Mutations occur at measurable rates. These rates vary depending on the type of mutation, the location of the mutation within a genome, from species to species, and also within species as a function of genotype as well as environment. Mutations can be spontaneous or instead caused by mutagenic processes. As mutations are the ultimate source of genetic variation within populations, mutation rates can play important roles in both the amount of genetic variation available to a population and otherwise rates of evolution, and particularly so when genetic migration is not prevalent and, indeed, given a lack of sex generally.
Mutations not only can be qualitatively distinguished, i.e., in terms of different types, but also quantitatively distinguished, such as in terms of their frequency. Organisms will tend to have characteristic mutation rates, and different organisms have different characteristic mutation rates. Especially, "Mutation rates in organisms reflect the need for accuracy to maintain critical genetic information and the requirement for flexibility to adapt to environmental changes," (Loh et al., 2010) (emphasis mine). Here greater accuracy represents lower mutation rates, a.k.a., higher fidelity in genome replication. Flexibility refers to genetic variation, that is, the presence of mutations, a.k.a., allelic variation within populations, which in turn is ultimately the result of infidelity in genome replication.
Certain strains of microorganisms have been described as general mutators, which is due to their display of unusually high mutation rates. In considering mutation rates, it is important to realize that over long periods, e.g., multiple generations/evolutionary time, it is not possible to determine actual mutation rates. This is because other evolutionary mechanisms, such as drift and selection, will have an impact on the retention of mutations by lineages. Indeed, such complications are a concern when measuring mutation rates even over short periods. Another way of stating this is that over long time frames mutation rate determinations are less a function of replication fidelities and more of mutation fixation rates, a subject covered in the subsequent section.
Some authors have come to describe organisms displaying very high mutation rates, e.g., RNA viruses, as existing as quasispecies. This descriptor may be interpreted as these organisms simply displaying an unusual high potential to explore sequence space. That is, like quasispecies, sensu stricto, populations experiencing very high mutation rates generate large numbers of mutations, such that every nucleotide has a potential to display every possible transition or transversion. Unlike non-biological quasispecies, however, the fitness of each of the resulting mutants will not be identical and in fact natural selection will tend to limit the exploration of sequence space no matter how high mutation frequencies might be.
There are a number of ways that mutation frequency may be described, that is, the number of mutations per unit something. These include as a function of time, per round of replication, per nucleotide, and per genome. Per unit time typically is what is determined when organisms are exposed to exogenous mutagens, i.e., chemicals or physical agents that are found in environments that can modify the sequence of nucleotides in hereditary material. Mutation rate in this case then would be the number of mutations present within the population following some duration of exposure, calculated as divided by that duration (e.g., 10 mutations following 10 min of exposure would be a mutation rate of 1 mutation per min). This is equivalent to mutation accumulation within a mutating environment. Alternatively, intrinsic mutation rates, ones not caused by exogenous mutagens, are typically determined per round of replication and the resulting per-nucleotide mutation rates can be viewed as measures of replication fidelity—the more mutations generated per round of replication, per nucleotide, then the more mistake-prone the polymerase (or, alternative, the greater the exposure to exogenous or, indeed, endogenous mutagens). Yet another means of considering mutation rates is the per-gene rate, which is similar to the per-nucleotide rate, except that the target size for the mutation is larger since genes consist of multiple nucleotides. Per-gene rates can be important when considering the likelihood of occurrence of mutations to specific phenotypes, such as one sees with host range shifts for symbiotic organisms. Mutation rates also can differ across genomes, with certain locations or certain nucleotides representing mutational hot spots, that is, where rates of mutation are higher than average. One additionally can speak of rates of accumulation of mutations across entire populations, which can limit the rate at which populations can evolve such as in the face of changing environments.
Of greatest importance evolutionarily is the per-genome mutation rate. This is because most mutations are detrimental and too-high per-genome mutation rates mean that lineages may accumulate mutations faster than they can eliminate them via natural selection. Sufficiently high mutation rates, such as may be induced via exposure to exogenous mutagens, thus can result in population extinction. This phenomenon is known as mutational meltdown, lethal mutagenesis, or error catastrophe. To avoid error catastrophe, organisms have to devote a certain fraction of their genomes to achieving reasonable levels of per-nucleotide replicative fidelity. If these fidelity-enhancing functions are more or less fixed in terms of the size and number of genes involved, then they may be less available to organisms with very small genomes versus those with larger genomes. Organisms with smaller genomes nonetheless will automatically display lower per-genome mutation rates than will equivalent larger-genomed organisms if both display the same per-nucleotide replication fidelity. Nonetheless, these smaller genomed organisms, mainly small viruses, can be sufficiently lacking in functions enhancing replication fidelity that they seem to perpetually exist near to error catastrophe.
The mutation rate that an organism displays is a product of natural selection meaning that it can change mutationally, either increasing or decreasing, plus some of these mutations will be associated with greater organismal fitness than will others. Specifically, optimal mutation rates should represent a balance between the metabolic or kinetic costs of increasing fidelity (i.e., energy or speed costs), the genetic costs of too-high per-genome mutation rates, and the evolutionary cost of too-low per-gene mutation rates. Too-low mutation rates, however, are costly mostly as a function of environmental stability: More-stable environments are less dependent on high rates of generation of novel, beneficial mutations for the sake of population survival and prosperity than are populations exposed to environments which are changing in unanticipated ways. We can thus expect selection to favor lower mutation rates in more stable environments and higher mutation rates given environmental change.
A given type of organism and/or a given species will have a characteristic mutation rate. This rate will be a function of the fidelity of the nucleic acid polymerases involved in the replication of its genome as well as the mechanisms of nucleic-acid repair available to the organism. Since polymerases and repair enzymes are gene products, it is possible for mutations to impact these functions, resulting in greater mutation rates as well as a potential for mutations in the same genes to give rise instead to mutation rate reductions: More mistakes and fewer mistakes, respectively. Lineages that possess mutations that lower their replication fidelity are described as general mutator strains. Such strains may be generated intentionally in the laboratory, but surprisingly are also relatively common in natural populations.
The argument for why general mutator strains would be present in wild populations involves the concept of genetic linkage – which is high among the genes found in many microorganisms due to low frequencies of recombination that in turn are a consequence of these organisms (typically) not being obligately sexual (Levin and Bergstrom, 2000) – and the utility of higher mutation rates given environmental change. In particular, if the generation of new beneficial mutations is substantially limiting to evolutionary adaptation, then an allele that bestows higher mutation rates may display a greater potential to be linked with a so-generated beneficial mutation. A mutator allele as a result of this linkage thus may achieve greater evolutionary success than would an allele that does not confer similarly higher mutation rates. In this case it is the potential to generate beneficial mutations that is selected for or, alternatively, the potential to generate detrimental alleles that is selected against, with selection strongest particularly when so-generated alleles are retained, via linkage, within the same genome over long periods.
Since the frequency of beneficial mutations also is a function of population size (i.e., the more individuals that are mutating then the higher the likelihood of the occurrence of a beneficial mutation), a mutator strain may be viewed as being able to generate a given beneficial mutation at lower population sizes than a non-mutator strain. Thus, in a sufficiently fluctuating environment, mutator strains may display both greater competitive ability (relative fitness) and greater overall adaptability, especially if population sizes are small. Mutators also have a greater potential to acquire simultaneous (or nearly simultaneous) beneficial mutations. This could be especially helpful towards the generation of multiple mutations that together are beneficial but individually are neutral or costly, and which otherwise would have a greater likelihood of occurring only in larger populations (given overall probabilities equal to per-individual probabilities multiplied by numbers of individuals). Yet another way of viewing the potential for general mutator strains to possess a selective advantage is that these strains will tend to accumulate neutral variation at a faster rate and that neutral variation can have some potential to be useful given environmental changes.
The cost of being a general mutator is one of higher rates of accumulation of deleterious mutations. Indeed, if there is a likelihood of simultaneous multiple mutations, then it presumably is more likely that a beneficial mutation will be paired with a detrimental mutation than that a mutually beneficial pair will be generated. In addition, and also an obvious cost to the extent that higher mutation rates push an organism toward error catastrophe, not only will detrimental mutations tend to accumulate within lineages, particularly mildly detrimental mutations, but so too will those detrimental mutations tend to result in declines specifically in organism fecundity. The latter occurs as a significant fraction of progeny show reduced levels of reproduction that are a consequence of acquired detrimental mutations.
In terms of the occurrence of specific mutations in specific populations, mutation rates are not everything. In fact, the likelihood of acquiring a specific mutation is also proportional to the rate of mutation to that specific change. This rate, while dependent on general mutation rates, is also dependent on a combination of the location within a genome of the mutational change (e.g., hot spots, as well as cold spots), the evolutionary history of the organism (which defines what is present that must be changed), and the specific mutation sought. That is, organisms are constrained not only by general and per-nucleotide mutational tendencies – keeping in mind that mutations can involve deletions, insertions, inversions, and translocations, as well as individual nucleotide changes – but also in terms of the distance different starting genotypes are from a specific ending genotypes. This is another way of viewing the concept of evolution being limited by rates of mutation, as natural selection can only act on what variation is currently present within a population, and what variation is present is dependent not only on mutation rates and how long as well as effectively new alleles have been accumulating, but is dependent also on what mutational change, in terms of resulting genotype, is reasonably probable. The old adage of "You can't get there from here" is a reasonable approximation of the problem of historical contingency on the potential for populations to evolve.
An additional and crucial issue is the number of individuals present within the population. Importantly, it should be kept in mind that this number is a function not just of organism density but also of environmental volume: A population displaying the same density but housed within a one-hundred-fold greater culture volume will, on average, display a one-hundred-fold greater likelihood of displaying a particular mutation. Thus, with mutations, it not only is organism properties and the specific mutations sought that are important, but so also is the absolute number of organisms mutating.
Lastly, the potential to go from one genotype to another often is dependent on the fitness of intermediate genotypes. Thus, unless multiple mutations occur simultaneously, or nearly so, then mutation accumulation will take time, resulting effectively in reduced rates of overall mutation from one genotype to another, owing to a potentially reduced likelihood for intermediate forms to successfully reproduce. Whether due to the number of mutations necessary to attain a given genotype from a particular starting genotype, or a function of the evolutionary fitness of intermediate steps, delays in mutation from one genotype to another can be described also as a consequence of historical contingencies (Blount et al., 2008) .