Back: Table of Contents | Why is genetic diversity important?
2. How is genetic diversity distributed in natural populations?
There are genetic differences among individuals within most (but not all) populations of plants and animals. There are also differences among populations across the range of each species. In this section we review some basic patterns of how genetic diversity of species is distributed, or partitioned.
2a. Monomorphic and polymorphic alleles.
By definition, individuals within a given species share some percentage of alleles; otherwise, they would not be considered members of the same species. This shared (or common) portion of the gene pool includes two basic classes of genes. Genes that are monomorphic (i.e., A = 1, where A is the number of alleles per locus) within a species are common to all populations, and indeed essentially to all individuals. (Technically, "monomorphic" genes are defined as those for which the frequency of the most common allele is > 99%; thus, some variation may be present even at these loci). The monomorphic proportion of the total genome varies among taxonomic groups, typically in the range of 85% in animals and 50% in plants (Hedrick 1985; Hartl and Clark 1997).
The remainder of loci (on average, 15% in animals and 50% in plants) is polymorphic (A > 1), varying among individuals in a population, and among populations within a species. These loci are of most concern to restoration, since polymorphic loci may or may not be adequately represented depending on the design and execution of a restoration program.
2b. Diversity within and among populations of a single species.
The genetic profile of whole populations typically varies from place to place across a species range. These differences may arise as the result of chance occurrences, such as the genetic composition of dispersing individuals that create a new population (founder effect), or changes in allele frequencies that result from chance matings in very small populations (genetic drift) (Primack and Kang 1989; Templeton 1991; Meffe and Carroll 1994; Eckert, Manicacci, and Barrett 1996; Husband and Schemske 1996). Differences among populations can also arise systematically, especially if the environment in various places exposes individuals to different optima for survival and reproduction (fitness). For these and other reasons, populations often diverge from one another in their genetic composition. Such divergence is especially strong and rapid when there is little gene flow between populations (e.g., limited dispersal of seeds or pollen, or limited movement of animals across physiographic barriers). Over evolutionary time, such among-population genetic differences can accumulate and eventually result in the development of a new species (allopatric speciation). Indeed, "populations" are defined as much (or more) by patterns of mating and gene flow as by the physical distribution of individuals, although the two are often closely related (Hartl and Clark 1997).
Each species distributes its genetic diversity (one measure of which is the total of all alleles at all loci) in a pattern reflecting both its biology and its history (Wright 1965; Nei 1975). For example, nearby populations of plants that are pollinated by bees may share many alleles because genes (packaged in pollen grains) can flow easily between sites. Such species may have fewer unique alleles in each population, so populations tend to be genetically similar. By contrast, there may be less gene flow among populations of species that are pollinated by ground-dwelling flightless beetles, or whose heavy fruits fall to the ground in the vicinity of the parent tree. Gene flow can also be obstructed by physical barriers (i.e., topography or habitat that a pollinator, disperser, or migrating individual cannot cross), as well as by disturbance (Levin 1981; Slatkin 1987).
Even if variation within a population is low, there may be considerable variability among populations. Imagine Species X, a plant that lives in high mountaintop alpine areas, in which every individual on a given mountain is nearly identical genetically to others. In this case, individuals in Population 1 will all be nearly identical to each other; individuals in Population 2 will also be identical to each other, but because populations are very isolated from one another, they may all be very different from all the individuals in Population 1; and so on. Under these circumstances, most of the variation in Species A is among populations; within-population variation is low or nonexistent.
Now imagine a contrary example. Species B lives in tallgrass prairie, where suitable habitats are relatively close to one another (or even continuous), and dispersal among populations is common. Here we might find a great deal of variation among individuals in each population (reflecting the benefits of variation discussed above), but because gene flow is high among sites, most populations are similar (that is, most polymorphic alleles are widely distributed). This illustrates a species with a high proportion of variation distributed within populations, while among-population variation is relatively small.
A variety of measures are used to quantify the distribution of genetic variation among individuals within populations, and among populations. These measures are the basis for describing how genetic variation is partitioned within species. Differences among populations are commonly quantified by the use of one of several statistics, including Wright's inbreeding coefficient (FST) and Nei's coefficient of gene variation (GST). These indices are functions of how heterozygosity is partitioned within and among populations, based on differences in allele frequencies (Wright 1969; Nei 1975; Chai 1976; Wright 1978). Where pi is the frequency of the ith allele, heterozygosity (H) = 1 - å pi2 for {i = 1...n} for populations within a larger sample. Then, the proportion of total variation that is distributed among populations (GST) is , or equivalently, 1- (HS / HT), where HS and HT are the mean heterozygosity within populations and in the entire species respectively. Values for these statistics can often (although not often enough!) be found in the population genetic literature, so restorationists do not have to generate them. Values of FST and GST vary from 0 to 1 (Nei 1975; Hedrick 1985; Crow 1986; Hartl and Clark 1997). GST in particular has a number of useful properties: it can be used for one or many loci, mutation rates do not alter the statistic significantly, the exact number of populations need not be specified, and the statistic is relatively responsive to changes in allele frequencies in time. Although they have important conceptual differences, in practice FST and GST are used in similar fashion as indices of genetic difference among populations (Crow 1986).
Species genetic structure can be modeled in a variety of ways. The simplest is a "finite island" conceptual model, in which alleles can flow between any pair of populations. In this case, the key parameters are the migration rate and effective population size (Crow 1986). However, the assumption that any two populations are equally likely to exchange alleles is probably unrealistic. Instead, a "stepping stone" model can be used, in which alleles spread by "stepping" from one population to an adjacent one (Wright 1969; Crow 1986).
In practical terms, G and F statistics tell us whether the majority of genetic variation is distributed among or within populations. In species with low GST (approaching 0), the majority of variation is found within populations; individuals within populations are likely to be genetically different, but each population contains the same complement of alleles in similar frequencies Where GST is high (approaching 1), individuals within a population are relatively similar but populations are significantly different. Most species fall somewhere in between these extremes.
The distribution of genetic variation within and among species is strongly linked to life-history traits, particularly dispersal and reproductive mode (Hamrick and Godt 1990; Hamrick et al. 1991). Species that disperse genes (in plants, this includes both pollen and seeds) widely and frequently will tend, other things being equal, to have lower GST (i.e., populations will be more similar). Even a moderate rate of gene movement among populations (one individual every a few generations) can "link" the gene pools of two populations. Mutation and drift (chance selection of genotypes, especially likely in small populations) can also lead to changes in allele frequencies, although in general these forces are believed to act more slowly than dispersal and selection.
2c. How is genetic variation detected and measured?
A variety of methods exist for the assessment of genetic variation (Schaal, Leverich, and Rogstad 1991). Traditionally, genetic variation was inferred by the shape (morphology) or growth responses of organisms. Such approaches often use common garden experiments or reciprocal transplants to distinguish genotypic and phenotypic variation. In this approach, individuals that appear phenotypically different (or that grow in different environments) are placed in a common environment. This eliminates the environmental component of variation; presumably, the remaining variation has a genetic basis.
While these methods can be useful, they measure higher-order effects of genetic variation (i.e., expression in the whole organism), and hence address genetic variation only indirectly. Moreover, the confounding effects of phenotypic variation are often difficult to separate from underlying genetic differences among organisms. In practical terms, it is often not possible to know whether organisms that look different are actually different genetically, or whether their differing phenotypes reflect the influences of environment or chance in development.
Since the 1960's the most widely applied method of estimating genetic differences among individuals is enzyme electrophoresis (also referred to as starch gel electrophoresis, including both allozyme and isozyme analysis). Electrophoresis detects variation in amino acids that are early products of gene translation (the process by which a cell translates a copy of its genetic code in messenger-RNA into amino acids). Since there may be multiple biochemical pathways leading to the synthesis of a single amino acid or enzyme, electrophoresis is relatively conservative in its estimates of genetic variation among individuals (Hamrick, Linhart, and Mitton 1979). Enzyme electrophoresis remains the most widely used method of estimating genetic variation, and literally thousands of studies have been conducted on a wide range of organisms. Electrophoretic analysis permits individuals to be distinguished from one another based on enzyme variation. These differences among individuals can be compiled to generate statistics about the degree of variation within and among populations (§2.b), a key consideration in restoration ecology.
In the past decade, powerful methods have developed and emerged to detect variation in DNA itself, not simply its products (Karp, Isaac, and Ingram 1998). While a bewildering variety of tools and techniques may be encountered in the literature, they all have in common the ability to reconstruct the sequence of nucleotides in the DNA molecule, arguably the most basic level of genetic variation (Britten 1986; Clegg 1990). The conceptually simplest (although experimentally most laborious) technique is gene sequencing -- that is, the exact reconstruction of the complete nucleotide sequence. Part of the difficulty in interpreting DNA sequences is that "genes" in a functional sense may be distributed over several sections of a chromosome; there are also large sections of all chromosomes that do not code for any known function, or that are redundant or regulatory (influence the transcription, translation, or function of other genes). Thus, unlike enzyme analysis, much of the information in a DNA sequence may be of little relevance to understanding ecologically significant genetic variation. Since in theory every sexually-produced individual is likely to be unique, DNA sequences can be almost too variable to detect the kinds of patterns required for restoration work. However, new analytical techniques allow DNA sequence variation to be interpreted in meaningful ways. Complete gene sequences have been complied for relatively few organisms, and generally only a few individuals each, although these numbers are increasing all the time.
It is important to note that isozymes and genetic markers may not reflect traits under strong natural selection. For example, Knapp and Rice (1998) evaluated patterns of variation in a native grass using both quantitative traits and isozymes, and compared these to geographic distance and climate. They found overall that isozymes did not reflect an adaptation to climate, whereas the quantitative traits did. In a restoration context, this suggests that climate may in some cases serve as a more useful guide for collecting and reintroduction zones than genetic markers per se.