Talk:Population bottleneck

Talk

Discussion for: Population bottleneck • View • Links • Related changes

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

Page scope

Population bottleneck covers sharp falls in population size: their size and length, the different loss of rare alleles and of heterozygosity, the effects on harmful mutations, the methods that read a past fall from genes, the debate over a domestication bottleneck, and the bottlenecks tested or proposed in cannabis. The process of chance loss and the arithmetic of recovery belong to Genetic drift; the harmonic mean, the sex ratio and genebank plant numbers to Effective population size; the values of heterozygosity in hemp and drug types to Heterozygosity; losses from crops and collections in general to Genetic erosion.

Data verification

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

Discussion

Sources

Felsenstein, Theoretical Evolutionary Genetics (text of December 2019), and Davinack, Molecular Ecology & Evolution: An Introduction (2024): the definition, the loss of 1/(2N) of heterozygosity a generation, 2N copies drawn each generation, slow recovery, faster coalescence at a small size, and the attribution of the first computations to Nei, Maruyama and Chakraborty (1975).
Cornuet and Luikart (1996), Luikart and colleagues (1998), Garza and Williamson (2001), Tajima (1989, two papers), Li and Durbin (2011): the methods.
Peery and colleagues (2012), Molecular Ecology 21: 3403–3418, and Mazet and colleagues (2016), Heredity 116: 362–371: how those methods can mislead.
Kirkpatrick and Jarne (2000), The American Naturalist 155: 154–167: inbreeding depression and genetic load after a bottleneck.
Eyre-Walker and colleagues (1998), Wright and colleagues (2005), Hyten and colleagues (2006), Wang and colleagues (2017), Gaut and colleagues (2018), Allaby and colleagues (2019): bottlenecks in the domestication of maize, soybean, barley and sorghum.
Ren and colleagues (2021), Science Advances 7: eabg2286: the demographic models. The sizes and dates on the article other than the interval for the first split are the point estimates in Figure 2B.
Woods and colleagues (2023), Zhang and colleagues (2018), Soorni and colleagues (2017): Tajima's D in cannabis. Zhang and colleagues (2020), Frontiers in Genetics 11: 958: the microsatellite test.
Clarke and Merlin (2016), Dufresnes and colleagues (2017), Soler and colleagues (2017), Henry and colleagues (2020), Welling and colleagues (2016), Sawler and colleagues (2015), McPartland (2018): bottlenecks proposed from breeding history, diversity and fossil pollen.
Crossa and colleagues (1993), Theoretical and Applied Genetics 86: 673–678: plant numbers for regeneration.

Where the sources disagree

Which group lost variation, hemp or drug type. Lynch and colleagues found lower heterozygosity in 16 fibre accessions. Sawler and colleagues (2015) found it lower in 81 marijuana varieties than in 43 fibre varieties. Welling and colleagues (2016) set the two side by side and note how few fibre accessions the first study held.
A contraction or stable populations. Woods and colleagues (2023) read positive Tajima's D as recent contraction in every group except the Asian one. Zhang and colleagues (2018) read non-significant D and F_s as stable populations. The first uses about 7,500 nuclear single-copy genes and the second five chloroplast regions.
The meaning of Tajima's D. Soorni and colleagues (2017) gloss a positive value as an excess of rare polymorphism. Woods and colleagues take a positive value as the mark of a contraction without time to expand.
Plant numbers for regeneration. Crossa and colleagues (1993) give 160 to 210 plants for alleles at 0.05 or more at 150 loci with a 90 to 95 per cent chance. The Vavilov Institute reports cite the same paper for 2,000 plants to keep 99 per cent of the gene pool of a dioecious accession. The aims stated differ.
Whether maize shows a bottleneck. Eyre-Walker and colleagues (1998) and Wang and colleagues (2017) model one from modern sequences; Wang and colleagues make it protracted. Allaby and colleagues (2019) find no dip at the level of individual ancient genomes.

How far the evidence goes

The groups in Zhang and colleagues (2020) are regions of origin, each holding accessions from many places. Cornuet and Luikart describe their tests for a sample from one population. The result also changes with the mutation model: significant under the infinite-allele model, not under the stepwise one.
The SNP set of Woods and colleagues was filtered to a minor allele frequency of at least 5 per cent. Soorni and colleagues filtered differently, so their values of D are not directly comparable.
Ren and colleagues merge feral and cultivated drug-type plants in their models and state that their groups are not panmictic populations. Mazet and colleagues show that structure of this kind can read as a change in size.
Ren and colleagues give no cause for the fall of the basal group in the last 54 years.
Clarke and Merlin's accounts of seed carried to new settlements and of the Afghan introductions are a historical reconstruction in a review. Their paper applies no bottleneck test to them.
Dufresnes, Soler and Henry each propose a bottleneck to explain diversity values. None of them applies a bottleneck test.
The probabilities in the section on rare alleles are calculated from the sampling of 2N gene copies, with every plant contributing equally. Unequal contributions would lower the chance of keeping a rare allele.

Terms

Bottleneck in some cannabis papers means any obstacle to breeding, such as scarce germplasm or the late appearance of flowering class. Those uses are not this concept.
Founder effect: a bottleneck at the founding of a new population.
Basal cannabis: the name Ren and colleagues give to the group of feral plants and landraces, mostly from China, closest to the ancestral gene pool.
Panmictic: mating at random throughout.
Mode shift: the change in the shape of the allele frequency distribution, from most alleles in the rarest class to more in a middle class.