GC low

Deep research

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

# Curation report: microbial trait **GC low**

## Executive curation recommendation

The identifier must be recorded verbatim as **`METPO:1000429`**, with parent **`METPO:1000127`** and synonym **GC_42.65_57.0**. Despite the upstream label “GC low,” the supplied authoritative definition is a **mid-range whole-genome GC-content bin of approximately 42.65–57.0%**. It is therefore an assay/computational genomic-composition class, not a physiological capacity or environmental preference.

The best-supported TraitMech graph is not a direct pathway ending uniquely in this interval. Rather, it is an evolutionary chain:

**DNA damage and replication errors → DNA-repair/replication machinery → mutation spectrum → fixation through drift, selection, and recombination/gene conversion → continuous genome-wide GC content → bin assignment to `METPO:1000429`.**

Direct experimental evidence is strongest for oxidative lesions, cytosine deamination, and corresponding base-excision/very-short-patch repair systems. Evidence connecting environmental conditions, polymerase inventories, or recombination directly to this particular numeric bin is comparative and should be marked uncertain.

## 1. Trait scope and boundaries

### Operational definition

Genome-wide GC content is the fraction of genomic DNA bases that are guanine or cytosine, normally computed as `(G+C)/(A+T+G+C) × 100`. For this trait, the measured value is classified as `METPO:1000429` when it lies approximately between **42.65% and 57.0%**. Endpoint inclusion should follow the METPO implementation; it should not be inferred from the rounded prose definition.

Across bacteria, reported genome-wide GC values span approximately 13–77%. A recent analysis of **11,083 representative bacterial genomes** reported a 16–77% range and a phylogenetically constrained bimodal distribution, with peaks below 45% and above 60%. Thus, the METPO interval occupies much of the intermediate region between those modes rather than a universally “low-GC” state. More than 60% of GC variance in that study was explained at phylum level, demonstrating strong phylogenetic inertia. (teng2023genomiclegaciesof pages 1-2)

### Boundary cases to exclude

* **True low-GC/AT-rich genomes:** values below approximately 42.65% are outside this class, even though the textual label might suggest otherwise.
* **Values above 57.0%:** these are also outside the class.
* **Local GC content:** a genomic island, gene, codon position, or sliding window may lie in the interval while the whole genome does not.
* **GC3:** GC at third codon positions is strongly affected by synonymous codon usage and is not equivalent to whole-genome GC.
* **GC skew:** `(G−C)/(G+C)` measures strand asymmetry, not total GC fraction.
* **Equilibrium GC:** a value inferred from a mutation spectrum is an evolutionary expectation and need not equal observed GC.
* **Assembly artifacts:** contamination, incomplete metagenome-assembled genomes, untrimmed plasmids, and biased sequencing can shift the calculated value.
* **Within-genome heterogeneity:** horizontally acquired and accessory regions may differ from the core genome; bin assignment should use the declared whole-genome measurement protocol.

## 2. Current mechanistic understanding

Mutation is broadly biased toward AT in bacteria, but observed genomes frequently contain more GC than mutation bias alone predicts. Consequently, present understanding invokes several interacting forces: mutation generated by replication and DNA damage; repair-system specificity; selection on coding and regulatory functions; genetic drift; homologous recombination and possible GC-biased gene conversion; and lineage history. A review reports mutation-accumulation rates of approximately **0.001 mutations/genome/generation in *Escherichia coli*** and **0.008 in *Mesoplasma florum***, illustrating that both mutation rate and spectrum are lineage dependent. Mismatch-repair-deficient *E. coli* can also reverse the wild-type direction of mutation bias, indicating that repair machinery is mechanistically upstream of long-term nucleotide composition. (hershberg2015mutation—theengineof pages 6-7)

The most recent directly relevant large-scale synthesis is Teng et al. (February 2023). Its analysis of 11,083 genomes supports a model in which ancient environmental adaptation changed DNA replication and repair inventories, whose resulting mutation biases subsequently shaped GC evolution. Associated modules included base-excision repair, nucleotide-excision repair, mismatch repair, homologous recombination, nonhomologous end joining, and translesion synthesis. This is an **indirect-selection model**, not evidence that contemporary environments rapidly force genomes into a defined GC bin. (teng2023genomiclegaciesof pages 8-10, teng2023genomiclegaciesof pages 1-2, teng2023genomiclegaciesof pages 10-12)

## 3. Candidate causal-graph nodes

### Trait and measurement nodes

* **Genome-wide GC-content phenotype:** `METPO:1000429`
* **Parent genomic-composition trait:** `METPO:1000127`
* **Continuous genome-wide GC percentage:** label-only candidate; retain the numerical value and calculation method as evidence metadata.
* **GC_42.65_57.0 binning process:** label-only assay/computational node.
* **AT-biased mutation spectrum**, **GC-biased mutation spectrum**, **GC→TA transversion**, and **GC→AT transition:** label-only molecular-event nodes.

### Genes, proteins, and complexes

Taxon-independent gene symbols should remain label-only until a species-specific locus or protein accession is selected.

* **mutM** — formamidopyrimidine-DNA glycosylase; removes 8-oxoG paired with C.
* **mutY** — adenine DNA glycosylase; removes A opposite 8-oxoG.
* **mutT** — oxidized-purine nucleotide sanitizer; hydrolyses 8-oxo-dGTP and can affect mutation direction.
* **ung** — uracil-DNA glycosylase.
* **mug** — mismatch-specific uracil-DNA glycosylase.
* **vsr** — very-short-patch repair endonuclease.
* **dnaE2** — error-prone/SOS-associated DNA polymerase III alpha-subunit homolog.
* **polC** — replicative DNA polymerase III alpha subunit in specific bacterial lineages.
* **Pol V** and **DinB/Pol IV** — translesion polymerases; label-only pending taxon-specific grounding.

Showing the first 60 of 207 lines of findings; the linked file also carries the run's front matter and the prompt it was given — read the full report.