One page per research project, summarising it and linking to its raw report. These are the evidence every other page on the site cites. Two cross-project digests sit here too.
Collection
Project reports
AI-generated wiki from AI-conducted research reports. Partially reviewed by scientists. See About This Wiki.
- Project reports
- 73
- Cross-project digests
- 2
75 items under this folder.
Discoveries LogCross-project digest of 2026 BERIL research findings — direct measurements, methodological corrections, null results, contradictions and reusable analytical rules spanning microbial fitness, pangenome evolution, metabolic modeling, resistance and defense, subsurface ecology, microbiome analysis and research-quality practice.KBase Data Lakehouse: Common Pitfalls & GotchasCross-project reference of practical pitfalls when querying and analyzing KBase Data Lakehouse databases, covering access, joins, typing, Spark scale limits, database-specific schema rules, statistical design, and reproducibility.Functional Dark Matter — Experimentally Prioritized Novel Genetic SystemsSummary of the functional_dark_matter project, which integrates fitness, essentiality, pangenome, environmental, pathway and reannotation evidence to prioritize experimentally actionable bacterial genes of unknown function across 48 Fitness Browser organisms.Plant Microbiome Ecotypes — Compartment-Specific Functional Guilds and Their Genetic ArchitectureSummary of the plant_microbiome_ecotypes project, which tests compartment effects, beneficial-versus-pathogenic genome architecture, HGT, complementarity, plant-enriched gene families, dual-nature classification and within-species subclade structure across 293,059 GTDB genomes, including its Phase 2b adversarial-review corrections.Metagenome-Prioritized Phage Cocktails for Crohn's Disease and IBDProject report compiling four gut-microbiome ecotypes, rigor-controlled Tier-A pathobiont prioritization, pathway/BGC/metabolite/strain/serology mechanism analyses, three layers of phage evidence and per-patient hybrid phage-cocktail drafts for 23 UC Davis Crohn's disease patients.Metabolic Capability vs DependencySummary of the pathway_capability_dependency project, which separates GapMind-predicted pathway capability from RB-TnSeq-observed dependency in 7 model bacteria and links pathway variation to pangenome openness and metabolic ecotypes across 2,810 GTDB species.Gene Function Ecological AgoraSummary of the Gene Function Ecological Agora project, a GTDB-r214-scale Producer × Participation and Sankoff-parsimony atlas of clade × KO gene-function innovation and acquisition depth, with its pre-registered hypothesis verdicts, ecological grounding, methodology revisions and caveats.Metabolic Capability vs Metabolic DependencySummary of a 48-organism study comparing GapMind metabolic pathway completeness with Fitness Browser fitness dependency, and relating fitness-neutral latent pathways to pathway conservation, pangenome openness and within-species metabolic ecotypes.Pan-Bacterial Metal Fitness AtlasCross-species atlas of bacterial gene fitness under metal stress, showing that metal-important genes are enriched in the core genome and proposing a two-tier core-stress versus accessory-resistance model.Pan-Bacterial AMR Gene LandscapePan-bacterial atlas of AMRFinderPlus antimicrobial-resistance genes across 27,690 pangenome species, covering core-versus-accessory conservation, mechanisms, taxonomic hotspots, environment associations, Fitness Browser fitness effects, caveats and proposed follow-up analyses.Metal Resistance Ecology — Phylogenetic Conservation vs. Environmental SelectionSummary of the microbeatlas_metal_ecology project, which links genus-level metal-resistance type diversity from pangenome AMR annotations to global MicrobeAtlas niche breadth using phylogenetic comparative methods, with groundwater and ENIGMA ORFRC field validation and extensive sensitivity caveats.Metal-Specific vs General Stress GenesSummary of the metal_specificity project, which classifies metal-important fitness genes from 24 organisms as metal-specific, metal+stress, or general sick, and tests their conservation, functional enrichment, candidate families, and module behavior.Rational Design of Protective Microbiome Formulations for Competitive Exclusion of *Pseudomonas aeruginosa* in Cystic Fibrosis AirwaysDesign of commensal consortia to competitively exclude *Pseudomonas aeruginosa* from cystic-fibrosis airways, integrating inhibition assays, carbon-utilization and growth-kinetics profiling, patient meta-omics, PA pangenome and virulence typing, and multi-criterion formulation optimization.Field vs Lab Gene Importance in *Desulfovibrio vulgaris* HildenboroughA single-organism analysis of Desulfovibrio vulgaris Hildenborough testing whether field-relevant versus lab-condition RB-TnSeq fitness importance predicts pangenome core status, finding that fitness importance itself matters more than ecological condition class.Genotype × Condition to Phenotype Prediction from ENIGMA Growth CurvesSummary of the ENIGMA genotype × condition project that predicts bacterial growth phenotypes from KEGG-ortholog content, finding binary growth partly predictable but growth kinetics not, and proposing field-relevant active-learning experiments for Oak Ridge.SNIPE Defense System in the KBase Data Lakehouse PangenomeProject report evaluating the SNIPE phage-defense system across the KBase Data Lakehouse pangenome, covering its corrected PF13250/PF13455 domain architecture, ManXYZ fitness trade-off, prevalence and mobility, environmental association, and phage-host and Klebsiella evidence.Annotation-Gap Discovery via Phenotype-Fitness-Pangenome-Gapfilling IntegrationSummary of the annotation_gap_discovery project, which triangulated gapfilling, Fitness Browser fitness, pangenome, GapMind and BLAST evidence to assign candidate genes to gapfilled metabolic reactions across 14 bacteria.Ecotype Reanalysis — Environmental vs Human-Associated SpeciesReanalysis testing whether environmental species show stronger environment–gene content correlations than human-associated species in the AlphaEarth subset, finding a null result despite confirmed clinical sampling bias.Fitness Effects vs Conservation — Quantitative AnalysisQuantitative comparison of Fitness Browser mutant fitness effects with pangenome conservation across ~194,000 genes from 43 bacteria, showing a weak but robust gradient in which fitness-important genes are more often core.Gene-Resolution Metal Cross-Resistance Across Diverse BacteriaSummary of the metal_cross_resistance project, which used gene-level fitness data across 28 bacteria to show universally positive metal cross-resistance, a three-tier gene conservation gradient, and a null BacDive isolation-environment validation at Fitness Browser scale.Ecotype Correlation AnalysisSummary of the ecotype_analysis project, which tested across 172 bacterial species whether environmental (AlphaEarth embedding) or phylogenetic similarity better predicts pangenome gene-content similarity.ENIGMA Carbon Census — Tiered Knowledge Census of 83 Enrichment CompoundsSummary of the ENIGMA Carbon Census, a tiered knowledge census of 83 groundwater and necromass enrichment compounds that links compound identity, pathways, ENIGMA-isolate utilizers, GTDB placement and environmental occurrence, and finds 74 of the 83 compounds organism-dark.Carbon Source Utilization Predicts Ecology and Lifestyle in PseudomonasSummary of a GapMind carbon-pathway analysis of 12,732 Pseudomonas genomes showing strong plant-sugar pathway loss in the P. aeruginosa group and a significant but modest ecological signal in carbon profiles among free-living species.Truly Dark Genes — What Remains Unknown After Modern Annotation?Summary of the truly_dark_genes project, which separates Fitness Browser dark genes into annotation-lag, truly dark and unlinked sets after Bakta reannotation, characterizes the truly dark genes' structural, evolutionary and phenotypic properties, and ranks candidates for experimental characterization.Within-Species AMR Strain VariationWithin-species antimicrobial-resistance gene variation across 1,305 bacterial species and 180,025 genomes, covering prevalence classes, resistance islands, phylogenetic signal, AMR ecotypes, temporal nulls and environment associations.Community Metabolic Ecology via NMDC × Pangenome IntegrationSummary of a project integrating NMDC community taxonomy and metabolomics with GTDB pangenome GapMind pathway completeness to test community-scale Black Queen dynamics and Soil–Freshwater differentiation of metabolic potential.Conservation vs Fitness — Linking FB Genes to Pangenome ClustersSummary of the conservation_vs_fitness project, which links Fitness Browser essentiality calls to KBase pangenome clusters across 33 bacteria and finds a modest enrichment of putative essential genes in core clusters.BERDL Data Atlas — Inventory, Topic Map, and Cross-Reference SynergiesSummary of the BERDL Data Atlas project, which inventories the KBase Data Lakehouse's tables, agency and topic coverage, cross-tenant join bridges, realized project use, and a sample-validated FitnessBrowser-to-AlphaFold join.Gene Conservation, Fitness, and the Architecture of Bacterial GenomesSummary of a synthesis that combines RB-TnSeq fitness and pangenome conservation across 43 organisms, covering a modest conservation gradient, a core-genome burden paradox, selection signatures, and core-enriched fitness modules.PGP Gene Distribution Across Environments & PangenomesSummary of a KBase Data Lakehouse pangenome analysis of plant-growth-promoting gene co-occurrence, soil/rhizosphere enrichment, core/accessory status and tryptophan-pathway coupling to ipdC.Self-Sufficiency, Anaerobic Toolkit, and Cultivation Bias in Clay-Confined Cultured Bacterial GenomesSummary of the clay_confined_subsurface project, which tests biosynthetic self-sufficiency, an anaerobic toolkit, and porewater-versus-rock-attached cultivation bias in cultured deep-clay, shallow-clay, and soil-baseline bacterial genomes.Metabolic Consistency of Pseudomonas FW300-N2E3 Across Four KBase Data Lakehouse DatabasesCross-database consistency of Pseudomonas FW300-N2E3 metabolism across Web of Microbes exometabolomics, Fitness Browser gene fitness, BacDive species-level utilization and GapMind pathway predictions, with sample-size-aware grading of discordances.Condition-Specific Respiratory Chain Wiring in ADP1Summary of the respiratory_chain_wiring project, which reports that carbon source selects qualitatively different respiratory-chain configurations in Acinetobacter baylyi ADP1, proposes an NADH flux-rate explanation, and records a null cross-species NDH-2 compensation test and standard-condition proteomics.Counter Ion Effects on Metal Fitness MeasurementsSummary of the counter_ion_effects project, which tested whether chloride from metal salts confounds genome-wide metal-fitness measurements and found substantial metal–NaCl shared-stress overlap but no chloride dose effect, with Metal Fitness Atlas core enrichment robust to correction.Fitness Cost of Antimicrobial Resistance GenesSummary of a Fitness Browser RB-TnSeq meta-analysis across 25 organisms that finds a small relative baseline fitness cost for antimicrobial-resistance genes, independent of mechanism and core/accessory status, with mechanism-dependent importance under antibiotic exposure.Acinetobacter baylyi ADP1 Data ExplorerSummary of the Acinetobacter baylyi ADP1 Data Explorer project, which analyzes a multi-omics SQLite database for ADP1 and 13 related genomes and links it to the KBase Data Lakehouse pangenome, biochemistry and fitness collections.Prophage Gene Modules and Terminase-Defined Lineages Across Bacterial Phylogeny and Environmental GradientsSummary of the prophage_ecology project, which maps eggNOG-annotated prophage gene modules and TerL lineages across 27,702 bacterial species and tests their environmental, phylogenetic, genome-size and NMDC metagenomic associations.Essential Metabolome: GapMind Pathway Analysis Across Essential-Gene OrganismsPilot GapMind analysis of amino-acid biosynthesis and carbon-source pathway completeness across 7 Fitness Browser organisms with essential-gene data, including an apparent Desulfovibrio vulgaris serine gap and E. coli coverage loss.Lanthanide Methylotrophy Atlas — SummaryA 293,059-genome atlas of lanthanide-dependent methanol oxidation in the KBase Data Lakehouse pangenome, covering xoxF predominance over mxaF, clade-restricted lanmodulin, soil/sediment enrichment, REE-AMD MAGs, PQQ annotation gaps and eggNOG-versus-Bakta marker calibration.Aromatic Catabolism Support Network in ADP1Summary of the aromatic_catabolism_network project, which defines a 51-gene quinate-specific support network around the β-ketoadipate pathway in Acinetobacter baylyi ADP1 spanning aromatic degradation, Complex I, iron acquisition, PQQ biosynthesis and regulation.BacDive Isolation Environment × Metal Tolerance PredictionSummary of the bacdive_metal_validation project, which tested whether BacDive isolation environments independently support genome-based metal-tolerance scores from GTDB pangenome species.BacDive Phenotype Signatures of Metal ToleranceSummary of a project testing whether classical BacDive microbiology phenotypes predict genome-based Metal Fitness Atlas metal tolerance scores independently of phylogeny.Core Gene Paradox — Why Are Core Genes More Burdensome?Summary of the core_gene_tradeoffs project, which links Fitness Browser RB-TnSeq fitness burden to pangenome conservation and finds a function-specific core-gene burden paradox, enrichment of condition-dependent trade-off genes among core genes, and a costly-versus-conserved selection-signature matrix.Soil Metal Concentrations Drive Functional Gene ShiftsPreliminary analysis of associations between nine soil metal concentrations and microbial COG functional gene content across 51,748 samples, with its confounding, effect-size and spatial-validation caveats.ADP1 Deletion Collection Phenotype AnalysisAnalysis of Acinetobacter baylyi ADP1 single-gene deletion growth across 8 carbon sources, integrated with TnSeq essentiality, functional annotation and pangenome status, showing a continuous condition-dependent phenotype gradient with one discrete quinate module.Environmental Resistome at Pangenome ScalePangenome-scale analysis of how antimicrobial-resistance gene-cluster counts, core/accessory composition and resistance mechanisms vary with ecological environment across 14,723 bacterial species, with its caveats, literature context and proposed follow-up work.Pan-bacterial Fitness Modules via Independent Component AnalysisSummary of the fitness_modules project, which used independent component analysis on RB-TnSeq fitness data from 32 bacteria to find co-regulated fitness modules, align them into cross-organism module families, and give hypothetical proteins process-level function context.Subsurface Bacillota_B Specialization — What Distinguishes Deep-Clay Lineages from Soil Congeners?Summary of a comparison of 10 deep-clay and 62 soil-baseline Bacillota_B genomes that finds 547 anchor-enriched orthologous groups and larger deep-clay genomes, and that corrects the clay project's iron-reduction marker analysis.Lab Fitness Predicts Field Ecology at Oak RidgeProject report testing whether Fitness Browser laboratory metal-tolerance fitness predicts genus abundance and community composition across a uranium gradient in 108 Oak Ridge groundwater sites.Polyhydroxybutyrate Granule Formation Pathways — Distribution Across Clades and Environmental SelectionSummary of the phb_granule_ecology project, which surveys polyhydroxybutyrate (PHB) pathway distribution across 27,690 GTDB species and tests environmental selection, genome-size confounding, and horizontal transfer of phaC.Regulatory and proteomic architecture of Δfur-permitted lipid A loss in *Caulobacter crescentus*Integration of transcriptomics, single-replicate outer-membrane proteomics, RB-TnSeq fitness, published regulons and comparative annotation explaining how Δfur ΔsspB permits lipid A loss in *Caulobacter crescentus*.Web of Microbes Data ExplorerSummary of the Web of Microbes Data Explorer project, which characterized a 2018 Web of Microbes exometabolomics snapshot and assessed its links to Fitness Browser, ModelSEED, GapMind and pangenome collections.COG Functional Category AnalysisCross-species analysis of COG functional categories across 32 bacterial species that contrasts conserved core-gene functions with mobile-, defense- and unknown-function enrichment among novel or singleton genes.Pangenome Openness AnalysisSummary of the pangenome_openness project, which found no significant correlation between species pangenome openness and environment or phylogeny effects on gene content.Prophage-AMR Co-mobilization AtlasPangenome-scale analysis of how antimicrobial-resistance genes co-localize with prophage markers across GTDB species, finding a strong species-level link between prophage density and AMR repertoire breadth but only a modest, heterogeneous gene-level proximity effect.The 5,526 Costly + Dispensable GenesCross-organism characterization of 5,526 bacterial genes that are costly in laboratory fitness assays yet dispensable in the pangenome, showing they are dominated by mobile genetic elements and poorly annotated, taxonomically restricted DNA that the report interprets as horizontally acquired.Triple Essentiality Concordance Analysis for *Acinetobacter baylyi* ADP1Summary of a two-stage Acinetobacter baylyi ADP1 analysis comparing FBA, RB-TnSeq, knockout, mutant growth and proteomics evidence for gene essentiality, including null FBA–growth results and systematic TnSeq–knockout discordance.Ecotype Functional DifferentiationSummary of a 15-species KBase Data Lakehouse pangenome study finding that within-species gene-content ecotypes are common and differ in COG functional categories, with larger effect sizes in adaptive than housekeeping categories.AlphaFold MSA Depth as a Lens on the Bacterial Annotation GapSummary of a project that used AlphaFold MSA depth as a proxy for functional annotation richness across bacterial pangenome gene clusters, covering the core/accessory gradient and conserved low-MSA-depth core 'paradox' proteins.Co-fitness and Co-inheritance in Bacterial PangenomesSummary of the cofitness_coinheritance project, which tested whether Fitness Browser gene co-fitness predicts gene co-occurrence across bacterial pangenomes at pairwise and ICA-module levels.SSO Subsurface Community Ecology — Spatial Structure, Functional Gradients, and Hydrogeological DriversSummary of the ENIGMA SSO subsurface 16S ASV study testing spatial structure, hydrogeological depth zonation, inferred biogeochemical gradients, groundwater–sediment differentiation, guild co-occurrence and short-term temporal stability across a 3×3 well grid.Fitness Modules × Pangenome ConservationSummary of the module_conservation project, which tests whether genes in ICA fitness modules and cross-organism module families are enriched in the bacterial core genome.Pan-Bacterial Anti-Phage Defense ArsenalSummary of the phage_defense_arsenal report, which surveys seven anti-phage defense-system families across the KBase Data Lakehouse pangenome and covers prevalence, the prophage-burden association, co-occurrence syndromes, accessory-pangenome enrichment and detection caveats.The Pan-Bacterial Essential GenomeSummary of the essential_genome project, which uses RB-TnSeq essentiality calls from the Fitness Browser, bidirectional-best-hit orthology, pangenome conservation links and ICA fitness modules across 48 organisms to classify universally, variably and orphan essential gene families and predict functions for hypothetical essentials.Contamination Gradient vs Functional Potential in ENIGMA CommunitiesSummary of an ENIGMA project linking groundwater metal-contamination gradients to genus-bridged pangenome functional proxies, in which confirmatory contamination–defense tests were null and exploratory coverage-aware associations were sensitive to coverage, covariates, taxonomic resolution and FDR correction.AlphaEarth Embeddings, Geography & Environment ExplorerCharacterization of 64-dimensional AlphaEarth satellite-derived environmental embeddings for 83,287 pangenome genomes, covering geographic distance-decay, human-associated sampling bias, coordinate quality, metadata harmonization, and UMAP/DBSCAN structure.AMR Co-Fitness Support NetworksPan-bacterial analysis of antimicrobial-resistance gene cofitness support networks across 28 Fitness Browser organisms, covering ICA module membership, InterProScan GO enrichment, organism-versus-mechanism specificity, and the unresolved split between co-regulation and shared dispensability.Harvard Forest Long-Term Warming — DNA vs RNA Functional ResponseSummary of a 25-year +5°C Harvard Forest soil-warming study comparing DNA- and RNA-pool functional responses, community composition, carbon-cycling genes and metabolite richness across organic and mineral horizons.PaperBLAST Data Explorer — Literature Coverage Bias in Protein Sequence SpaceSummary of the paperblast_explorer project, which quantifies how unevenly literature covers organisms, genes and MMseqs2 protein families in the KBase Data Lakehouse PaperBLAST collection and identifies dark protein families.Lignin Enrichment and Ecological Memory in Microbial CommunitiesSummary of a 16S and ITS amplicon study of how lignin enrichment, labile-carbon co-supplementation and sequential passaging restructure bacterial and fungal communities, including a quantified bacterial ecological-memory effect.Metadata Correlates of Eukaryotic Contamination in NMDC Prokaryote-Targeted MetagenomesSummary of a project that quantified eukaryotic read fractions in 2,759 NMDC prokaryote-targeted metagenome runs and showed that cross-study environmental correlates of eukaryotic signal are confounded with study batch, while vegetation and geography predict it within one batch-controlled soil study.NMDC Context AuditAudit of the KBase Data Lakehouse resources labeled `nmdc`, showing that 20 database names resolve to 7 maintained resources that differ in provenance, tenant placement, scale, currency and authority.Soil Microbial Dark Matter and the Clay Shield Null ResultPreliminary global soil analysis of 5,441 samples finding no support for the clay shield hypothesis, and introducing a Genomic Discovery Index that flags forest, cropland and alkaline soils as genomically under-represented.T4SS–CAZy Environmental HGTPreliminary analysis of type IV secretion system (T4SS) machinery in environmental MAGs, its genomic co-localization with carbohydrate-active enzyme (CAZy) genes, cross-phylum horizontal transfer of GT2 glycosyltransferases, and an association with metal-resistance genes.Global Biogeography of Environmental Bacterial Metal ResistancePreliminary global map of metal-resistance prevalence in 22,356 geolocated environmental MGnify MAGs, with 5° grid hotspots, biome enrichment and depletion, and ENA coordinate-coverage limits.