pangenome openness

Deep research

Generated by just research-trait; source: research/traits/genomics/pangenome_openness-deep-research-falcon.md

# TraitMech curation report: pangenome openness

## Trait record and scope

- **Trait label:** pangenome openness
- **Trait identifier:** `traitmech:000102`
- **Category / kind / status:** GENOMICS / CLASS / REVIEWED
- **Parent:** `METPO:1000188`
- **Recommended operational definition:** A population- or taxon-level genomic property describing whether the cumulative number of nonredundant gene families continues to increase as additional genomes are sampled, and therefore the balance between conserved core genes and variably distributed accessory genes. An **open** pangenome continues to yield novel gene families; a **closed** pangenome approaches saturation. Core genes occur in all—or, operationally, nearly all—sampled genomes; accessory genes occur in a subset, and strain-specific/singleton genes occur in one sampled genome. Thresholds such as 95% for “core” are conventions rather than biological invariants. (medini2020thepangenomea pages 7-10, tonkinhill2023challengesinprokaryote pages 4-6)

Openness is usually inferred from randomized genome-accumulation/rarefaction curves or a Heaps-law model. Under the convention used by Pardeshi et al., a decay exponent **α < 1** denotes an open pangenome and **α > 1** a closed one. However, notation differs among implementations, so the equation, fitted parameter, confidence interval, number of genomes, taxonomic scope, and software must accompany any value. (medini2020thepangenomea pages 7-10, tonkinhill2023challengesinprokaryote pages 4-6, pardeshi2024pangenomicstounderstand pages 3-7)

### What the trait is—and is not

1. **It is a property of a delimited collection and model, not an intrinsic binary state observable in one cell.** Only the pangenome parameters of the sampled genomes can be estimated; unsampled lineages, population structure, and taxonomic boundaries affect the result. (tonkinhill2023challengesinprokaryote pages 4-6)
2. **It is not genome size.** A species may have similarly sized individual genomes yet a large collective accessory pool maintained by turnover.
3. **It is not simply accessory-genome fraction.** Accessory fraction describes the current sample; openness describes the expected accumulation of new families with further sampling.
4. **It is not identical to pangenome fluidity.** Fluidity is the mean proportion of genes not shared by pairs of genomes. It is correlated with pangenome variability and is useful comparative evidence, but it does not directly measure asymptotic gene-family accumulation. Dewar et al. observed fluidities from 0.012 to 0.41 across 126 species. (dewar2024bacteriallifestyleshapes pages 2-3, dewar2024bacteriallifestyleshapes media 30d7ccd4)
5. **It is not nucleotide diversity, recombination rate, HGT rate, genome plasticity, or taxonomic diversity**, although each may influence or correlate with it.
6. **Species-level and genus-level pangenomes must not be merged.** For example, the 2024 *Pectobacterium* result is primarily a genus-level graph over 22 species, with separate species-level sub-pangenomes. (pardeshi2024pangenomicstounderstand pages 3-7)

## Current mechanistic understanding

The proximal biological determinants are **gene-family gain and loss across lineages**. HGT through transformation, conjugation, transduction, and mobile genetic elements supplies gene gains; deletion and lineage-specific loss remove genes. Selection, drift, ecological exposure, and barriers to exchange determine which gains persist. Tonkin-Hill et al. therefore recommend gain/loss-rate models as a more mechanistic alternative to treating open/closed status as a simple binary. (tonkinhill2023challengesinprokaryote pages 4-6)

The strongest recent cross-species result is Dewar et al. (May 2024), a phylogenetically controlled analysis of 126 bacterial species. Free-living species had higher fluidity than host-associated species (**R²=0.139, pMCMC=0.004**). Among 115 host-associated species, facultative host reliance, extracellular location, mutualism, and motility jointly explained **25.7%** of fluidity variation; all five lifestyle variables explained **29.9%** across 119 species. The authors interpret variable lifestyles as increasing exposure to novel genes and/or selection for niche-specific gains and losses, but could not separate adaptive from neutral contributions. These are strong comparative associations, not experimental proof that any lifestyle state directly changes Heaps-law openness. (dewar2024bacteriallifestyleshapes pages 2-3, dewar2024bacteriallifestyleshapes pages 5-5, dewar2024bacteriallifestyleshapes pages 5-7, dewar2024bacteriallifestyleshapes media 20431c56)

Selection also structures which accessory genes coexist. Across 40 *Pseudomonas* species, **86.7% of common accessory genes** participated in significant co-occurrence or avoidance relationships; non-vertically inherited coincident genes were more likely to share functions, be co-transcribed, and encode interacting proteins. This supports selection on functional combinations, but not a simple claim that selection always increases openness. (whelan2021evidenceforselection pages 1-2)

## Candidate nodes

### Trait and measurable genomic-state nodes

| Candidate node | Grounding | Curation note |
|---|---|---|
| pangenome openness | `traitmech:000102` | Target node; retain identifier verbatim. |
| pangenome | `METPO:1000188` only if this parent denotes the intended pangenome concept | Verify the parent’s label before reuse as an entity node. |
| core genome / core gene family | Label only | Prevalence threshold is assay-dependent. |
| accessory genome / accessory gene family | Label only | Includes shell/cloud partitions in some pipelines. |
| singleton or strain-specific gene family | Label only | Highly sensitive to errors and sample size. |
| pangenome fluidity | Label only | Comparative proxy, not synonymous with openness. |
| gene-family accumulation curve | Label only | Assay/output node. |
| Heaps-law exponent | Label only | Store equation and convention with the value. |
| gene gain rate; gene loss rate | Label only | Prefer these as proximal mechanistic nodes. |

### Biological processes and molecular mechanisms

| Candidate node | Suggested grounding | Role |
|---|---|---|
| horizontal gene transfer | `GO:0042710` | Umbrella process supplying nonvertical gene gains; verify ontology version. |
| natural transformation | `GO:0009294` | Uptake and incorporation of exogenous DNA; verify label/version. |
| DNA recombination | `GO:0006310` | Incorporation/reshuffling of imported DNA. |
| homologous recombination | `GO:0035825` | Candidate child process; verify before YAML insertion. |
| conjugation | Label only pending ontology verification | Plasmid/ICE-mediated transfer. |
| transduction | Label only pending ontology verification | Phage-mediated transfer. |
| DNA integration | `GO:0015074` may be relevant | Verify that the intended ontology meaning fits genomic integration. |
| transposition | `GO:0032196` | Movement of transposable elements; verify version. |
| gene deletion / gene loss | Label only | Proximal process reducing retained repertoire. |
| natural competence | Label only | Cell state enabling transformation. |
| CSP–ComD/ComE competence signaling | Taxon-specific, label only | Pneumococcal regulatory module. |

Showing the first 60 of 213 lines of findings; the linked file also carries the run's front matter and the prompt it was given — read the full report.