluidity, while the correlations remain vulnerable to sampling and incomplete phylogenetic control. 3 Ecotype profiles were clustered from binary completeness across 80 pathways; among 225 species meeting the study filters, the median number of ecotypes was 4 and the maximum was 8. 3
The new analysis supports this pangenome-linked interpretation at the fitness-tested scale: after aggregation to 22 species clades, latent capability rate correlated positively with pangenome openness (Spearman ρ = 0.69, p = 0.0004, n = 22 clades). 4 However, pathway-level conservation did not distinguish latent capabilities from active dependencies: mean conservation was 0.829 for 755 active dependencies, 0.907 for 508 intermediate pathways, and 0.869 for 248 latent capabilities, with Mann–Whitney U p = 0.94 for active > latent and rank-biserial r = 0.052. 4 This refines the broader pangenome result by indicating that the association is more evident at the clade-level relationship between latent rate and genome openness than in simple per-pathway conservation comparisons. 4
Only 7 of 48 Fitness Browser organisms had matching GapMind genome data, and the model organisms had near-complete core genomes; therefore, the comparative capability–dependency evidence is not a replacement for expanded condition measurements in ADP1. 3 GapMind assessed 18 amino acid biosynthesis pathways and 62 carbon-source utilization pathways, leaving other metabolic axes untested. 3 The pathway analysis also did not demonstrate metabolite exchange from accessory-dependent biosynthesis, and its condition-specific reclassification is limited by the median-based importance threshold. 3
The new analysis adds related limitations: it used GapMind completeness with SEED subsystem annotations as a pathway-membership proxy, excluded deoxyribonate and myoinositol because no matching SEED role descriptions were found, included phenylalanine and tyrosine through phe and tyr abbreviation matching, and excluded alanine because fewer than 3 SEED-annotated genes met the coverage threshold. 4 Only 48 organisms were fitness-tested among 293,000 genomes with pathway predictions, and laboratory fitness conditions may make environmentally important pathways appear latent. 4 These limitations refine the interpretation of comparative condition-space dimensions by showing that both condition coverage and pathway mapping constrain which axes can be observed. 4
The comparison with discrete chemical-genetic modules in E. coli is interpretive rather than a direct reanalysis here, so it does not resolve whether perturbation modality or ADP1 metabolic interconnectedness explains the apparent balance between continuous and modular architectures. 1 The low silhouette score and absence of FDR-significant enrichments support the gradient interpretation, but they do not establish that all ADP1 phenotypic structure is continuous. 1 Proposed independent component analysis and an expanded condition panel remained future work rather than completed analyses. 1
Metabolic clustering provides a broader analogue of condition-space structure: all 10 target species in the new analysis had silhouette scores greater than 0.2, ranging from 0.35 for PALSA-747 sp. to 0.89 for Salmonella enterica. 4 Cluster associations with isolation environment were significant in Salmonella enterica (χ²=1570.2, df=25, p<0.0001) and Phenylobacterium sp. (χ²=12.2, df=1, p=0.0005), but not in Stutzerimonas, Alteromonas, Pelagibacter, or Prochlorococcus. 4 This supports the idea that metabolic dimensions can align with ecological context in some species, while refining it with a clear warning that marine non-associations may reflect coarse isolation-source metadata, depth- or nutrient-zone structure, or insufficient pathway resolution; these observational associations do not establish causality. 4
This finding supports condition-specific-fitness by showing that fitness effects are distributed across multiple substrate-linked dimensions rather than captured by a single global essentiality score. 1 It also supports the use of principal-component-analysis for separating shared sensitivity from condition-specific variation in deletion-phenotype matrices. 1 The combined evidence supports metabolic-model-gapfilling because condition-specific experimental defects can remain informative even when condition-specific FBA flux is weakly or incorrectly aligned with growth. 5 The respiratory findings refine metabolic-model-gapfilling by identifying growth optimization and respiratory capacity constraints as a specific route by which FBA can miss condition-dependent requirements. 2 The pathway-capability results support pangenome-integration by linking variable pathway complements and metabolic ecotypes to pangenome openness, while refining the interpretation through explicit genome-count controls. 3 The new clade-level association supports this connection, but the null pathway-conservation comparison and limited fitness-tested sampling caution against treating openness as evidence that any particular latent pathway is ecologically maintained. 4
Tensions
The approximately 5 independent dimensions are inferred from only 8 tested carbon sources, so the dimensionality may increase when additional conditions are measured. 1 The low clustering silhouette score and absence of FDR-significant functional enrichments favor a gradient interpretation, but they do not exclude additional discrete modules that would emerge under a broader condition panel or with improved measurement precision. 1
The single-gene deletion and chemical-genetic comparison is interpretive rather than directly tested here, so it does not resolve whether perturbation modality or ADP1 metabolic interconnectedness explains the apparent balance between continuous and modular architectures. 1 The ADP1 dataset alone therefore does not establish a cross-organism rule about genetic deletions versus chemical perturbations. 1
The deletion-phenotype analysis identifies aromatic degradation as a concentrated quinate-specific module, whereas the triple-essentiality analysis finds aromatic degradation strongly enriched among FBA-discordant genes and reports directional FBA under-prediction. 1 5 These findings are compatible but leave unresolved whether the apparent pathway-specific requirement primarily reflects biological substrate dependence, missing model inputs, or both; the new analysis proposes trace aromatics and mismatched environmental assumptions as hypotheses requiring direct testing rather than as established explanations. 5
The respiratory analysis supports the existence of substrate-specific respiratory requirements, but it also leaves unresolved whether the apparent quinate Complex I requirement reflects a concentrated NADH flux burst, another function of ACIAD3522, unmeasured environmental inputs, or limitations of the fitness and FBA datasets. 2 Its cross-species test found validated NDH-2 in 5 of 14 organisms; organisms with validated NDH-2 had a larger mean Complex I aromatic deficit, −0.297, than organisms without validated NDH-2, −0.156, but the difference was not statistically significant (p = 0.52) and was opposite to the predicted compensation pattern. 2 Thus the ADP1-specific respiratory interpretation should not yet be generalized across species. 2
The pathway-capability results create a related calibration tension: all 66 aggregate Latent Capability pairs became important under at least one condition type, which supports condition dependence, but the analysis itself cautions that median thresholds applied within condition subsets can force crossings. 3 This does not contradict the ADP1 evidence for multiple phenotype axes, but it limits how strongly condition-specific pathway reclassification can be treated as independent biological confirmation. 3
The new capability–dependency analysis creates a second, distinct tension with a simple conservation interpretation. Its clade-level result associates higher latent capability rates with more open pangenomes (ρ = 0.69, p = 0.0004, n = 22 clades), but its pathway-level comparison did not distinguish latent from active pathways (Mann–Whitney U p = 0.94; rank-biserial r = 0.052). 4 These results should not be reconciled by treating openness as proof of pathway-level conservation or loss: they support different levels of inference and require analyses that connect individual pathway trajectories to clade-level genome dynamics. 4
Two reports summarize pairwise condition correlations across 8 carbon sources differently: the deletion-phenotype analysis reports a median of r = 0.25 across all 28 pairs, whereas the ADP1 explorer reports a mean pairwise r = 0.44, although both report r = 0.58 for the acetate–butanediol pair. 1,6 Mean-versus-median summaries or different gene filtering could account for the gap, but neither report establishes this, so the moderate-correlation interpretation should be cited with the statistic each report used rather than averaged. 1,6
Open Directions
- Measure the same deletion collection across an expanded condition panel and use PCA or independent component analysis (ICA), a method that separates statistically independent signal sources, to test whether the approximately 5-dimensional structure remains stable or increases with substrate diversity and whether additional pathway-specific modules emerge. 1
- Recompute mean and median pairwise condition correlations on one matched gene set from the deletion collection to test whether the reported r = 0.44 and r = 0.25 summaries differ because of statistic choice or gene filtering. 6,1
- Replicate the single-timepoint growth measurements with technical replicates and time-course assays, quantify measurement error, and test whether condition-specificity scores ≥ 1.0 and the 24-gene quinate module with mean quinate z-score -7.28 exceed measurement noise. 1
- Map the 625 condition-specific genes onto pathway annotations and compare their loadings across conditions to determine which metabolic modules define the additional axes beyond general sensitivity. 1
- Test the hypothesis that quinate and protocatechuate catabolism impose distinctive respiratory-chain demands by measuring respiratory Complex I mutants, including the 51-gene quinate-specific set, under quinate, glucose, and related aromatic substrates. 1
- Reanalyze the 24-gene quinate module with additional aromatic substrates to test whether it is specific to quinate or represents a broader aromatic-catabolism axis. 1
- Compare matched ADP1 single-gene deletions and chemical perturbations across the same 8 carbon sources using identical growth measurements, hierarchical clustering, silhouette, and FDR procedures to test whether chemical perturbations produce more discrete modules than the deletion collection. 1
- Integrate deletion phenotypes with RB-TnSeq, or random barcode transposon sequencing, measurements under matched conditions and compare condition-specificity scores to test whether perturbation modality changes the apparent continuity of fitness effects. 1
- Compare ADP1 and E. coli using matched perturbation types, condition panels, and clustering metrics to ask whether organismal metabolic interconnectedness or perturbation modality better explains the difference between continuous and discrete architectures. 1
- Measure experimental media composition, add trace aromatic compounds to FBA media definitions, and rerun condition-specific flux simulations to test whether the aromatic FBA-discordance pattern and the positive glucarate correlation are caused by missing substrates or incorrect environmental constraints. 5
- Construct an ADP1 NDH-2 deletion mutant and measure NADH/NAD⁺ ratios, growth, and respiratory-chain phenotypes on quinate, glucose, acetate, lactate, and urea to test whether flux rate and capacity explain the condition-specific Complex I pattern. 2
- Reanalyze quinate-versus-succinate proteomics and characterize ACIAD3522 to distinguish passive flux-based respiratory wiring from transcriptional regulation and to determine whether ACIAD3522 has a respiratory function. 2
- Expand KO-based K03885 and K00330–K00343 searches across the planned 27K-species comparison and perform pangenome co-occurrence analysis to test whether NDH-2/Complex I relationships generalize beyond ADP1. 2
- Perform condition-matched TnSeq and complete-knockout experiments, including domain- and insertion-position analysis, to determine whether continuous fitness effects and condition-space dimensions differ systematically between perturbation modalities. 5
- Validate the 66 Latent Capability reclassifications against an independent essentiality set and condition-specific held-out data, testing whether pathway requirements recur across environmental axes without median-threshold inflation. 3
- Extend the ADP1 condition-space analysis to matched species with GapMind and Fitness Browser data, testing whether pathway variation and metabolic ecotype structure predict the number and stability of experimentally observed fitness dimensions after controlling for genome count and phylogeny. 3
- Recompute the 1,695 capability–dependency classifications using direct GapMind per-step gene assignments rather than SEED-proxy pathway membership, and test whether the 15.8% latent fraction and the carbon-versus-amino-acid contrast persist. 4
- Test the 22-clade latent-rate/open-pangenome association with pathway-specific gene-loss trajectories, matched fitness conditions, and phylogenetic correction to determine whether genome openness predicts individual latent pathways or only aggregate capability profiles. 4
- Recluster the 10 metabolic-ecotype species using finer environmental metadata, phylogenetic correction, and gene-level rather than pathway-level features to distinguish genuine environmental dimensions from isolation-source and annotation confounding. 4
Triple Essentiality Concordance Analysis for *Acinetobacter baylyi* ADP1 provides the detailed concordance analysis underlying the integrated evidence. Condition-Specific Respiratory Chain Wiring in ADP1 provides the detailed respiratory-chain analysis. Metabolic Capability vs Dependency provides the detailed capability–dependency and comparative pangenome analysis. Metabolic Capability vs Metabolic Dependency provides the new 48-organism capability–dependency, pangenome-openness, and metabolic-ecotype analysis.