phenotype

Deep research

Generated by just research-trait; source: research/traits/upper/phenotype-deep-research-falcon.md

# Curation report: microbial phenotype (`METPO:1000059`)

## Executive curation recommendation

`METPO:1000059` should remain a **broad upper-level phenotype class**, not be assigned one universal biochemical pathway. Its supplied definition—“A quality that differentiates specific instances of a species from other instances of the same species”—is compatible with entity–quality (EQ) modeling: a microbial entity, structure, or process **bears a quality** whose state differs among instances. Contemporary phenotype modeling likewise treats phenotype as an outcome of genetic endowment interacting with environmental conditions over time. (thessen2020transformingthestudy pages 8-11, thessen2020transformingthestudy pages 11-12, thessen2020transformingthestudy pages 1-2)

For `data/traits/upper/phenotype.yaml`, the safest graph is therefore a small schema-level graph connecting **biological entity → quality → phenotype representation**, with genotype, environment, and time/context as contributors. Specific genes, nutrients, pathways, drugs, and instruments belong in contextual exemplar subgraphs or narrower child traits, not as universal causes of phenotype.

## 1. Scope and boundaries

### Intended scope

The target denotes an **observable or inferable quality of a microbial instance**—cell, strain, isolate, or population—that distinguishes it from another instance of the same species. Examples include cell morphology, colony morphology, growth rate, substrate-utilization state, metabolite-production level, antimicrobial susceptibility, stress robustness, motility, and a differentiated cell state. EQ models can use anatomical entities, biological processes, or physiological entities, and can represent qualitative states or be extended with quantitative measurements. (thessen2020transformingthestudy pages 2-4, thessen2020transformingthestudy pages 8-11)

At the class level, `METPO:1000059` identifies the kind of quality. At the assertion level, a curator should record the bearer, quality/state, organism or strain, environment, assay, time point, and—when quantitative—the value and unit. This reflects the distinction between terminological classes and instance-specific assertions and the need to retain microbial origin and treatment metadata. (thessen2020transformingthestudy pages 7-8)

### Nearby concepts that should remain distinct

- **Trait/character:** the dimension being compared, such as growth rate or cell length. A **phenotype** is the state manifested by an instance, such as reduced growth rate or elongated cells.
- **Physiological capacity:** a potential ability, such as growth using cellobiose. It becomes an observed phenotype only under stated conditions and an appropriate assay.
- **Environmental preference:** a pattern of comparatively better occurrence or performance across environments. Presence in one environment does not by itself establish preference.
- **Assay result or measurement value:** an information artifact about a phenotype, not the phenotype itself. Measurement models should preserve who, how, when, and where the observation was made. (thessen2020transformingthestudy pages 5-7, thessen2020transformingthestudy pages 12-14)
- **Fitness:** performance or reproductive success in a specified environment. In yeast phenomics, fitness was explicitly distinguished from robustness. (trivellin2024robustnessquantificationof pages 1-2)
- **Robustness:** stability of one or more phenotypes across perturbations, rather than the value of a phenotype in one environment. It can be high even when absolute fitness is low. (trivellin2024robustnessquantificationof pages 1-2, trivellin2024robustnessquantificationof pages 10-11)
- **Genotype or gene presence:** a possible determinant or predictor, not itself a phenotype. Incomplete penetrance and environmental dependence preclude automatically converting genotype associations into phenotype-causation edges. (yu2024decipheringcomplexantibiotic pages 1-2)

### Boundary cases

Population-average phenotypes can conceal distinct states among genetically identical cells. A 2024 review notes that bacterial populations often contain mixtures of phenotypic responses and identifies transcriptional heterogeneity as a principal driver of distinct cell states. Consequently, population, subpopulation, and single-cell phenotypes should not be merged without an aggregation qualifier. (walls2024bacterialphenotypicheterogeneity pages 1-2)

Similarly, morphology, Raman spectrum, optical density, colony size, and MIC are not interchangeable. Morphology is a quality; a Raman spectrum is an assay output reflecting biomolecular composition; colony size is a measurement proxy; and MIC is an operational susceptibility endpoint defined by a protocol.

## 2. Candidate nodes grouped by type

### Core ontology/schema nodes

| Node | Suggested grounding | Curation role |
|---|---|---|
| phenotype | `METPO:1000059` | Target class; quote exactly in YAML. |
| parent trait | `METPO:1000188` | Supplied parent; retain unless ontology review indicates otherwise. |
| biological entity / microbial bearer | Label-only unless the bearer is specified | The cell, strain, isolate, population, structure, or process bearing the quality. |
| quality | PATO candidate; exact child CURIE should be selected for each concrete quality | EQ quality component. Do not force one PATO term onto the upper class. |
| phenotype observation | Label-only candidate | Observation event linking bearer, quality, assay, and context. |
| measurement datum | Label-only candidate | Value/unit output; distinct from phenotype. |
| genotype | Label-only candidate | Contributing biological context. |
| environment | ENVO candidate, selected per experiment | Contributing context; use precise medium, habitat, temperature, oxygen, pH, or exposure nodes when known. |
| developmental/growth phase and time point | Label-only candidates | Essential temporal context. |

### Genes, proteins, and complexes

- **MET28** and the **CBF1–MET4–MET28 transcription-factor complex**: yeast-specific candidates regulating sulfur metabolism. The 2024 study found that `met28` deletion produced the largest robustness increase but substantially reduced fitness; this is a trade-off, not a universally beneficial effect. (trivellin2024robustnessquantificationof pages 10-11)
- **TIR3, WWM1, BCH1**: candidate yeast robustness modifiers, but their effects varied with perturbation space and some mechanistic assignments remain incomplete. (trivellin2024robustnessquantificationof pages 1-2, trivellin2024robustnessquantificationof pages 10-11)
- **AMR-associated variants/loci**: use only when a study establishes the exact organism, allele, drug, and assay. The retrieved *H. pylori* study shows that multivariant models can outperform individual known sites, warning against simplistic single-locus causation. (yu2024decipheringcomplexantibiotic pages 1-2)

No UniProt accession should be assigned without specifying species and protein record; gene symbols alone are safer at this upper level.

### Pathways, modules, and biological processes

- Gene expression and transcriptional regulation.
- Transcriptional heterogeneity and cell-state differentiation.

Showing the first 60 of 233 lines of findings; the linked file also carries the run's front matter and the prompt it was given — read the full report.