4.4.4C: Gene Flow and Mutation - Biology

A population’s genetic variation changes as individuals migrate into or out of a population and when mutations introduce new alleles.

Learning Objectives

  • Explain how gene flow and mutations can influence the allele frequencies of a population

Key Points

  • Plant populations experience gene flow by spreading their pollen long distances.
  • Animals experience gene flow when individuals leave a family group or herd to join other populations.
  • The flow of individuals in and out of a population introduces new alleles and increases genetic variation within that population.
  • Mutations are changes to an organism’s DNA that create diversity within a population by introducing new alleles.
  • Some mutations are harmful and are quickly eliminated from the population by natural selection; harmful mutations prevent organisms from reaching sexual maturity and reproducing.
  • Other mutations are beneficial and can increase in a population if they help organisms reach sexual maturity and reproduce.

Key Terms

  • gene flow: the transfer of alleles or genes from one population to another
  • mutation: any heritable change of the base-pair sequence of genetic material

Gene Flow

An important evolutionary force is gene flow: the flow of alleles in and out of a population due to the migration of individuals or gametes. While some populations are fairly stable, others experience more movement and fluctuation. Many plants, for example, send their pollen by wind, insects, or birds to pollinate other populations of the same species some distance away. Even a population that may initially appear to be stable, such as a pride of lions, can receive new genetic variation as developing males leave their mothers to form new prides with genetically-unrelated females. This variable flow of individuals in and out of the group not only changes the gene structure of the population, but can also introduce new genetic variation to populations in different geological locations and habitats.

Maintained gene flow between two populations can also lead to a combination of the two gene pools, reducing the genetic variation between the two groups. Gene flow strongly acts against speciation, by recombining the gene pools of the groups, and thus, repairing the developing differences in genetic variation that would have led to full speciation and creation of daughter species.

For example, if a species of grass grows on both sides of a highway, pollen is likely to be transported from one side to the other and vice versa. If this pollen is able to fertilize the plant where it ends up and produce viable offspring, then the alleles in the pollen have effectively linked the population on one side of the highway with the other.


Mutations are changes to an organism’s DNA and are an important driver of diversity in populations. Species evolve because of the accumulation of mutations that occur over time. The appearance of new mutations is the most common way to introduce novel genotypic and phenotypic variance. Some mutations are unfavorable or harmful and are quickly eliminated from the population by natural selection. Others are beneficial and will spread through the population. Whether or not a mutation is beneficial or harmful is determined by whether it helps an organism survive to sexual maturity and reproduce. Some mutations have no effect on an organism and can linger, unaffected by natural selection, in the genome while others can have a dramatic effect on a gene and the resulting phenotype.


Autosomal recessive bestrophinopathy (ARB) is a disease that results from the mutations in the BEST1 gene. It is characterized by multifocal yellowish lipofuscin deposits, cystoid macular edema, and subretinal fluid. Among approximately 270 BEST1 mutations, only 40 that include both heterozygous and homozygous mutations are associated with ARB. However, very few ARB-related mutations have been reported in the Japanese population. Therefore, in this study, we aimed to identify BEST1 mutations and describe the genotype–phenotype relationship in Japanese dizygotic twins presenting with ARB.


We performed clinical examinations in Japanese dizygotic twin patients (male: 29 years) with ARB as well as whole-exome sequencing in seven family members of these twins.


In this study, we have reported on a novel BEST1 mutation, the p. Phe151Cys mutation, associated with ARB in Japanese dizygotic twins who had bi-allelic p. Ala160Pro mutations in BEST1. The clinical features observed were binocular abnormalities of the fundus, such as multifocal yellowish subretinal deposits, cystoid macular edema, and subretinal fluid. The full-field electroretinography results were subnormal.


It was indicated that the novel BEST1 mutations identified may be strongly correlated with binocular ARB. This study provides significant information of the genotype–phenotype association in Japanese ARB patients. Further, the genetic analysis that we performed was very useful for the differential diagnosis and might have implications in the development of future treatment modalities.


Study species

Viburnum lantanoides is an understory shrub that inhabits cool, mesic habitats in mixed hardwood and boreal forests from the southern Appalachians to the Canadian Maritime Provinces. In the northern portion of its range, V. lantanoides can be locally abundant, often occurring in dense thickets in cool moist forests throughout northern New England, the Canadian Maritimes to the east, and southeastern Ontario to the west. It is more patchily distributed in the southern portion of its range, occurring exclusively at higher elevations in spruce–fir forests in the southern Appalachians. Flowers are borne in compound umbel-like inflorescences in which a cluster of small perfect flowers is encircled by sterile flowers with greatly enlarged corollas. Plants flower for a short time during early spring and are visited primarily by andrenid bees, syrphid flies, and elaterid beetles (Park et al., 2019 2019). Fruits are mainly produced through outcrossing (Park et al., 2019 2019) and typically are dispersed by birds, although rates of frugivory are low (Gould, 1966 1966).

Sampling and sequencing

One to three plants were sampled from 70 localities distributed throughout the geographic range of V. lantanoides (Appendix S1). Latitude/longitude coordinates for each locality were obtained using a Garmin GPSmap 62s handheld GPS unit (Olathe, Kansas, USA) to 2–2000 m precision. Leaves were preserved in silica gel, and voucher specimens for each population were deposited in the Yale University Herbarium (YU). A modified CTAB method was used to extract genomic DNA from our collections (Healey et al., 2014 2014). The resulting DNA extractions were prepared for RAD-sequencing (Baird et al., 2008 2008) with the restriction enzyme Pst1 by Floragenex (Portland, Oregon, USA). RAD-libraries (two libraries consisting of 95 samples each) were sequenced at the GC3F sequencing facility at the University of Oregon (Eugene, Oregon, USA) over two lanes of Illumina Hi-Seq 4000 to produce 100 bp, single-end reads.

Sequence processing

The Python program iPyrad version 0.7.12 (Eaton, 2014 2014) was used to demultiplex, filter, and generate homologous de novo RAD-loci. Raw reads were demultiplexed to individual samples by unique 10 bp sample-specific barcodes and filtered to remove adapter-contaminated reads and reads with more than four low-quality bases (phred scores <20). Filtered reads were clustered within samples at 88% sequence similarity into de novo loci, and loci with fewer than six or >10,000 reads were removed from further processing. Resulting loci were clustered across samples at 88% sequence similarity and processed into a “base” assembly, where all loci were shared between four or more ingroup and outgroup samples. In total, we included 190 V. lantanoides accessions. For rooting purposes we included one sample each of two closely related species within the Pseudotinus clade (V. sympodiale of China and Taiwan, and V. furcatum of Japan). Twelve samples with <10,000 loci were excluded from further processing because they were found to contribute disproportionately to the amount of missing data in downstream analyses.

The “base” assembly was then processed to generate two assemblies for analysis: a “phylogenomics” assembly for phylogenetic analyses, where all loci were shared among ≥90 ingroup and outgroup samples (178 V. lantanoides samples, two outgroup samples) and an “ingroup” assembly for population genomic analyses, where all loci were shared among ≥89 ingroup samples (n = 178 samples). The “ingroup” assembly was subsequently filtered using the R package “radiator” (Gosselin, 2017 2017) to randomly select one single-nucleotide polymorphism (SNP) per RAD-locus, to remove SNPs with more than two alleles (V. lantanoides is diploid, with 2N = 18 Egolf, 1962), and to remove SNPs with minor allele frequency <0.01. Our relatively low sample coverage threshold was chosen because it allowed for the inclusion of high-mutation-rate loci, which may be useful in resolving relationships among more recently diverged populations (Huang and Knowles, 2014 2014 Eaton et al., 2017 2017), while allowing for reasonable levels of missing data, an excess of which may bias phylogeographic signal (Rubin et al., 2012 2012).

Genetic diversity

Global population genetic parameters (e.g., Fis, Ho, Ht, Hs) were estimated using the R package “hierfstat” (Goudet, 2005 2005). Overall Fst was estimated using Weir and Cockerham's ( 1984 1984) method in the R package “assigner” (Gosselin et al., 2016 2016), with 100 bootstrap replicates to assess significance. We assessed patterns of isolation-by-distance (IBD) by performing a Mantel test using the “mantel.randtest” function from the R package “ade4” (Dray and Dufour, 2007 2007) and 1 × 10 6 Monte Carlo simulations to test for significance. We calculated pairwise genetic distances between samples (i.e., the proportion of allelic differences) using the “bitwise.dist” function in the R package “poppR” (Kamvar et al., 2014 2014), and geographic distances between collection localities using the “distm” function in the R package “geosphere” (Hijmans, 2017 2016). We regressed sample observed heterozygosity (Ho) against latitude and longitude to identify geographic patterns in the distribution of genetic diversity. We used BCFtools (Li et al., 2009 2009) to estimate Ho by dividing the number of heterozygous sites by the sum of homozygous and heterozygous sites sequenced per sample. Finally, we performed a hierarchical analysis of molecular variance (AMOVA) in “poppR,” with 10,000 permutations to assess significance. AMOVA was performed for the following levels: between genetic clusters (described below) among localities within genetic clusters among samples at each collecting locality and within samples.

Population structure

Four approaches were used to assess population structure. We first used sNMF (Frichot et al., 2014 2014) as implemented in the R package “LEA” (Frichot and François, 2015 2015) to assign samples to genetic clusters. sNMF estimates individual ancestry coefficients utilizing the same likelihood model underlying Structure (Pritchard et al., 2000 2000) and Admixture (Alexander et al., 2009 2009), but uses non-negative matrix factorization and least-squares optimization to accommodate large, genome-scale datasets. sNMF is also robust to deviations from Hardy-Weinberg equilibrium, making it an effective tool to infer population structure in organisms with diffuse population structure. We complemented the sNMF analyses with tess3r (Caye et al., 2016 2016), which utilizes the same algorithm as sNMF but incorporates geographic coordinates in estimating sample ancestry coefficients. Both sNMF and tess3r analyses were performed using default parameters for K = 1–10, with 10 repetitions for each value of K, and the optimal value of K assessed using the cross-entropy criterion. We also used two model-free approaches to infer genetic clusters. We used discriminant analysis of principal components (DAPC Jombart et al., 2010 2010) as implemented in the R package “adegenet” (Jombart, 2008 2008), using K-means clustering to identify the optimal number of clusters from K = 1–10, and estimating individual admixture coefficients using n/3 principal components (n = 178 samples). We then calculated Bayesian information criterion scores for each value of K to identify the best-fit number of clusters. We also conducted principal component analysis (PCA) as implemented in the R package “LEA” to visualize samples in two-dimensional genetic space.

Phylogenetic relationships

Supermatrix and species-tree approaches were used to infer phylogenetic relationships among samples in the “phylogenomic” assembly. Supermatrix analyses were conducted using RAxML version 8.2.1 (Stamatakis, 2014 2014) under the GTR + CAT substitution model with 100 rapid bootstrap replicates to assess support. Species-tree analyses were conducted in “tetrad,” an implementation of the program SVDquartets (Chifman and Kubatko, 2014 2014) in iPyrad tetrad estimates topologies for a user-specified number of quartets present in a set of taxa and assembles the resulting quartets into a supertree. For our analyses, we sampled all quartets and assessed node support with 100 nonparametric bootstrap replicates.

Demographic modeling

We used demographic model testing to identify the scenario that best describes the demographic history of V. lantanoides. Based on the results of our population structure and phylogenetic analyses, we focused on determining the timing of the origin of the northern lineage and its mode of population growth. Our models were based on a simple scenario in which ancestral lineages of V. lantanoides differentiate into two independently evolving entities: a southern Appalachian cluster comprising several lineages from the southern Appalachian mountains and a northern lineage representing the single genetic lineage that appears to have colonized the northern portion of the range of V. lantanoides (see below). Models were then formulated to identify the mode of postdivergence population growth, where we allowed for nonlinear changes in effective population size in none (Appendix S2A), one (Appendix S2B, S2C), or both (Appendix S2D) the southern Appalachian cluster and the northern lineage. Our models estimated three parameters: contemporary effective population sizes (Ne) of the southern Appalachian cluster (N1) and northern lineage (N2), and divergence time for the northern lineage (T). We did not consider island models (Wright, 1943 1943) because the significant pattern of IBD and diffuse population structure are poor fits to V. lantanoides demographic history. Furthermore, considering the monophyly of the northern samples and a significant pattern of IBD, we did not include a migration parameter in our models.

We fit models using the diffusion approximation approach in δaδi (Gutenkunst et al., 2009 2009). We generated a folded, two-dimensional joint site frequency spectrum from the “ingroup” assembly using the Python script “easySFS” (, sampling 14 alleles from the southern Appalachian cluster and 16 alleles from the northern lineage to maximize the number of segregating sites for the analysis. Initial optimizations were conducted using 20 independent replicates of twofold perturbed parameters and were optimized using the L-BFGS-B algorithm for a maximum of 100 repetitions. Parameter estimates from the replicates with the highest log-likelihood were used to generate 20 additional sets of onefold perturbed starting points for final optimization, and the repetition with the highest log-likelihood was used to compare models and estimate parameter uncertainties.

We compared models using the Akaike information criterion (AIC). Calculation of AIC scores and model selection was conducted following Burnham and Anderson ( 2007 2007). Standard deviations for the maximum-likelihood parameter estimates were obtained using the Fisher information matrix (FIM) approach (Coffman et al., 2016 2016), which is a computationally efficient alternative to traditional bootstrapping. The FIM approach may not be appropriate for datasets composed of highly linked SNPs or when fitting complex models. However, considering that we sampled one SNP per RAD-locus, and that our models estimated four free parameters, we believe that the FIM approach provided reasonable uncertainty estimates for our analysis. Parameter estimates and uncertainties were transformed into biological units from model-derived estimates of ϴ (population mutation rate). Mutation rates in Viburnum are unknown, so we conservatively assumed a rate obtained for the woody plant Populus (2.5 × 10 −9 substitutions site −1 yr −1 Ingvarsson, 2008 2008). We also assumed a generation time of 10 yr, which we based on the observation that V. lantanoides produces “double-dormant” seeds (i.e., seeds need to be exposed to two successive winters before emergence of the epicotyl) and that several years of growth are needed before plants become reproductive.

Species distribution models

We used MaxEnt version 3.3.3k (Phillips et al., 2004 2004, 2006 2006) as implemented in the R package “dismo” (Hijmans et al., 2017 2017) to estimate the current and historical geographic range of V. lantanoides. Data for 19 bioclimatic variables under current climates (1950–1990) and three LGM climate models (22 kya CCSM-4, MIROC, and MPI-ESM) were downloaded at 2.5 arcminute resolution (

5 km 2 at the equator) from WorldClim ( The three LGM models are based on different general circulation models that vary in how they simulate climate dynamics (Varela et al., 2015 2015). We pruned the climate data used to estimate distribution models by sampling 10,000 random points spanning our study region (latitude: 25–58°N, longitude: 100–50°W) and removing variables with a Pearson R 2 > 0.7. Using this approach, we retained the following six variables: mean annual temperature (bio1), mean diurnal range (bio2), mean temperature of the wettest quarter (bio8), mean annual precipitation (bio12), precipitation seasonality (bio15), and precipitation of the warmest quarter (bio18).

Locality data were from our own collections, which cover much of the known range of V. lantanoides (Appendix S3 n = 86 localities). We chose not to include data from other sources (e.g., Global Biodiversity Information Facility) because those data are less reliable and contain misidentifications. We sampled 10,000 pseudo-absence points (Barbet-Massin et al., 2012 2012) from a minimum-convex polygon drawn two decimal degrees around each occurrence. Distribution models were generated using current climate data and then evaluated using fivefold cross-validation and the area-under-curve (AUC) statistic. Finally, a current climate model using all occurrences was used to project the distribution of V. lantanoides using the three models of LGM climate.

Tying It All Together

While the biggest leap forward in understanding how evolution works came with the joining (synthesis) of Darwin’s concept of natural selection with Mendel’s insights about particulate inheritance, there were some other big contributions that were crucial to making sense of the variation that was being observed. R.A. Fisher (1919) and John Burdon Sanderson Haldane (1924) developed and tested mathematical models for evolutionary change that provided the tools to study variation and became the basis for the study of population genetics. Sewall Wright (1932) and Theodosius Dobzhansky (1937) performed studies that revealed the existence of chromosomes as carriers of collections of genes. Edmund Brisco Ford (1949) conducted studies on wild butterflies that confirmed Fisher’s mathematical predictions and also led to his definition of the concept of polymorphisms to describe alternative phenotypes, or multiple forms of a trait. Ford (1942) also correctly predicted that human blood type polymorphisms were maintained in the population because they were involved in disease resistance. Julian Huxley’s 1942 book, Evolution: The Modern Synthesis, provided an easy-to-read summary of the evolutionary studies that had come before. It was with this book that the term Modern Synthesis was first used to describe the integration of Darwin’s, Mendel’s, and subsequent research into a unified theory of evolution. In appealing to the general public, Huxley’s book also found new success establishing a wide acceptance of the process of evolution.


Some conservation biologists and conservationists have used genetic pollution for a number of years as a term to describe gene flow from a non-native, invasive subspecies, domestic, or genetically-engineered population to a wild indigenous population. [3] [7] [8]

Importance Edit

The introduction of genetic material into the gene pool of a population by human intervention can have both positive and negative effects on populations. When genetic material is intentionally introduced to increase the fitness of a population, this is called genetic rescue. When genetic material is unintentionally introduced to a population, this is called genetic pollution and can negatively affect the fitness of a population (primarily through outbreeding depression), introduce other unwanted phenotypes, or theoretically lead to extinction.

Introduced species Edit

An introduced species is one that is not native to a given population that is either intentionally or accidentally brought into a given ecosystem. Effects of introduction are highly variable, but if an introduced species has a major negative impact on its new environment, it can be considered an invasive species. One such example is the introduction of the Asian Longhorned beetle in North America, which was first detected in 1996 in Brooklyn, New York. It is believed that these beetles were introduced through cargo at trade ports. The beetles are highly damaging to the environment, and are estimated to cause risk to 35% of urban trees, excluding natural forests. [9] These beetles cause severe damage to the wood of trees by larval funneling. Their presence in the ecosystem destabilizes community structure, having a negative influence on many species in the system.

Introduced species are not always disruptive to an environment, however. Tomás Carlo and Jason Gleditch of Penn State University found that the number of "invasive" honeysuckle plants in the area correlated with the number and diversity of the birds in the Happy Valley Region of Pennsylvania, suggesting introduced honeysuckle plants and birds formed a mutually beneficial relationship. [10] Presence of introduced honeysuckle was associated with higher diversity of the bird populations in that area, demonstrating that introduced species are not always detrimental to a given environment and it is completely context dependent.

Invasive species Edit

Conservation biologists and conservationists have, for a number of years, used the term to describe gene flow from domestic, feral, and non-native species into wild indigenous species, which they consider undesirable. [3] [7] [8] For example, TRAFFIC is the international wildlife trade monitoring network that works to limit trade in wild plants and animals so that it is not a threat to conservationist goals. They promote awareness of the effects of introduced invasive species that may "hybridize with native species, causing genetic pollution". [11] Furthermore, the Joint Nature Conservation Committee, the statutory adviser to the UK government, has stated that invasive species "will alter the genetic pool (a process called genetic pollution), which is an irreversible change." [12]

Invasive species can invade both large and small native populations and have a profound effect. Upon invasion, invasive species interbreed with native species to form sterile or more evolutionarily fit hybrids that can outcompete the native populations. Invasive species can cause extinctions of small populations on islands that are particularly vulnerable due to their smaller amounts of genetic diversity. In these populations, local adaptations can be disrupted by the introduction of new genes that may not be as suitable for the small island environments. For example, the Cercocarpus traskiae of the Catalina Island off the coast of California has faced near extinction with only a single population remaining due to the hybridization of its offspring with Cercocarpus betuloides. [13]

Domestic populations Edit

Increased contact between wild and domesticated populations of organisms can lead to reproductive interactions that are detrimental to the wild population's ability to survive. A wild population is one that lives in natural areas and is not regularly looked after by humans. This contrasts with domesticated populations that live in human controlled areas and are regularly, and historically, in contact with humans. Genes from domesticated populations are added to wild populations as a result of reproduction. In many crop populations this can be the result of pollen traveling from farmed crops to neighboring wild plants of the same species. For farmed animals, this reproduction may happen as the result of escaped or released animals.

Aquaculture Edit

Aquaculture is the practice of farming aquatic animals or plants for the purpose of consumption. This practice is becoming increasingly common for the production of salmon. This is specifically termed aquaculture of salmonoids. One of the dangers of this practice is the possibility of domesticated salmon breaking free from their containment. The occurrence of escaping incidents is becoming increasingly common as aquaculture gains popularity. [14] [15] [16] Farming structures may be ineffective at holding the vast number of fast growing animals they house. [17] Natural disasters, high tides, and other environmental occurrences can also trigger aquatic animal escapes. [18] [19] The reason these escapes are considered dangers is the impact they pose for the wild population they reproduce with after escaping. In many instances the wild population experiences a decreased likelihood of survival after reproducing with domesticated populations of salmon. [20] [21]

The Washington Department of Fish and Wildlife cites that "commonly expressed concerns surrounding escaped Atlantic salmon include competition with native salmon, predation, disease transfer, hybridization, and colonization." [22] A report done by that organization in 1999 did not find that escaped salmon posed a significant risk to wild populations. [23]

Crops Edit

Crops refer to groups of plants grown for consumption. Despite domestication over many years, these plants are not so far removed from their wild relatives that they could reproduce if brought together. Many crops are still grown in the areas they originated and gene flow between crops and wild relatives impacts the evolution of wild populations. [24] Farmers can avoid reproduction between the different populations by timing their planting of crops so that crops are not flowering when wild relatives would be. Domesticated crops have been changed through artificial selection and genetic engineering. The genetic make up of many crops is different than that of its wild relatives, [25] but the closer they grow to one another the more likely they are to share genes through pollen. Gene flow persists between crops and wild counterparts.

Genetically engineered organisms Edit

Genetically engineered organisms are genetically modified in a laboratory, and therefore distinct from those that were bred through artificial selection. In the fields of agriculture, agroforestry and animal husbandry, genetic pollution is being used to describe gene flows between GE species and wild relatives. [26] An early use of the term "genetic pollution" in this later sense appears in a wide-ranging review of the potential ecological effects of genetic engineering in Jeremy Rifkin in his 1998 book The Biotech Century. [27] While intentional crossbreeding between two genetically distinct varieties is described as hybridization with the subsequent introgression of genes, Rifkin, who had played a leading role in the ethical debate for over a decade before, used genetic pollution to describe what he considered to be problems that might occur due to the unintentional process of (modernly) genetically modified organisms (GMOs) dispersing their genes into the natural environment by breeding with wild plants or animals. [26] [28] [29]

Concerns about negative consequences from gene flow between genetically engineered organisms and wild populations are valid. Most corn and soybean crops grown in the midwestern USA are genetically modified. There are corn and soybean varieties that are resistant to herbicides like glyphosate [30] and corn that produces neonicotinoid pesticide within all of its tissues. [31] These genetic modifications are meant to increase yields of crops but there is little evidence that yields actually increase. [31] While scientists are concerned genetically engineered organisms can have negative effects on surrounding plant and animal communities, the risk of gene flow between genetically engineered organisms and wild populations is yet another concern. Many farmed crops may be weed resistant and reproduce with wild relatives. [32] More research is necessary to understand how much gene flow between genetically engineered crops and wild populations occurs, and the impacts of genetic mixing.

Mutated organisms Edit

Mutations within organisms can be executed through the process of exposing the organism to chemicals or radiation in order to generate mutations. This has been done in plants in order to create mutants that have a desired trait. These mutants can then be bred with other mutants or individuals that are not mutated in order to maintain the mutant trait. However, similar to the risks associated with introducing individuals to a certain environment, the variation created by mutated individuals could have a negative impact on native populations as well.

Preventive measures Edit

Since 2005 there has existed a GM Contamination Register, launched for GeneWatch UK and Greenpeace International that records all incidents of intentional or accidental [33] [34] release of organisms genetically modified using modern techniques. [35]

Genetic use restriction technologies (GURTs) were developed for the purpose of property protection, but could be beneficial in preventing the dispersal of transgenes. GeneSafe technologies introduced a method that became known as "Terminator." This method is based on seeds that produce sterile plants. This would prevent movement of transgenes into wild populations as hybridization would not be possible. [36] However, this technology has never been deployed as it disproportionately negatively affects farmers in developing countries, who save seeds to use each year (whereas in developed countries, farmers generally buy seeds from seed production companies). [36]

Physical containment has also been utilized to prevent the escape of transgenes. Physical containment includes barriers such as filters in labs, screens in greenhouses, and isolation distances in the field. Isolation distances have not always been successful, such as transgene escape from an isolated field into the wild in herbicide-resistant bentgrass Agrostis stolonifera. [37]

Another suggested method that applies specifically to protection traits (e.g. pathogen resistance) is mitigation. Mitigation involves linking the positive trait (beneficial to fitness) to a trait that is negative (harmful to fitness) to wild but not domesticated individuals. [37] In this case, if the protection trait was introduced to a weed, the negative trait would also be introduced in order to decrease overall fitness of the weed and decrease possibility of the individual’s reproduction and thus propagation of the transgene.

Risks Edit

Not all genetically engineered organisms cause genetic pollution. Genetic engineering has a variety of uses and is specifically defined as a direct manipulation of the genome of an organism. Genetic pollution can occur in response to the introduction of a species that is not native to a particular environment, and genetically engineered organisms are examples of individuals that could cause genetic pollution following introduction. Due to these risks, studies have been done in order to assess the risks of genetic pollution associated with organisms that have been genetically engineered:

  1. Genetic In a 10-year study of four different crops, none of the genetically engineered plants were found to be more invasive or more persistent than their conventional counterparts. [38] An often cited claimed example of genetic pollution is the reputed discovery of transgenes from GE maize in landraces of maize in Oaxaca, Mexico. The report from Quist and Chapela, [39] has since been discredited on methodological grounds. [40] The scientific journal that originally published the study concluded that "the evidence available is not sufficient to justify the publication of the original paper." [41] More recent attempts to replicate the original studies have concluded that genetically modified corn is absent from southern Mexico in 2003 and 2004. [42]
  2. A 2009 study verified the original findings of the controversial 2001 study, by finding transgenes in about 1% of 2000 samples of wild maize in Oaxaca, Mexico, despite Nature retracting the 2001 study and a second study failing to back up the findings of the initial study. The study found that the transgenes are common in some fields, but non-existent in others, hence explaining why a previous study failed to find them. Furthermore, not every laboratory method managed to find the transgenes. [43]
  3. A 2004 study performed near an Oregon field trial for a genetically modified variety of creeping bentgrass (Agrostis stolonifera) revealed that the transgene and its associate trait (resistance to the glyphosate herbicide) could be transmitted by wind pollination to resident plants of different Agrostis species, up to 14 km from the test field. [44] In 2007, the Scotts Company, producer of the genetically modified bentgrass, agreed to pay a civil penalty of $500,000 to the United States Department of Agriculture (USDA). The USDA alleged that Scotts "failed to conduct a 2003 Oregon field trial in a manner which ensured that neither glyphosate-tolerant creeping bentgrass nor its offspring would persist in the environment". [45]

Not only are there risks in terms of genetic engineering, but there are risks that emerge from species hybridization. In Czechoslovakia, ibex were introduced from Turkey and Sinai to help promote the ibex population there, which caused hybrids that produced offspring too early, which caused the overall population to disappear completely. [46] The genes of each population of the ibex in Turkey and Sinai were locally adapted to their environments so when placed in a new environmental context did not flourish. Additionally, the environmental toll that may arise from the introduction of a new species may be so disruptive that the ecosystem is no longer able to sustain certain populations.

Environmentalist perspectives Edit

The use of the word "pollution" in the term genetic pollution has a deliberate negative connotation and is meant to convey the idea that mixing genetic information is bad for the environment. However, because the mixing of genetic information can lead to a variety of outcomes, "pollution" may not be the most accurate descriptor. Gene flow is undesirable according to some environmentalists and conservationists, including groups such as Greenpeace, TRAFFIC, and GeneWatch UK. [47] [33] [35] [48] [7] [11] [49]

"Invasive species have been a major cause of extinction throughout the world in the past few hundred years. Some of them prey on native wildlife, compete with it for resources, or spread disease, while others may hybridize with native species, causing "genetic pollution". In these ways, invasive species are as big a threat to the balance of nature as the direct overexploitation by humans of some species." [1]

It can also be considered undesirable if it leads to a loss of fitness in the wild populations. [50] The term can be associated with the gene flow from a mutation bred, synthetic organism or genetically engineered organism to a non GE organism, [26] by those who consider such gene flow detrimental. [47] These environmentalist groups stand in complete opposition to the development and production of genetically engineered organisms.

Governmental definition Edit

From a governmental perspective, genetic pollution is defined as follows by the Food and Agriculture Organization of the United Nations:

"Uncontrolled spread of genetic information (frequently referring to transgenes) into the genomes of organisms in which such genes are not present in nature." [51]

Scientific perspectives Edit

Use of the term 'genetic pollution' and similar phrases such as genetic deterioration, genetic swamping, genetic takeover, and genetic aggression, are being debated by scientists as many do not find it scientifically appropriate. Rhymer and Simberloff argue that these types of terms:

". imply either that hybrids are less fit than the parentals, which need not be the case, or that there is an inherent value in "pure" gene pools." [1]

They recommend that gene flow from invasive species be termed genetic mixing since:

" "Mixing" need not be value-laden, and we use it here to denote mixing of gene pools whether or not associated with a decline in fitness." [1]

Patrick Moore has questioned whether the term "genetic pollution" is more political than scientific. The term is considered to arouse emotional feelings towards the subject matter. [9] In an interview he comments:

"If you take a term used quite frequently these days, the term "genetic pollution," otherwise referred to as genetic contamination, it is a propaganda term, not a technical or scientific term. Pollution and contamination are both value judgments. By using the word "genetic" it gives the public the impression that they are talking about something scientific or technical--as if there were such a thing as genes that amount to pollution." [2]

Thus, using the term "genetic pollution" is inherently political. A scientific approach to discussing gene flow between introduced and native species would be to use terms like genetic mixing or gene flow. Such mixing can definitely have negative consequences on the fitness of native populations, so it is important not to minimize the risk. However, because genetic mixing can also lead to fitness recovery in cases that could be described as "genetic rescue", it is important to distinguish that just mixing genes from introduced into native populations can lead to variable outcomes for the fitness of native populations.


The information contained in DNA determines protein function in the cells of all organisms. Transcription and translation allow this information to be communicated into making proteins. However, an error in reading this communication can cause protein function to be incorrect and eventually cause disease even as the cell incorporates a variety of corrective measures.

Central dogma Edit

In 1956 Francis Crick described the flow of genetic information from DNA to a specific amino acid arrangement for making a protein as the central dogma. [1] For a cell to properly function, proteins are required to be produced accurately for structural and for catalytic activities. An incorrectly made protein can have detrimental effects on cell viability and in most cases cause the higher organism to become unhealthy by abnormal cellular functions. To ensure that the genome successfully passes the information on, proofreading mechanisms such as exonucleases and mismatch repair systems are incorporated in DNA replication . [1]

Transcription and translation Edit

After DNA replication, the reading of a selected section of genetic information is accomplished by transcription. [1] Nucleotides containing the genetic information are now on a single strand messenger template called mRNA. The mRNA is incorporated with a subunit of the ribosome and interacts with an rRNA. The genetic information carried in the codons of the mRNA are now read (decoded) by anticodons of the tRNA. As each codon (triplet) is read, amino acids are being joined together until a stop codon (UAG, UGA or UAA) is reached. At this point the polypeptide (protein) has been synthesised and is released. [1] For every 1000 amino acid incorporated into the protein, no more than one is incorrect. This fidelity of codon recognition, maintaining the importance of the proper reading frame, is accomplished by proper base pairing at the ribosome A site, GTP hydrolysis activity of EF-Tu a form of kinetic stability, and a proofreading mechanism as EF-Tu is released. [1]

Frameshifting may also occur during prophase translation, producing different proteins from overlapping open reading frames, such as the gag-pol-env retroviral proteins. This is fairly common in viruses and also occurs in bacteria and yeast (Farabaugh, 1996). Reverse transcriptase, as opposed to RNA Polymerase II, is thought to be a stronger cause of the occurrence of frameshift mutations. In experiments only 3–13% of all frameshift mutations occurred because of RNA Polymerase II. In prokaryotes the error rate inducing frameshift mutations is only somewhere in the range of .0001 and .00001. [5]

There are several biological processes that help to prevent frameshift mutations. Reverse mutations occur which change the mutated sequence back to the original wild type sequence. Another possibility for mutation correction is the use of a suppressor mutation. This offsets the effect of the original mutation by creating a secondary mutation, shifting the sequence to allow for the correct amino acids to be read. Guide RNA can also be used to insert or delete Uridine into the mRNA after transcription, this allows for the correct reading frame. [1]

Codon-triplet importance Edit

A codon is a set of three nucleotides, a triplet that code for a certain amino acid. The first codon establishes the reading frame, whereby a new codon begins. A protein's amino acid backbone sequence is defined by contiguous triplets. [6] Codons are key to translation of genetic information for the synthesis of proteins. The reading frame is set when translating the mRNA begins and is maintained as it reads one triplet to the next. The reading of the genetic code is subject to three rules the monitor codons in mRNA. First, codons are read in a 5' to 3' direction. Second, codons are nonoverlapping and the message has no gaps. The last rule, as stated above, that the message is translated in a fixed reading frame. [1]

Frameshift mutations can occur randomly or be caused by an external stimulus. The detection of frameshift mutations can occur via several different methods. Frameshifts are just one type of mutation that can lead to incomplete or incorrect proteins, but they account for a significant percentage of errors in DNA.

Genetic or environmental Edit

This is a genetic mutation at the level of nucleotide bases. Why and how frameshift mutations occur are continually being sought after. An environmental study, specifically the production of UV-induced frameshift mutations by DNA polymerases deficient in 3′ → 5′ exonuclease activity was done. The normal sequence 5′ GTC GTT TTA CAA 3′ was changed to GTC GTT T TTA CAA (MIDT) of GTC GTT C TTA CAA (MIDC) to study frameshifts. E. coli pol I Kf and T7 DNA polymerase mutant enzymes devoid of 3′ → 5′ exonuclease activity produce UV-induced revertants at higher frequency than did their exonuclease proficient counterparts. The data indicates that loss of proofreading activity increases the frequency of UV-induced frameshifts. [7]

Detection Edit

Fluorescence Edit

The effects of neighboring bases and secondary structure to detect the frequency of frameshift mutations has been investigated in depth using fluorescence. Fluorescently tagged DNA, by means of base analogues, permits one to study the local changes of a DNA sequence. [8] Studies on the effects of the length of the primer strand reveal that an equilibrium mixture of four hybridization conformations was observed when template bases looped-out as a bulge, i.e. a structure flanked on both sides by duplex DNA. In contrast, a double-loop structure with an unusual unstacked DNA conformation at its downstream edge was observed when the extruded bases were positioned at the primer–template junction, showing that misalignments can be modified by neighboring DNA secondary structure. [9]

Sequencing Edit

Sanger sequencing and pyrosequencing are two methods that have been used to detect frameshift mutations, however, it is likely that data generated will not be of the highest quality. Even still, 1.96 million indels have been identified through Sanger sequencing that do not overlap with other databases. When a frameshift mutation is observed it is compared against the Human Genome Mutation Database (HGMD) to determine if the mutation has a damaging effect. This is done by looking at four features. First, the ratio between the affected and conserved DNA, second the location of the mutation relative to the transcript, third the ratio of conserved and affected amino acids and finally the distance of the indel to the end of the exon. [10]

Massively Parallel Sequencing is a newer method that can be used to detect mutations. Using this method, up to 17 gigabases can be sequenced at once, as opposed to limited ranges for Sanger sequencing of only about 1 kilobase. Several technologies are available to perform this test and it is being looked at to be used in clinical applications. [11] When testing for different carcinomas, current methods only allow for looking at one gene at a time. Massively Parallel Sequencing can test for a variety of cancer causing mutations at once as opposed to several specific tests. [12] An experiment to determine the accuracy of this newer sequencing method tested for 21 genes and had no false positive calls for frameshift mutations. [13]

Diagnosis Edit

A US patent (5,958,684) in 1999 by Leeuwen, details the methods and reagents for diagnosis of diseases caused by or associated with a gene having a somatic mutation giving rise to a frameshift mutation. The methods include providing a tissue or fluid sample and conducting gene analysis for frameshift mutation or a protein from this type of mutation. The nucleotide sequence of the suspected gene is provided from published gene sequences or from cloning and sequencing of the suspect gene. The amino acid sequence encoded by the gene is then predicted. [14]

Frequency Edit

Despite the rules that govern the genetic code and the various mechanisms present in a cell to ensure the correct transfer of genetic information during the process of DNA replication as well as during translation, mutations do occur frameshift mutation is not the only type. There are at least two other types of recognized point mutations, specifically missense mutation and nonsense mutation. [1] A frameshift mutation can drastically change the coding capacity (genetic information) of the message. [1] Small insertions or deletions (those less than 20 base pairs) make up 24% of mutations that manifest in currently recognized genetic disease. [10]

Frameshift mutations are found to be more common in repeat regions of DNA. A reason for this is because of slipping of the polymerase enzyme in repeat regions, allowing for mutations to enter the sequence. [15] Experiments can be run to determine the frequency of the frameshift mutation by adding or removing a pre-set number of nucleotides. Experiments have been run by adding four basepairs, called the +4 experiments, but a team from Emory University looked at the difference in frequency of the mutation by both adding and deleting a base pair. It was shown that there was no difference in the frequency between the addition and deletion of a base pair. There is however, a difference in the end result of the protein. [15]

Huntington's disease is one of the nine codon reiteration disorders caused by polyglutamine expansion mutations that include spino-cerebellar ataxia (SCA) 1, 2, 6, 7 and 3, spinobulbar muscular atrophy and dentatorubal-pallidoluysianatrophy. There may be a link between diseases caused by polyglutamine and polyalanine expansion mutations, as frame shifting of the original SCA3 gene product encoding CAG/polyglutamines to GCA/polyalanines. Ribosomal slippage during translation of the SCA3 protein has been proposed as the mechanism resulting in shifting from the polyglutamine to the polyalanine-encoding frame. A dinucleotide deletion or single nucleotide insertion within the polyglutamine tract of huntingtin exon 1 would shift the CAG, polyglutamineen coding frame by +1 (+1 frame shift) to the GCA, polyalanine-encoding frame and introduce a novel epitope to the C terminus of Htt exon 1 (APAAAPAATRPGCG). [16]

Several diseases have frameshift mutations as at least part of the cause. Knowing prevalent mutations can also aid in the diagnosis of the disease. Currently there are attempts to use frameshift mutations beneficially in the treatment of diseases, changing the reading frame of the amino acids.

<p>This section provides any useful information about the protein, mostly biological knowledge.<p><a href='/help/function_section' target='_top'>More. </a></p> Function i

Probable core component of the endosomal sorting required for transport complex III (ESCRT-III) which is involved in multivesicular bodies (MVBs) formation and sorting of endosomal cargo proteins into MVBs. MVBs contain intraluminal vesicles (ILVs) that are generated by invagination and scission from the limiting membrane of the endosome and mostly are delivered to lysosomes enabling degradation of membrane proteins, such as stimulated growth factor receptors, lysosomal enzymes and lipids. The MVB pathway appears to require the sequential function of ESCRT-O, -I,-II and -III complexes. ESCRT-III proteins mostly dissociate from the invaginating membrane before the ILV is released. The ESCRT machinery also functions in topologically equivalent membrane fission events, such as the terminal stages of cytokinesis and the budding of enveloped viruses (HIV-1 and other lentiviruses). Key component of the cytokinesis checkpoint, a process required to delay abscission to prevent both premature resolution of intercellular chromosome bridges and accumulation of DNA damage: upon phosphorylation by AURKB, together with ZFYVE19/ANCHR, retains abscission-competent VPS4 (VPS4A and/or VPS4B) at the midbody ring until abscission checkpoint signaling is terminated at late cytokinesis. Deactivation of AURKB results in dephosphorylation of CHMP4C followed by its dissociation from ANCHR and VPS4 and subsequent abscission (PubMed:22422861, PubMed:24814515).

ESCRT-III proteins are believed to mediate the necessary vesicle extrusion and/or membrane fission activities, possibly in conjunction with the AAA ATPase VPS4. Involved in HIV-1 p6- and p9-dependent virus release. CHMP4A/B/C are required for the exosomal release of SDCBP, CD63 and syndecan (PubMed:22660413).

<p>Manually curated information for which there is published experimental evidence.</p> <p><a href="/manual/evidences#ECO:0000269">More. </a></p> Manual assertion based on experiment in i

Supporting Information

S1 Fig. Aversive olfactory learning and/or memory is independent of de-novo protein synthesis.

Training and different treatment protocols are shown at the top of each panel. A: Cycloheximide (CXM) treatment applied before training did not reduce aversive olfactory learning and/or memory of wild type larvae. For all three groups aversive olfactory learning and/or memory tested immediately after three cycle standard training was significantly different from random distribution (One sample t test, p<0.0001, respectively) and not significantly different from each other (One way ANOVA, p = 0.33). For all three groups aversive olfactory learning and/or memory tested 60 minutes after three cycle standard training was significantly different from random distribution (One sample t test, p<0.0001, p = 0.0004, and p<0.0001, respectively) and not significantly different from each other (Kruskal-Wallis, p = 0.29). B: CXM treatment prevented wild type larvae from pupation (after 5 days) and eclosion (after 10 days) after metamorphosis (red line). Control groups (blue lines) that were raised on standard food or on a sucrose diet showed no effect. Results are shown as means and s.e.m. For each group 10 repetitions were done. A significant number of surviving animals is indicated in black (p<0.05), whereas a non-significant number of pupae or flies is marked in light grey (p≥0.05).C: Aversive olfactory learning and/or memory was not affected when interfering with CREB function of mushroom body Kenyon cells (MB KCs) using UAS-dCREb2-b and OK107-Gal4. Experimental OK107-Gal4/UAS-dCREB2b larvae showed learning and/or memory (One sample t test, p<0.0001, respectively) comparable to two genetic controls (Kruskal-Wallis, p = 0.15) when tested immediately after three cycle standard training. Experimental OK107-Gal4/UAS-dCREB2b larvae showed learning and memory tested 60 minutes after three cycle training (One sample t test, p = 0.0004, p = 0.006 and p = 0.002, respectively) comparable to two genetic controls (One way ANOVA, p = 0.64). D: We trained larvae using a spaced training protocol consisting of five training cycles with 15 minutes rest intervals in between. Aversive olfactory learning and/or memory is completely abolished when interfering with CREB function of MB KCs using UAS-dCREb2-b/OK107-Gal4 larvae (One sample t test, p = 0.180). Both genetic controls showed aversive olfactory learning and memory indistinguishable from each other (Tukey post hoc test, p = 0.916). Sample size is n = 16 for each group if not indicated otherwise. In S1A,S1C,S1D Fig. differences between groups are depicted above the respective box plots ns indicates p≥0.05. Different lowercase letters indicate statistical significant differences from each other at level p<0.05. Grey boxes show memory performance above chance level (p<0.05), whereas white boxes indicate random distribution (p≥0.05). Small circles indicate outliers in S1B Fig.

S2 Fig. Drosophila larvae establish an anesthesia resistant type of memory following odor-high salt conditioning.

Training and different treatment protocols are shown at the top of each panel. A: Wild type larvae were trained using the standard three cycle odor-high salt conditioning paradigm. Three different groups were tested: An experimental group received a cold shock (one minute in a 4°C ice water bath) directly after conditioning. A second group received the same cold shock treatment before the conditioning phase. A third group was not cold shocked. Memory was tested after a short recovering phase of 10 minutes after conditioning. Cold shock treatment before or after three cycle standard training did not affect learning and memory of the larvae (One sample t test, p<0.0001, p = 0.0002, p<0.0001 and p<0.0001, respectively). All three groups performed on the same level (One way ANOVA, p = 0.68). B: In addition, we also tested if a harsh cold shock treatment of 5 minutes at 4°C that completely paralyses the larvae affected the performance of the animals. Also under these conditions experimental larvae showed learning and memory that was resistant to cold shock anesthesia (Two way ANOVA, p = 0.07 comparing duration of the cold shock, p = 0.76 comparing if cold shock treatment was applied or not). C: Aversive olfactory learning and memory tested at 10, 60, 120 and 180 minutes after conditioning. Experimental groups received a cold shock directly after three cycle standard training. In all four cases cold shock treated larvae behaved on a comparable level as control groups (Unpaired t test, p = 0.4, p = 0.5, p = 0.68 and p = 0.16, respectively). D: Aversive olfactory memory was tested 60 minutes after three cycle standard training experimental groups received a cold shock 0, 10, 20 or 40 minutes after conditioning. In all four cases cold shock treated larvae behaved as control groups (Unpaired t test, p = 0.5, p = 0.88, p = 0.71 and p = 0.79, respectively). Sample size is n = 16 for each group if not indicated otherwise. In S2A,S2C,S2D Fig. differences between groups are depicted above the respective box plots ns indicates p≥0.05. Grey boxes indicate a performance above chance level (p<0.05). In S2B Fig. differences between groups are depicted below the symbols ns indicates p≥0.05. A performance significant different from random distribution (p<0.05) is indicated in black. The data are shown as means ± s.e.m.

S3 Fig. Sensory acuity tests for larvae with impaired radish gene function.

A: Schematic representation of chemosensory acuity tests. Olfactory perception is analyzed by putting 30 larvae in the middle of a Petri dish with either an amyl acetate (AM) or a benzaldehyde (BA) containing odor container on one side and an empty container (EC) on the other side. After 5 minutes larvae are counted to calculate an olfactory preference index. For gustatory acuity tests, 30 larvae are put in the middle of a Petri dish that contained pure agarose on one side and agarose plus a high salt concentration on the other side. After 5 minutes larvae were counted to calculate a gustatory preference index. B: Naive olfactory and gustatory acuity tests for rsh 1 mutant and wild type control larvae. Olfactory preference for AM of rsh 1 mutant larvae were not different from the one of wild type controls (Unpaired t test, p = 0.30). rsh 1 mutant larvae, however, did not show any significant preference for BA (One sample t test, p = 0.23). Gustatory avoidance for high-salt concentration of rsh 1 mutant larvae is not different from the one of wild type controls (Unpaired t test, p = 0.84). C: Due to the fact that rsh1 mutants showed an impaired BA preference, we applied a one odor paradigm. In contrast to three cycle standard training, larvae received only AM and instead of BA paraffin oil (no odor information) during odor-high salt conditioning (one odor paradigm). In line with the results for two odor conditioning rsh 1 mutant larvae behaved significantly different compared to wild type control larvae (Unpaired t test, p = 0.0001) and showed no learning and/or memory (One sample t test, p = 0.26). The training protocol is shown at the top of the panel. D: Naive olfactory and gustatory acuity tests of experimental and control larvae used to rescue rsh gene function. Olfactory preference for AM for rsh 1 and w,rsh 1 hs-rsh were comparable to their controls at both temperatures (Mann-Whitney test, p = 0.24, p = 0.82, p = 0.22 and p = 0.10, respectively). However, again rsh 1 larvae showed significant differences in their BA preference compared to controls (Unpaired t test, p<0.0001, p = 0.49, p<0.0001 and p = 0.68, respectively). High-salt avoidance of rsh 1 and w,rsh 1 hs-rsh larvae was indistinguishable from the behavior of controls (Mann-Whitney test, p = 0.06, p = 0.36, p = 0.11 and p = 0.34, respectively). Sample size is n = 16 for each group if not indicated otherwise. Differences between groups are depicted above the respective box plots ns indicates p≥0.05 and * p<0.05. Grey boxes indicate a memory performance above chance level (p<0.05), whereas white boxes indicate a memory performance at chance level (p≥0.05). Small circles indicate outliers.

S4 Fig. Presynaptic output of mushroom body Kenyon cells is not necessary for naïve behaviors towards olfactory and gustatory stimuli.

A: Sensory acuity tests when interfering with neuronal output of mushroom body Kenyon cells (MB KCs) using UAS-shi ts1 and OK107-Gal4. Experimental (OK107-Gal4 /UAS-shi ts1 ) and control larvae (OK107-Gal4 /+ and UAS-shi ts1 ) showed no difference in their naïve responses to AM, BA and high salt (for AM: One way ANOVA, p = 0.25, for BA: Kruskal-Wallis, p = 0.09 and for high salt: One way ANOVA, p = 0.08). B: Sensory acuity tests when knocking down brp in the MB KCs via UAS-brp-RNAi B3C8 and OK107-Gal4. Experimental (OK107-Gal4/UAS-brp-RNAi B3C8 ) and control larvae (OK107-Gal4/+, UAS-brp-RNAi B3C8 /+) showed no difference in their naïve responses to AM, BA and high salt (Kruskal-Wallis, p = 0.54, p = 0.27 and p = 0.68, respectively). C: Blockade of presynaptic output of MB KCs via UAS-shi ts1 using another driver line H24 completely impaired aversive olfactory learning and/or memory. Larvae were raised at the permissive temperature (19°C) and shifted to restrictive temperature before and during three cycle standard training and testing. In contrast to both genetic controls aversive olfactory learning and/or memory tested immediately after three cycle standard training was completely abolished in H24-Gal4/UAS-shi ts1 larvae (One sample t test, p<0.0001 for both control groups and p = 0.853 for H24-Gal4/UAS-shi ts1 ). D: Sensory acuity tests when interfering with neuronal output of mushroom body Kenyon cells (MB KCs) using UAS-shi ts1 and H24-Gal4. Experimental (H24-Gal4/UAS-shi ts1 ) and control larvae (H24-Gal4/+, UAS-shi ts1 /+) showed no difference in their naïve responses to AM, BA and high salt (One way ANOVA, p = 0.517, p = 0.184 and p = 0.753, respectively). E: Shows a frontal view projection (left) of a H24-Gal4/UAS-mCD8::GFP larval hemispheres labeling the entire set of MB KCs (anti-GFP in green and anti-FasII, anti-ChAT neuropil staining in magenta). The observed staining is nearly specific for the larval MB. Below a zoom in of the MB is shown. Further below only the GFP channel is depicted. Scale bars: upper panel 50μm, middle and lower panel 25μm. Sample size is n = 16 for each group if not indicated otherwise. Differences between groups are depicted above and below the respective box plots ns indicates p≥0.05. Grey boxes indicate a memory performance above chance level (p<0.05). Small circles indicate outliers.

S5 Fig. Sensory acuity tests of larvae with impaired PKC function.

A: Sensory acuity tests when suppressing PKC activity in the MB KCs using an inhibitory pseudo substrate of PKC (PKCi). For more details see S3A Fig. The inhibition of PKC did not alter the perception of AM, BA or high salt. Experimental and control groups were indistinguishable from each other (One way ANOVA, p = 0.99, p = 0.40 and p = 0.50, respectively). Sample size is n = 16 for each group if not indicated otherwise. Differences between groups are depicted above or below the respective box plots, ns indicates p≥0.05. Grey boxes indicate a memory performance above chance level (p<0.05). Small circles indicate outliers.

S6 Fig. Sensory acuity test for larvae with impaired dopaminergic signaling.

Training and methylphenidate (MPH) treatment protocols are shown at the top of each panel. A: Sensory acuity tests of the dopamine (DA) receptor mutants dumb 2 and damb. Both receptors mutants perceived AM, BA and high salt stimuli comparable to controls (for AM: unpaired t test, p = 0.25 and p = 0.83, for BA: Mann-Whitney test, p = 0.07 and unpaired t test, p = 0.41, for high salt: unpaired t test, p = 0.57 and p = 0.28). B: Sensory acuity tests of the DAT mutant fumin (fmn). fmn mutant larvae showed no difference in their naïve responses to AM and high salt (Unpaired t test, p = 0.10 and p = 0.64, respectively). However, the naïve response to BA was impaired (Unpaired t test, p = 0.06). C: In line with the results for two odor conditioning, fmn mutant larvae using a one odor learning paradigm (to omit BA as a sensory stimulus) behaved significantly different compared to wild type control larvae (paired t test, p = 0.0003). They showed no aversive olfactory learning and/or memory (One sample t test, p = 0.06). D: Aversive olfactory memory in rsh 1 and wild type control larvae after methylphenidate (MPH) treatment using different concentrations. Memory was tested directly after three cycle standard training. Aversive olfactory learning and memory was indistinguishable from random distribution for the rsh 1 mutant without MPH application (One sample t test, p = 0.94). After application of 0.5 mM MPH rsh 1 mutant larvae showed reduced learning and/or memory compared to wild type controls (One sample t test, p = 0.03, unpaired t test, p = 0.01). After application of 2.0 mM MPH rsh 1 mutant larvae performed as wild type controls (Unpaired t test, p = 0.11). Sample size is n = 16 for each group if not indicated otherwise. Differences between groups are depicted above or below the respective box plots ns indicates p≥0.05 and * p<0.05. Grey boxes indicate a memory performance above chance level (p<0.05), whereas white boxes indicate a memory performance at chance level (p≥0.05). Small circles indicate outliers.

S7 Fig. Aversive olfactory high salt reinforced learning and memory using only one training trial.

Training and cold shock treatment protocols are shown at the top of each panel. A: Aversive olfactory learning and/or memory using a one cycle training protocol was tested at four different retention times after conditioning (0, 10, 20 and 60 minutes). Statistical significant differences were revealed between the groups (One way ANOVA, p<0.0001). The performance indices measured immediately and 10 minutes after one cycle training were indistinguishable from each other (Tukey post hoc test, p = 0.945). Alike the performance indices measured 20 and 60 minutes after once cycle training showed no statistical significance differences (Tukey post hoc test, p = 0.874). However the performance indices measured 0 and 10 minutes after conditioning where at a higher level than the ones measured 20 and 60 minutes after conditioning when analyzed with Tukey post hoc test (p<0.0001 for 0 and 20, p<0.0001 for 0 and 60, p = 0.0001 for 10 and 20 and p<0.0001 for 10 and 60). B: Aversive olfactory learning and memory established with increasing training cycles (One cycle, two cycles and three cycles) revealed significant differences between one cycle training and two or three cycle training (One way ANOVA, p = 0.003 Tukey post hoc test p = 0.037, p = 0.033, respectively). For two and three cycle training no difference was detected (p = 0.705). C: Cold shock treatment directly after training partially impaired aversive olfactory learning and/or memory when tested 10 minutes after one cycle training. No difference was detected between cold shock treated and control larvae tested 10 minutes after two and three cycle training (Unpaired t-test, p = 0.0001 for one training cycle, p = 0.623 for two training cycles and p = 0.396 for three training cycles). D: Aversive olfactory learning and memory of cold shock treated larvae and control groups tested 10, 20 and 60 minutes after one cycle training. Only cold shocked treated larvae tested 10 minutes after one cycle training showed a significant reduction compared to controls that perceived no cold shock. No effect was seen between cold shocked and control groups tested 20 and 60 minutes after one cycle training (Unpaired t test, p = 0.0001 for 10 minutes, p = 0.934 for 20 minutes and p = 0.681 for 60 minutes). E: Aversive olfactory learning and/or memory of rsh 1 mutants measured immediately and 20 minutes after one cycle training was reduced compared to wild type larvae (Unpaired t test, p<0.0001 for both). Furthermore, rsh 1 mutant revealed a complete loss of aversive olfactory learning and/or memory only when measured 20 minutes after one cycle training but not when measured immediately after one cycle training (One sample t-test, p = 0.383). F: Aversive olfactory learning and/or memory of rut 2080 and dnc 1 mutants was reduced compared to wild type larvae when measured directly after one cycle training (One way ANOVA, p<0.0001 Tukey post hoc test, p<0.0001 for both compared to wild type control and p = 0.902 compared to each other). For both mutants learning and memory was not completely abolished (One sample t test, p<0.0001 for both). Aversive olfactory learning and memory of both mutants tested 20 minutes after one cycle training was indistinguishable from the one of wild type larvae (One way ANOVA, p = 0.106).


Fungal strains, culture conditions, and transformation:

C. cinerea strains were standardly grown at 37° on YMG/T complete medium and minimal medium (G ranado et al. 1997) supplemented with p-aminobenzoate (PABA, 5 mg/liter) when required. Strain 6-031 (A43mut, B43mut, pab1-1, skn1) is a fruiting body initiation mutant generated from homokaryon AmutBmut (A43mut, B43mut, pab1-1) by UV-mutagenesis (U. K ües , J. D. G ranado and M. A ebi , unpublished results). Monokaryons PG78 (A6, B42, pab1) and JV6 (A42, B42), both unrelated to homokaryon AmutBmut, and the AmutBmut co-isogenic monokaryons PS001-1 (A42, B42) and PS002-1 (A3, B1) were used in crosses (K ertesz -C haloupková et al. 1998 P. S rivilai and U. K ües , unpublished results). Matings were performed on YMG/T plates by placing two mycelial blocks of inoculum 5 mm apart. For growth and induction of fruiting bodies, mating plates were incubated at 12-hr light/12-hr dark, 25°, 90% humidity under standard fruiting conditions (G ranado et al. 1997). Randomly isolated basidiospores were germinated on YMG/T medium at 37°. Progeny of crosses were analyzed on minimal media for PABA auxotrophy. Presence of unfused and fused clamp cells, indicators of activated A and B mating-type pathways, respectively (K ües 2000), was determined by microscopy.

Monokaryon JV6 served to confirm mating types in A43mut, B43mut progenies from crosses with PS001-1 and PS002-1 that subsequently were submitted to fruiting tests. Frequencies of phenotypic distributions in A43mut, B43mut progenies were tested by a chi-square method. A skn1 + , mat + , bad clone (PS-Mu1-3) and a skn1 + , mat, bad clone (PS-Mu1-2) within the A42, B42 progeny of cross 6-031 × PS001-1 (defined by crosses with monokaryon PG78) were identified through mating with mutant 6-031. Mating of these two strains with A43mut, B43mut clones of the progeny 6-031 × PS001-1 that did not initiate fruiting identified homokaryon OU3-1 (A43mut, B43mut, pab1-1, skn1).

The F1 progeny of cross 6-031 × JV6 was randomly analyzed for fruiting ability by individually inoculating clones on YMG/T agar, growing them for 4 days at 37° in the dark, and subsequently transferring them to standard fruiting conditions. Dikaryons among the clones were identified by light inducing oidia production, germinating the spores on YMG/T agar, and analyzing pab-auxotrophy on minimal medium. For oidia induction of dikaryons, mutant 6-031 and other A43mut, B43mut strains, dark-grown cultures were exposed to light for 2 days (K ertesz -C haloupková et al. 1998) and the number of oidia per plate was determined by a spectrophotometer.

DNA transformation was performed as described (G ranado et al. 1997). For selecting PABA prototrophs in cotransformations, 1 μg of plasmid pPAB1-2 (G ranado et al. 1997) was added. Upon germination on regeneration agar, transformants were individually transferred onto minimal medium for further growth. Subsequently, three or four individual transformants were inoculated on YMG/T agar per single petri dish and grown in the dark at 37° for 2 days to a colony size of 3–3.5 cm in diameter. To induce fruiting, plates were moved for 2 weeks to standard fruiting conditions. The number and size of primordia per transformant were scored. A small piece of gill tissue from primordia developed upon transformation with cosmid 40-5A was spread and nuclei in basidia stained with hematoxylin (L u and R aju 1970).

DNA and RNA techniques:

An indexed genomic cosmid library derived from homokaryon AmutBmut was transformed into mutant 6-031 and screened for cosmids that were able to restore fruiting ability in this strain, following a SIB-selection procedure. The pab1 + wild-type gene of C. cinerea present in the cosmid backbone was used as a selection marker. Cosmid DNAs from 60 pools of each 96-well microtiter dish-arranged E. coli clones, and from subpools and individual clones of microtiter dish 40 were isolated (B ottoli et al. 1999).

Cloning was performed by standard methods (S ambrook et al. 1989). Plasmids were propagated in E. coli strain XL1-Blue (Stratagene, La Jolla, CA). Derivative pSphA of cosmid 40-5A is a ligation product between a 16-kb SphI fragment (13-kb genomic DNA + 3-kb cosmid backbone) and a 7.5-kb SphI fragment (2-kb genomic DNA + 5.5-kb cosmid backbone) in their natural order. NotI fragments of cosmid 40-5A were cloned into the NotI site of pBC SK (+) (Stratagene). Plasmids pNotB5 and pNotB7 contained the same DNA insert but in opposite orientation. The insert in pNotB5 was sequenced on both strands by primer walking (Microsynth, Balgach, Switzerland). Sequences were assembled with the program DNASTAR and analyzed with OMIGA 2.0, BLAST (NCBI). The whole sequence but 32 bp originating from the linker of the cosmid was submitted to GenBank (AF338438).

pNotB5 and pNotB7 gave rise to the following pBC SK (+) subclones: p5SmaCS and p5BamCS (used in Northern analysis) carry gene arf1 on a 1.4-kb NotI–SmaI and a 3.5-kb NotI–BamHI fragment, respectively. p5EcoCS and p5XbaCS contain arf1 and a truncated cfs1 + gene on a 3.8-kb NotI–EcoRI and a 5.5-kb NotI–XbaI fragment, respectively. p5SpeCS includes both arf1 and cfs1 + on a 7-kb SpeI fragment, whereas cfs1 + is truncated in p5KpnCS on the shorter 4.4-kb NotI–KpnI fragment. p7XbaCS carries truncated cfs1 + and kin1 copies on a 5-kb XbaI–NotI fragment. p7SpeCS contains a truncated kin1 on a 3.5-kb SpeI–NotI fragment. Subclones constructed in pBluescript KS (−) (Stratagene) were as follows: pPvu8.5 contains an 8.5-kb PvuII fragment covering the complete cfs1 + gene and the 3′ end of kin1. pBam3.5 and pSmaSpe5.5 carry cfs1 + on a 3.5-kb BamHI and a 5.5-kb SmaI–SpeI insert, respectively. pEco4.4 contains truncated cfs1 + and kin1 copies on a 4.4-kb EcoRI fragment. Furthermore, the 8.5-kb PvuII fragment, 3.5-kb BamHI fragment, and 4.4-kb EcoRI fragment were also cloned into pPAB1-2 containing the C. cinerea pab1 + gene, resulting in pPvu8.5-pab, pBam3.5-pab, and pEco4.4-pab, respectively.

A cut-and-shut strategy using NcoI and plasmid pNotB7 resulted in p7NcoCSΔcfs with a deletion in cfs1 + (Δbp 5296–5476). Similarly, p5BstCSΔkin with a deletion in kin1 (Δbp 8502–8619) were created from plasmid p5NotB5 by using BstEII. An AatII deletion (Δbp 6567–6803) in p5SpeCS yielded p5SpeCSΔgtl. An NruI deletion in arf1 (Δbp 676–972) in p5SmaCS gave rise to p5SmaCSΔarf, from which the SmaI insert was cloned into pSmaSpe5.5 to generate p5SpeCSΔarf. The insertion of a 3.5-kb SpeI fragment from pNotB7 at the SpeI site in p5SpeCSΔarf resulted in pNotB5Δarf. The T-to-G transversion found in the cfs1 allele of mutant 6-031 was introduced into plasmid p5SpeCS by exchanging a 1-kb PCR-amplified StuI–NdeI fragment with the wild-type sequence, yielding p5SpeCS/6-031. pNotB5/6-031 distinguishes from pNotB5 by the same T-to-G transversion.

Genomic DNA of C. cinerea strains was isolated from powdered lyophilized mycelium (Z olan and P ukkila 1986). Two overlapping fragments containing the cfs1 allele of mutant 6-031 were independently amplified six times from genomic DNA with specific primers (a 3.1-kb fragment using primers 5′ TCAAGTCGGGTCGGTAGAAG 3′ and 5′ TTTGTTTCGGAGCTTGACTG 3′ and a 1.1-kb fragment using primers 5′ GGACGCTTCAAGATTAGATC 3′ and 5′ CTCTGAAGGAATCGCTCTTG 3′) and sequenced using a ABI PRISM DNA Sequencing Kit and a Model 373A DNA sequencer (Perkin-Elmer). Sequences of PCR products separately amplified with the same primer set were identical. Presence of the same sequence in p5SpeCS/6-031 has also been verified by sequencing.

Southern blot analysis was performed with 10 μg of genomic DNA per sample following basic protocols (S ambrook et al. 1989). Total RNA of strain AmutBmut was extracted with a guanidinium isothiocyanate procedure (C homczynski and S acchi 1987) from powdered lyophilized C. cinerea mycelia or tissues of different fruiting stages. Poly(A) + RNA was isolated with the Oligotex mRNA Midi kit (QIAGEN). Per sample, 10 μg of total RNA or 2.5 μg of poly(A) + RNA were used for Northern blot analysis (S ambrook et al. 1989). Hybridization signals in Southern and Northern blot analyses were produced with DNA fragments labeled with [α- 32 P]dCTP by random primed DNA labeling (Boehringer Mannheim).

The 5′ and 3′ cDNA ends of the cfs1 + gene were determined with the 5′/3′ RACE kit (Roche Molecular Biochemicals) following the instructions of the manufacturer. Poly(A) + RNA from 5-mm-sized primordia of homokaryon AmutBmut was used for cDNA synthesis. In the 5′ RACE, a cfs1 specific primer sp1 (5′ ACAATGCACAGGAGTACATC 3′) was employed to synthesize the first strand cDNA. Two cfs1-specific primers, sp2 (5′ GCAATGGCATTGAGTCGAG 3′) and sp3 (5′ TAGACGATAGGGTCATCTCC 3′), were applied in subsequent PCR reactions. In the 3′ RACE, two cfs1-specific primers, sp4 (5′ GATTTTGCCCTCAAGCCAC 3′) and sp5 (5′ CAATTCGAGCCTGCCCAG 3′), were used. RACE products were cloned into pBluescript KS (−) by T/A cloning (M archuk et al. 1991) and sequenced with a Model 373A DNA sequencer. The full coding length of the cfs1 + cDNA was obtained by PCR, using the two primers cfsATG (5′ ATGCCGGCCCACCACCACCCTTC 3′) and cfsREV (5′ CGCCGAGGCCGCCGTGTAAACAC 3′). For sequencing, the PCR product was cloned into the EcoRV site in the β-galactosidase gene of pBluescript KS (−) via T/A cloning, resulting in construct pYL28 having the cfs1 + cDNA inserted in frame to the β-galactosidase gene.

Computer analysis of protein sequences:

Proteomics tools provided by ExPaSy Molecular Biology Server (Swiss Institute of Bioinformatics, Geneva) were used to perform protein pattern and profile searches (InterPro), transmembrane region detection (TMpred and TMHMM), and secondary structure predication (PSA and PSIpred). Hydrophilicity profile was calculated with Goldman/Engelman/Steitz parameters in OMIGA 2.0.


The purpose of this investigation was to improve our understanding of how genetic architecture, in particular recombination and locus effects, as well as the pattern and amount of migration determine polymorphism, local adaptation, and differentiation in a subdivided population inhabiting a heterogeneous environment. For simplicity, we restricted attention to two linked, diallelic loci and to migration between two demes. The study of diversifying selection in just two demes may also shape our intuition about clinal variation if the two subpopulations are from different ends of the cline. If alleles are beneficial in only one environment and detrimental in the other, local adaptation of subpopulations and differentiation between them can be obtained only if a (multilocus) polymorphism is maintained. Therefore, most of our mathematical results focus on existence and stability of polymorphic equilibria and on the dependence of the equilibrium configurations on the model parameters (migration rates, selection coefficients, recombination rate).

The model is introduced in Sect. 2. Sections 3 and 4 are devoted to the derivation of the possible equilibrium configurations and bifurcation patterns. They contain our main mathematical results. Explicit analytical results about existence and stability of equilibria were obtained for several limiting or special cases and are complemented by numerical work.

The conditions for admissibility of all single-locus polymorphisms (SLPs) are given in Sect. 3.1, those for asymptotic stability of the monomorphic equilibria (ME) in Proposition 3.1 in Sect. 3.2. The stability of SLPs could not generally be determined (Sect. 3.3). Weak migration is treated by perturbation methods in Sect. 3.4. For sufficiently weak migration, there exists a globally attracting fully polymorphic equilibrium, (mathsf F ) (Proposition 3.2). Its approximate coordinates are given by (3.16).

The complete equilibrium and stability structure could be derived under the assumption of linkage equilibrium (Sect. 3.5). The unique, fully polymorphic equilibrium (mathsf F =mathsf) is admissible and globally attracting if and only if all four SLPs are admissible. Otherwise, one boundary equilibrium (SLP or ME) is globally asymptotically stable (Proposition 3.3). These results extend straightforwardly to an arbitrary number of diallelic loci. Based on these results, nonlinear perturbation theory establishes the existence of a globally stable, fully polymorphic equilibrium in a perturbed parameter range if recombination is sufficiently strong (Sect. 3.6). This equilibrium is in quasi-linkage equilibrium and given by (3.21).

Also for completely linked loci all equilibria and their local stability properties could be derived (Sect. 3.7). In this case, the fully polymorphic equilibrium (mathsf) (3.24) may lose stability while it is admissible (3.28). At this threshold a boundary equilibrium becomes stable by a ‘jump bifurcation’ (Proposition 3.4). In general, however, more complicated equilibrium patterns than determined by Propositions 3.3 and 3.4 can occur, in particular, multiple stable equilibria.

In Sect. 3.8, we apply perturbation theory to infer the equilibrium properties under highly asymmetric migration from those derived for the continent-island model in Bürger and Akerman (2011) and Bank et al. (2012). There, a stable ( (mathsf F ) ) and an unstable fully polymorphic equilibrium may exist if recombination is intermediate, and (mathsf F ) is simultaneously stable with a boundary equilibrium. In general (Sect. 3.11), we cannot exclude the existence of more than two internal equilibria or complicated dynamical behavior. Numerical searches produced no such instances. What can be shown easily is that, if ( ho <infty ) , any fully polymorphic equilibrium exhibits LD. In all cases, where an internal equilibrium was calculated (numerically or analytically), it exhibited positive LD.

In the super-symmetric case, in which selection in deme 2 mirrors that in deme 1 and migration is symmetric, an assumption made in several applications, a fully polymorphic equilibrium exists always and, presumably, is stable (Sect. 3.10). This is a highly degenerate situation because if ( heta e 0) , only a monomorphic equilibrium can be stable for sufficiently large migration rates (Proposition 4.3). If ( heta =0) (Sect. 3.9), then a fully polymorphic equilibrium can exist for arbitrarily large migration rates if (phi =phi ^mathsf) (see also Sect. 4.11).

Whereas in Sect. 3 the focus was on the efficient presentation of the existence and stability results of equilibria, in Sect. 4 these results are used to derive the possible bifurcation patterns with the total migration rate (m) as the bifurcation parameter. All possible bifurcation patterns could be derived under the assumption of LE (Theorem 4.4, Figs. 2, 3), and under the assumption of complete linkage (Theorem 4.6, Figs. 4, 5). The latter case is considerably more complex. Interestingly, in each case, every bifurcation pattern can occur for every ratio (phi =m_1/m) of migration rates by choosing the selection coefficients appropriately. Hence, the assumption of symmetric migration does not yield simpler equilibrium configurations than general migration if arbitrary selection coefficients are admitted.

In each of these cases (LE or ( ho =0) ), we determined the maximum migration rate (m_) admitting an asymptotically stable, fully polymorphic equilibrium (Corollaries 4.5 and 4.7). The maximum migration rate (m_^0) for ( ho =0) always exceeds or equals that ( (m_^infty ) ) for LE, i.e., (m_^infty le m_^0) . Although for strong recombination, (m_) can be very slightly smaller than (m_^infty ) (Sect. 4.7), in the vast majority of investigated cases, (m_) is bracketed by (m_^infty ) and (m_^0) (Fig. 6, Sect. 4.11).

Proposition 4.3 demonstrates that a ME is globally attracting if migration is sufficiently strong (except in the degenerate case noted above). If we interpret the equilibria (mathsf M _2) and (mathsf M _3) as fixation of a generalist ( (A_1B_2) and (A_2B_1) are haplotypes of intermediate fitness), and (mathsf M _1) and (mathsf M _4) as fixation of a specialist ( (A_1B_1) and (A_2B_2) are the locally adapted haplotypes), then depending on the sign of ( heta ) one of the generalists becomes fixed for high (m) if (phi ) is intermediate (i.e., (phi ^mathsf<phi <phi ^mathsf) if ( heta >0) , (phi ^mathsf<phi <phi ^mathsf) if ( heta <0) but note that, depending on the selection coefficients, both (phi ^mathsf) and (phi ^mathsf) can be arbitrarily close to 0 or 1.). The critical value (m) as well as (phi ^mathsf) and (phi ^mathsf) are independent of ( ho ) . Otherwise, one of the specialists becomes fixed for large (m) .

The fact that a generalist becomes fixed for strong migration is a distinct feature of (balanced) two-way migration: in the CI model or if migration is sufficiently asymmetric ( (phi <phi ^mathsf) or (phi >phi ^mathsf) if ( heta >0) ), one of the specialist haplotypes swamps the populations and becomes fixed. Another difference between highly asymmetric and more symmetric migration patterns is that in the first case, it is always the locus under weaker selection that first loses its polymorphism while (m) increases, whereas this not necessarily so in the latter case (see Sect. 4.2 and Theorem 4.6, cases A3 and A4).

In summary, we determined quantitatively when the following three evolutionarily stable states discussed by Kawecki and Ebert (2004) occur: (i) existence of a single specialist optimally adapted to one deme and poorly to the other, (ii) existence of a single generalist type which has higher average fitness in the whole population than than any of the specialists, and (iii) existence of a set of specialists each adapted to its deme, i.e., coexistence in a polymorphism. Local adaptation and differentiation occur only in case (iii).

In Sect. 5, we used the migration load in each deme to quantify the degree of local adaptation. In Sect. 6 we introduced a new multilocus version of (F_mathrm) to measure differentiation. If migration is weak, then local adaptation and differentiation decrease with increasing migration rate and increase with increasing linkage between the loci (Fig. 7). In particular, for given (small) migration rate, local adaptation and differentiation are maximized if the fitness effects are concentrated on a single locus (corresponding to ( ho =0) in our model). However, as discussed in Sect. 5, for high migration rates, the migration load of the total population can decrease with increasing recombination or migration rate. Similarly, at high recombination and migration rates, (F_mathrm) can increase with increasing migration or recombination rate. Thus, for given, relatively high migration rate, (F_mathrm) may be minimized at intermediate recombination rates. Apparently, it is always maximized in the absence of recombination.

In Sect. 7, we investigated the conditions for invasion of locally beneficial mutants. At an isolated locus, such a mutant can invade and become established in a migration-selection equilibrium if and only if its advantage exceeds a threshold that increases with the immigration rate of the wild type see (7.3b). If, however, this mutant occurs at a locus that is linked to a locus that is already in migration-selection balance, then its invasion is facilitated, i.e., its local selective advantage can be smaller (Fig. 8b). Equivalently, for given selection coefficients and total migration rate, the minimum recombination rate needed for invasion increases if (phi ) , or the influx of the (deleterious) wild type relative to the efflux of the new mutant, increases (Fig. 8a). For the extreme case of one-way migration from a ‘continental’ population to an ‘island’ population that is adapting to a new environment, Bürger and Akerman (2011) proved that invasion of a locally beneficial mutant is always facilitated by increased linkage to a locus in migration-selection balance.

Thus, our results complement the numerical finding by Yeaman and Whitlock (2011) for a multilocus quantitative-genetic model that clusters of locally adaptive mutations, or concentrated genetic architectures, build up in spatially structured populations with opposing selection pressures in two demes. Because tighter linkage is required for invasion under increasingly asymmetric migration rates, more concentrated architectures and a greater advantage for recombination-reducing mechanisms (such as chromosome inversions) should be expected for highly asymmetric migration. In finite populations, invasion of new mutants occurs only with a certain probability, and genetic drift may erase polymorphism. Numerical work, supported by analytical methods, has already shed some light on the dependence of the probability of establishment of new, locally adaptive mutations on the recombination rate and other factors (Yeaman and Otto 2011 Feder et al. 2012). Analytical work on the role of genetic drift and finite population size on these issues is in progress.

Our results also show that, in the absence of epistasis and under the present form of balancing selection, reduced recombination between selected loci is favored, except when migration rates are sufficiently symmetric and high (Sect. 5). Selection inducing certain forms of epistasis may favor high recombination in structured populations more easily (Pylkov et al. 1998 Lenormand and Otto 2000 Bank et al. 2012). Therefore, general predictions about the emergence of clusters of locally adaptive mutations in regions of reduced recombination, or of genomic islands of speciation (Wu and Ting 2004) or of differentiation (Feder et al. 2012), can not be made in the absence of detailed information about epistasis and the spatial pattern of selection and migration. At least in the absence of epistasis, the most favorable situation for the emergence of such clusters should occur in populations that are adapting to a new environment, still receiving maladaptive gene flow but sending out only very few or no migrants (corresponding to a continent-island model).

In Sect. 8, we derived the approximation (8.3) for the effective migration rate at a linked neutral locus that is located between the selected loci. This approximation is simply the sum of the two effective migration rates under one-way migration (Bürger and Akerman 2011). Because in the present model, polymorphism at the selected loci is maintained by balancing selection, the effective migration rate may be greatly reduced compared with the actual migration rate (see Fig. 9). Thus, strong barriers against gene flow may build up at such neutral sites and enhance (neutral) differentiation (see Charlesworth and Charlesworth 2010, Chap. 8.3). Future work will have to study the actual amount and pattern of neutral diversity at such sites in finite populations.

Watch the video: Mutations. Genetics. Biology. FuseSchool (December 2021).