Talk:Selective sweep
Discussion page of Selective sweep
More actions
Talk
Discussion for: Selective sweep • View • Links • Related changes
Use this page for verification notes, disputes, and action items. Keep narrative content on the article.
Page scope
Selective sweep covers the loss of diversity around a favoured allele: the hitch-hiking effect, the signs a sweep leaves and how long they last, hard and soft sweeps, other causes of low diversity, one sweep under crop domestication, and the scans that have reported candidate sweep regions in cannabis and how each was made. Falls in population size belong to Population bottleneck, chance changes in allele frequency to Genetic drift, selection by people to Artificial selection, adaptation to place to Local adaptation, the change in a trait's mean to Response to selection and the division of cannabis into genetic groups to Population structure. The traits behind the cannabis results are read on Cannabis genetics.
Data verification
Record source conflicts, figures that differ between studies and sources not used here. Keep narrative content on the article.
Discussion
Sources
- Maynard Smith and Haigh (1974), Genetical Research: the hitch-hiking effect and its weight against random fixation.
- Booker, Jackson and Keightley (2017), BMC Biology: the predictions of recurrent sweeps, the link between diversity and recombination, how long each signal lasts, hard, soft, incomplete and partial sweeps, and background selection.
- Pennings and Hermisson (2006), PLoS Genetics: the classical footprint and the weaker signal of sweeps from recurrent mutation.
- Voight and colleagues (2006), PLoS Biology: a haplotype statistic for sweeps still under way, applied to HapMap data.
- Garud and colleagues (2015), PLOS Genetics; Harris, Sackman and Jensen (2018), PLOS Genetics: whether soft sweeps are common.
- Olsen and colleagues (2006), Genetics: the sweep around Waxy in temperate japonica rice.
- Allaby, Ware and Kistler (2019), Evolutionary Applications: what the domestication bottleneck was expected to leave and how archaeological evidence bears on it.
- Ren and colleagues (2021), Science Advances; Chen and colleagues (2022), BMC Plant Biology; Lynch and colleagues (2025), Nature; Wang S. and colleagues (2025), International Journal of Molecular Sciences; Wang B. and colleagues (2026), BMC Plant Biology: scans of cannabis genomes.
Where the sources disagree
- How common soft sweeps are. Garud and colleagues find that the strongest candidate sweeps in North American Drosophila melanogaster all look soft. Harris, Sackman and Jensen hold that the methods behind this and two other such studies give very high rates of false positives, and prefer demographic histories with rare hard sweeps. Booker and colleagues note that soft sweeps from new mutation need high beneficial mutation rates, and that selection on standing variation may still give a hard sweep.
- Strength of selection under domestication. Olsen and colleagues estimate a selection coefficient of about 4 at rice Waxy and rank selection under domestication one to two orders of magnitude above strong natural selection. Allaby and colleagues read the archaeological evidence from cereals as showing weak selection on domestication traits, over a long transition, at few loci. The two concern different traits: a grain-quality mutation favoured in one variety group of rice, and traits of the domestication syndrome such as the loss of seed shattering.
- What counts as a sweep in cannabis. Ren and colleagues call a window a putative sweep only when it stands out on FST, the ratio of diversity and XP-CLR together, in order to separate sweeps from the effects of bottlenecks. Wang S. and colleagues use two of these statistics, Lynch and colleagues report XP-CLR and FST separately, and Wang B. and colleagues use XP-CLR alone for their regional comparisons. Wang B. and colleagues call regions found by FST alone between plants grouped by their own cannabinoid content chemotype-associated selective sweeps. Chen and colleagues use FST and the ratio of diversity and call the genes found positively selected genes, without the word sweep.
- Position of THCA synthase. Wang B. and colleagues place the synthase cluster on chromosome 7 of the CS10 assembly in their introduction and the THCA synthase gene on chromosome 2 in their discussion.
How far the evidence goes
- Each cannabis scan keeps a fixed share of windows, the top 1, 5 or 10 per cent. An outlier threshold of this kind always returns some windows, whether or not selection acted. Ren and colleagues report that no functional category of genes was significantly enriched after correction for multiple tests, and Wang S. and colleagues found no significant enrichment in their drug-type comparison.
- Ren and colleagues note that their groups are not randomly mating populations and that most were bred separately at a probably small effective size. Their basal group held 16 plants.
- Wang S. and colleagues drew 110 of their 113 genomes from published resequencing data from around the world. Ren and colleagues resequenced 110 accessions of worldwide origin, and Wang S. and colleagues call their own four groups similar to Ren and colleagues' results. The two scans may draw largely on the same plants.
- The genes in the cannabis windows are named from homologues in other plants. Their functions in cannabis were not tested in these studies, and the trait each window might affect is the authors' reading.
- Lynch and colleagues report one gene, GIGANTEA, from their XP-CLR scan and give no count of significant windows in the main text.
- The loss of complete synthase sequences reported by Ren and colleagues comes from mapping reads to a reference genome carrying both genes, not from the sweep scan. Ren and colleagues add that the loss in hemp may also reflect breeding for low THC under licensing rules.
- The paper by Wang B. and colleagues is an article in press, published ahead of its final version. Its chemotype groups are 90 plants from 14 populations, and the regions it calls chemotype-associated sweeps total about 41 and 40 Mb.
- The cannabis scans use FST, ratios of diversity and XP-CLR. None of them reports a haplotype statistic or a test for soft sweeps.
- The results on soft sweeps come from Drosophila, humans and HIV, and the classic crop example from rice. None of them is a result in cannabis.
Terms
- Hitch-hiking: the change in frequency of a variant caused by selection on a linked allele.
- Hard sweep: a sweep in which one new mutation, on one haplotype, rises to fixation.
- Soft sweep: a sweep in which the favoured allele rises on several haplotypes, from standing variation or from repeated mutation.
- FST: a measure of differentiation in allele frequencies between groups.
- XP-CLR: the cross-population composite likelihood ratio test, which looks for allele-frequency differentiation at linked loci between two groups.
- Tajima's D: a statistic that compares two estimates of diversity and is negative when rare variants are in excess.
- Background selection: the loss of linked neutral diversity caused by the removal of harmful mutations.