Talk:Overdominance

Talk

Discussion for: Overdominance • View • Links • Related changes

Use this page for verification notes, disputes, and action items. Keep narrative content on the article.

Page scope

Overdominance covers heterozygote advantage at one locus: the two scales it is measured on, how selection holds an overdominant polymorphism, the rise and decline of overdominance as a general explanation, cases measured at single genes, patterns that imitate it, and what cannabis studies report. The superiority of hybrids over their parents belongs to Heterosis, the relation between two alleles and the scale of dominance to Dominance and recessivity, loss of fitness on inbreeding to Inbreeding depression, other selection that keeps alleles to Balancing selection and the measure of heterozygotes in a population to Heterozygosity.

Data verification

Record source conflicts, figures that differ between studies and sources not used here. Keep narrative content on the article.

Discussion

Sources

Felsenstein (2019), Theoretical Evolutionary Genetics, pp. 85, 88–92, 99–101, 104–105, 242–244, 348 and 411–414: the fitness parameters, equilibrium, protected polymorphism, history, segregational load, the sickle-cell fitnesses, inbreeding, finite populations, and the word and its separation from heterosis.
Delph and Kelly (2014), New Phytologist: the history of the idea in plants, the evidence against it as a cause of inbreeding depression, simple and emergent overdominance, and protected polymorphism against frequency dependence. Charlesworth (2006), PLoS Genetics: overdominance and balancing selection.
Hedrick (2012), Trends in Ecology & Evolution, and Sellis and colleagues (2011), PNAS: how common heterozygote advantage is.
Krieger and colleagues (2010), Nature Genetics, and Jiang and colleagues (2013), PLoS Genetics: SFT in tomato. Kellenberger and colleagues (2019), Nature Communications: Gymnadenia rhellicani. Luzzatto (2012), Mediterranean Journal of Hematology and Infectious Diseases, and Piel and colleagues (2010), Nature Communications: sickle-cell haemoglobin and malaria.
Mackay and colleagues (2021), Plant Biotechnology Journal; Li and colleagues (2015), PNAS; Larièpe and colleagues (2012), Genetics; Ohta (1971), Genetical Research; Zhao and Charlesworth (2016), Genetics: patterns that imitate overdominance.
Woods and colleagues (2021), Genetics, with its supplement: the Carmagnola × USO31 F₂. Mishchenko and colleagues (2022), Taurian Scientific Herald 125, in Ukrainian: the coefficient h_p. Woods and colleagues (2023), G3, and Lynch and colleagues (2025), Nature: genome scans and the inversion.

Where the sources disagree

How common it is. Hedrick holds that loci with heterozygote advantage are a small minority of loci and Felsenstein calls overdominance very rare in nature. Sellis and colleagues expect it to be frequent, and short-lived, during adaptation in diploids. Their result is a model, and they say that the abundance of incomplete sweeps they point to is consistent with it without predicting it uniquely. Kellenberger and colleagues describe the known cases as mostly short-lived. Delph and Kelly point out that the experiments behind the low estimates mostly exclude environmental variation, so they do not speak to emergent overdominance.
What counts as overdominance. Kellenberger and colleagues and Delph and Kelly define it by fitness. Beavis and colleagues define it on a scale of genotypic values for any character, and Krieger and colleagues speak of yield overdominance. Mishchenko and colleagues use it for a coefficient computed from means of a cross.
Who proposed pseudo-overdominance. Mackay and colleagues and Li and colleagues date the idea to 1917, Mackay and colleagues citing Jones. Delph and Kelly define it through recessive deleterious alleles at two linked genes. Li and colleagues and Mackay and colleagues describe dominant favourable alleles dispersed between the parents. The two descriptions are the same arrangement seen from opposite alleles.
Maize chromosome 3. Larièpe and colleagues found a grain yield region on bin 3.05 that seemed overdominant, as had Lu and colleagues (2003) and Frascaroli and colleagues (2007) in other material. Pea and colleagues (2009) found complete dominance in near-isogenic lines. Their account of it comes through Larièpe and colleagues.
The sickle-cell fitnesses. Felsenstein's values, 0.685, 0.763 and 0.151, include a third allele, haemoglobin C, in his full example. The table takes the two-allele side of it. The calculated equilibrium of 0.113 for the S allele and the load of 0.091 reproduce his equilibrium and his mean fitness of 0.6938.

How far the cannabis evidence goes

TSM.2 comes from one F₂ family in one field season, Fort Collins in 2017, with one clone of each F₂ line in each of three replicate blocks. Woods and colleagues call the pattern suggestive. Their Supplementary Figure 4 plots quantile-normalised values, so the +0.33, −0.33 and −0.71 are read from a chart on that scale and are not grams.
The measure at TSM.2 is seed mass. The cannabis studies give no fitness for genotypes at a locus.
The two homozygotes at TSM.2 are both below the heterozygote, and Supplementary Table 3 marks both parental homozygotes with a minus sign. The F₁ mean equals that of Carmagnola, 17.2 g, with three clones of each parent and of the F₁ in each block. The other three thousand seed mass QTL favour the Carmagnola allele.
Does TSM.2 belong to the cluster LG5.05? Woods and colleagues do not say so of TSM.2 by name. They say that each agronomic trait model contains a QTL within LG5.05, and their Table 1 lists TSM.2 as the only thousand seed mass QTL on linkage group 5.
Whether one gene or several linked genes produce the pattern at TSM.2 is open. Only one of the 34 agronomic QTL, SB.3, has a narrow interval of about 50,000 base pairs. The TINY SNP lies in that interval and is a hypothesis of the authors for LG5.05.
Woods and colleagues (2021) state that earlier work on cannabinoid inheritance shows only dominance and additivity.
The coefficient of Mishchenko and colleagues describes where a hybrid's mean falls against the means of its parents. A value above 1 says nothing direct about overdominance at any locus.
Woods and colleagues (2023) offer heterozygote advantage as a hypothesis from Tajima's D and heterozygosity. Lynch and colleagues write "perhaps" of the inversion. Cai and colleagues (2026, Horticulture Research) suggest balancing selection behind the two haplotypes of one wild plant, without naming overdominance.
García-de Heer and colleagues (2026) describe overdominance among the general causes of inbreeding depression. Their selfed lines do not separate it from dominance.

Terms

Heterozygote advantage and overdominance are used here as one term, the use of Kellenberger and colleagues and of Delph and Kelly. Felsenstein and Charlesworth warn against two neighbouring confusions, with heterosis and with balancing selection.
Load in the table is the fraction by which segregation lowers mean fitness below that of the heterozygote, st/(s + t). Felsenstein stresses that it is relative and need not be a burden.
Thousand seed mass is the term of Woods and colleagues. Mass of thousands of seeds in the English summary of Mishchenko and colleagues is the same measure.

Images reserved in the article

One image has a place in the article but is not yet uploaded:

File:Overdominance-change-in-allele-frequency.jpg (Selection at an overdominant locus), a chart to be drawn: the change in the frequency of A per generation against that frequency, for fitnesses of 0.85, 1 and 0.70, crossing zero at 0.667. An arrow on each side of the crossing points towards it.