Population bottlenecks

We have noted already that a population bottleneck will reduce the effective size of a population and hence its overall level of genetic diversity. We are now returning to bottlenecks for a more in-depth discussion because, as mentioned previously, they are one of the most important determinants of a population's genetic diversity and therefore merit further discussion. The methods outlined previously for quantifying the effect of bottlenecks on Ne, and hence genetic diversity, required long-term demographic data. Because such data are seldom available, potential relationships between bottlenecks and genetic variation are more likely to be inferred from situations in which biologists discover a genetically depauperate population for which no records exist, and then have to try to decide whether or not a past bottleneck can explain why current genetic diversity is so low.

Detecting past reductions in population size is seldom straightforward. For one thing, the severity of any population bottleneck depends on both the size that a population is reduced to and the speed at which it recovers. In general, the initial loss of alleles is proportional to the reduction in population size. Genetic drift means that diversity will continue to deplete while the population size remains low, and as a result populations that take longer to rebound will generally lose more genetic diversity than populations that recover rapidly.

The detection of bottlenecks is confounded further by the fact that not all measures of genetic diversity will show a uniform decrease. On the one hand, allelic diversity usually decreases after a bottleneck because rare alleles will be lost. This often is associated with a drop in He because fewer alleles typically lead to reduced expectations of heterozygosity under HWE. At the same time, Ho values may not deplete and in fact there may be a temporary increase in observed heterozygosity compared with expectations under HWE. In some cases, observed heterozygosity excess can be a useful indicator of past bottlenecks, although this test requires a large number of polymorphic loci and can detect only relatively recent bottlenecks (Luikart and Cornuet, 1998).

If samples are available from two or more generations then we may be able to infer past bottlenecks from the variation in allele frequencies between generations. We know that if a population has undergone a bottleneck its Ne will be reduced. Because the rate of genetic drift is inversely proportional to a population's Ne, a bottleneck will accelerate drift. This will lead to an inflated variance in allele frequencies, which may be taken as evidence that a bottleneck occurred between two sampling periods. Simulations suggest that the temporal variance method provides an 85 per cent probability of detecting a bottleneck after a single generation, provided that data can be obtained from at least five highly variable loci and at least 30 individuals from each sampling period (Luikart et al., 1998). However, the same study also suggested that the temporal variance test may not be useful for populations that have undergone prolonged bottlenecks (>3-- 5 generations) because these populations will continue to lose alleles. Once an allele is lost, its frequency must remain at zero and therefore it cannot exhibit any variance between generations. Prolonged bottlenecks mean that a relatively large proportion of alleles will go extinct and therefore will show no change in frequency over time, and as a result Ne will be overestimated.

Several methods for inferring past bottlenecks were evaluated in a series of experimental populations of the western mosquitofish Gambusia affinis (Spencer, Neigel and Leberg, 2000). The authors of this study set up multiple populations that were founded by 2--16 individuals, and after two or three generations they quantified genetic diversity in a number of different ways. The most useful measure was the temporal variance in allele frequencies, because the degree to which allele frequencies changed was inversely proportional to the size of the founding population. Allelic diversity and He also proved to be suitable for detecting bottlenecks (Figure 3.8), but neither Ho nor the proportion of polymorphic loci showed any relationship with the severity of the bottleneck.

B)

0.9

16

0.8

Founder Population

14

0.7

2?

12

0.6

10

0.5

<u >

8

0.4

t5

0.3

6

0.2

<

4

0.1

2

0.0

0

0 2 4 6 8 10 12 14 16 18 Number of founders

0 2 4 6 8 10 12 14 16 18 Number of founders

0 2 4 6 8 10 12 14 16 18 Number of founders

Figure 3.8 Relationship between the size of a founding population and (A) expected heterozygosity (He) and (B) allelic diversity (A), both of which were measured after two or three generations in experimental populations of the western mosquitofish Gambusia affinis. Dashed lines show the relevant measures of genetic diversity in the founding population (data from Spencer, Neigel and Leberg, 2000)

It is important to keep in mind that the results from the experimental western mosquitofish populations will not be transferable to all populations. The most appropriate indicator of past bottlenecks will depend on a number of factors, including the variability of the marker that is used and both the recency and severity of the bottleneck. A related study of the eastern mosquitofish (G. holbrooki), based on allozyme data, found that temporal variance in allele frequencies may be misleading if only a few loci are characterized or if the population lost many alleles during a bottleneck (Richards and Leberg, 1996). In this study, changes in allelic diversity were tied more closely to bottlenecks than were changes in allele frequencies. It is therefore advisable, whenever possible, to obtain multiple estimates of genetic diversity before inferring past bottlenecks (see also Box 3.4).

Conflicting results have also been found when nuclear and organelle markers are used to evaluate the effects of bottlenecks. The relatively low population size of organelle genomes means that they are more sensitive to bottlenecks and may provide exaggerated estimates of post-bottleneck reductions in genetic diversity. This is illustrated by Scandinavian brown bears (Ursus arctos), which underwent a well-documented population bottleneck between the mid-1800s and 1930, during which time the population of around 5000 plummeted to approximately 130 bears distributed among four subpopulations. Today, the total population is thriving (approximately 1000 bears), but there has been some concern about current levels of genetic diversity.

A recent investigation revealed only two mtDNA haplotypes in 380 bears, which likely reflects the loss of most haplotypes during the bottleneck of the late 19th century. At the same time, data from 19 microsatellite loci revealed estimates of He comparable with those found in North American brown bear populations that have not undergone severe bottlenecks (Waits et al., 2000) (Table 3.7 and Figure 3.9). Although the bottleneck in Scandinavia was severe enough for most mtDNA haplotypes to be lost, multiple microsatellite alleles appeared to be maintained within each of the four subpopulations. In this case, mitochondrial data suggest that bottlenecks have had a lasting impact on genetic diversity in Scandinavian brown bears, whereas nuclear data show that the impact on highly polymorphic markers has been minimal.

Table 3.7 Comparison of He values (based on microsatellite data) from four post-bottleneck subpopulations of Scandinavian brown bears (Waits et al., 2000) and five North American brown bear populations that have not experienced severe bottlenecks (Paetkau etal., 1998); Nc is the estimated census size of each population at the time that samples were taken for genetic analysis

Table 3.7 Comparison of He values (based on microsatellite data) from four post-bottleneck subpopulations of Scandinavian brown bears (Waits et al., 2000) and five North American brown bear populations that have not experienced severe bottlenecks (Paetkau etal., 1998); Nc is the estimated census size of each population at the time that samples were taken for genetic analysis

Population

Nc

He

Scandinavian subpopulation 1

108

0.70

Scandinavian subpopulation 2

29

0.69

Scandinavian subpopulation 3

156

0.68

Scandinavian subpopulation 4

88

0.67

Kluane, Yukon

50

0.76

Flathead River, British Columbia/Montana

40

0.69

Richardson Mountains, Northwest Territories

119

0.75

East Slope, Alberta

45

0.67

Yellowstone, Montana/Wyoming

57

0.55

Figure 3.9 A North American brown bear (Ursus arctos). The genetic diversity of North American and Scandinavian brown bears is comparable at microsatellite loci, even though the latter underwent a severe bottleneck in the late 19th century. This bottleneck did, however, drastically deplete mitochondrial haplotype diversity, which is now much lower in Scandinavian compared with North American populations. Author's photograph

Figure 3.9 A North American brown bear (Ursus arctos). The genetic diversity of North American and Scandinavian brown bears is comparable at microsatellite loci, even though the latter underwent a severe bottleneck in the late 19th century. This bottleneck did, however, drastically deplete mitochondrial haplotype diversity, which is now much lower in Scandinavian compared with North American populations. Author's photograph

Box 3.4 The long-term decline of the Hawaiian nene

Knowledge of a past bottleneck, coupled with currently low levels of genetic variation, does not necessarily imply cause and effect because genetic variation may have been reduced by either multiple bottlenecks or a gradual and prolonged decline in population size. Long-term genetic data produced some surprising results in a study of the population genetic history of the Hawaiian goose, or nene (Branta sandvicensis). The nene once occurred on most of the main Hawaiian islands, but by the time Captain Cook arrived in 1778 it was found only on the island of Hawaii. By the middle of the 20th century there were fewer than 30 surviving nene.

In order to assess the impact of this recent decline on genetic diversity, Paxinos et al. (2002) compared sequence variation in the mitochondrial control region from samples representing four different time periods: current, museum specimens (dating from 1833 to 1928), archaeological samples (160--500 radiocarbon years before present) and palaeontological samples (500-2540 radiocarbon years before present). All but one of the current, museum and archaeological samples had the same haplotype, but 7 haplotypes were found in the 14 palaeontological samples. This translates into haplotype diversity values of zero in the current population, 0.067 in the archaeological and museum specimens combined and 0.802 in the palaeontological samples. The researchers concluded that the nene lost most of its haplotype diversity during a period of prehistoric human population growth, and as a result had very little mitochondrial diversity even before its recent population decline. A direct approach such as this can provide valuable insight into how historical processes have affected genetic diversity, although unfortunately obtaining data from multiple time periods is seldom a practical option.

Founder effects

The effects that both the size of a bottleneck and the speed of the population's recovery can have on the long-term genetic diversity of a population have been well illustrated by a number of studies on species that have been introduced by humans into new geographical areas such as islands. These introductions involve a particular type of bottleneck known as a founder effect, so-called because the founders of a new population will carry only a portion of the genetic diversity that was present in the source population.

The advantage of using human-mediated introductions to investigate the effects of bottlenecks is that they often occurred on known dates and involved a founding population of a known size. On the Hawaiian archipelago, the Mauna Kea silversword (Argyroxiphium sandwicense ssp sandwicense) has undergone at least two bottlenecks in recent times. The first known bottleneck occurred approximately 100-150 years ago when the spread of introduced ungulates caused the population to plummet to fewer than 50 plants. Microsatellite data showed that this bottleneck did not significantly deplete genetic diversity according to either allelic diversity (A) or heterozygosity (He) values (Friar et al. 2002). This retention of genetic diversity in a population of <50 individuals is partly attributable to the species' reproductive ecology; silverswords often grow for 30--50 years before flowering and reproducing, and therefore recovery was rapid because the bottleneck lasted for only one or two generations

The second known Mauna Kea silversword bottleneck occurred in a reintro-duced population that originated from two or three transplanted individuals in 1973. Today this population numbers more than 1500 individuals, but the 1973 bottleneck was so extreme that its effects on genetic diversity could not be ameliorated by rapid population growth and long generation times, and levels of A and He remain very low in the current population (Friar et al., 2002). This example illustrates how a population's rapid recovery can ameliorate a moderately severe, but not an extremely severe, bottleneck.

The effects that founding population size and post-bottleneck growth rates had on genetic variation were also investigated in a review of several studies of avian introductions onto islands that used allozyme data to compare the genetic diversity of mainland (source) and island populations (Baker, 1992). These data showed that introduced island populations do not necessarily have lower levels of genetic diversity than the source mainland populations. The genetic diversity of island populations was correlated with the size of the founding population and also with the rate at which the population grew (Figure 3.10), a finding that supports the idea that bottlenecks may reduce overall genetic diversity only if severe and prolonged.

0 100 200 300 400 500 600 700 Size of founding population

0.04

0.02

0.00

0.02

0.00

0 10 20 30 40 50 60 Time to abundance

Figure 3.10 The effects of (a) size of founding population and (b) time to abundance on the He of founded (island) populations compared with their source (mainland) populations. The values of He were calculated from allozyme data representing populations of starlings (Sturnus vulgaris), greenfinches (Carduelis chloris) and chaffinches (Fringilla coelebs) that were introduced to New Zealand, and common mynas (Acridotheres tristis) and house sparrows (Passer domesticus) that were introduced to both New Zealand and Australia. Time to abundance refers to the number of years between the founding event and the time when populations were first described as abundant. Dashed lines represent parity between He of the source and founded populations. Data points above dashed lines represent founded populations with lower He than source populations; data points below dashed lines represent founded populations with higher He than source populations. Generally, the He of the founded population is positively correlated with the number of founders and the population growth rate. Data from Merila, Bjorklund and Baker (1996) and references therein

Was this article helpful?

0 0

Responses

  • GUY
    What are bottleneck in population ecology?
    1 month ago

Post a comment