BIOL 4160

Evolution

Phil Ganter

301 Harned Hall

963-5782

07 - Evolution of the Phenotype

Email me

Link to a list of Specific Objectives for lectures

Back to:

Academic Page 
Tennessee State Home page 
BIOL4160 Page 
Ganter home page 

Phenotypic Evolution

  • The previous lecture presented a model of natural selection that was genetically explicit but only looked at a single locus or two linked loci
    • Most phenotypic traits (and all complex phenotypic traits) are polygenic (affected by many loci)
    • Here we will look at understanding evolution from the perspective of the phenotype without knowing what the exact genetic system that produces it
  • Phenotypic traits are often Quantitative traits in that they are described by measurement of some continuous variable, such as weight or length or concentration or by counting something like bristle number on Drosophila
  • Darwin examined many cases of artificial selection and concluded that there was sufficient genetic variation in natural populations to allow selection for many complex traits that would modify the phenotype so much as to take it outside of the observed range in phenotypes found in the natural population
    • This must mean that, assuming no mutation (reasonable in the time span of the average artificial selection program), that these striking phenotype changes are mostly the result of recombination of already existing genes
  • Phenotypic Variation
    • Geographical Variation is an important component of within-species variation
    • Local populations may have adapted to specific local conditions not found in locales occupied by other populations of a single species

A Model of Phenotype and Phenotypic Variation

  • Each trait that can be measured or counted has a mean value in a population
  • With respect to a single locus that affects whatever trait under consideration, each homozygote has a separate mean
    • If the mean of the heterozygotes is at the midpoint between the two homozygote means, then the effect of each allele at that locus is Additive
      • Let a = the distance between any homozygote mean and the midpoint (so a is half of the interval between the two homozygote means)
    • As the value of the heterozygote departs from the midpoint, the effect of the alleles becomes less and less linear (which means it is less and less additive)
      • Dominance is a departure from additivity
      • For some very non-linear traits, the homozygote lies outside the interval between the two homozygote means
  • A model of additive effects (we will not deal with the more complicated case of non-additive (non-linear) effects here
    • Additive effects allows us to make some predictions about the relationship between parents and offspring
    • The expected average for a trait among all of the offspring from two parents is the average (= midpoint because there are two parents to average) of the parent's trait values
      • Natural Selection will change this expectation (the offspring average will be closer to the favored allele's homozygote average)
        • The difference between the midpoint of the parents and the offspring average is the Response to Selection
        • The offspring should be at the midpoint and selection is the reason for the difference
      • If more than one locus affects the value of a trait in an additive fashion, then the average value for that trait for a particular multi-locus genotype is the sum of the additive values of each locus of the genotype
    • Recall (from lecture 03) that the variation in a trait within a population or species can be divided into two additive portions:

Phenotypic Variation (Vp) =Genetic Variation (VG) + Environmental Variation(VE)

  • Variation is the average of the squared differences between individual values and a mean value (the numerator of the equation is called a "sum of squares corrected for the mean" or more simply and usual the "sum of squares")

    • If you have learned this formula in a stat class, you might be expecting N-1 in the denominator, but that is not needed here as we are describing the population and not a sample
  • Remember, variance is a measure of the differences between all of the individuals in a population and their collective mean - as they cluster closer to the mean, variance shrinks (if they were all the same as the mean, then there is no variation and variance is zero because the sum of squares becomes zero)
  • The additive portion of genetic variation due to a single locus (with two alleles) at the level of a population (VA) depends on the effect of the alleles (a) and the frequency of the alleles (p for A1 and q for A2) as follows:

VA = 2pqa2

      • Thus, we have a simple way to calculate VA
      • As in the case of the expected trait values above, when more than one locus affects the value of a trait in an additive fashion, then VA is the sum of the variance of the trait at each locus
    • Narrow-Sense Heritability -- The partition of additive variance into environmental and genetic components allows the estimation of heritability (in the narrow, or additive, sense) as

h2N = VA  / (VG + VE)

      • Narrow sense heritability can also be estimated by plotting the offspring average versus the midpoint of the parents
      • The slope of the regression line is equal to the narrow-sense heritability of that trait

Estimating the number of loci affecting a trait

  • Current methods allow the detection of loci for which there is variation present
    • Done by associating variation in the phenotype inherited with previously identified and mapped marker loci, one can separate the effects of each variable locus that affects the trait
  • When the marker loci are recombined, so to will be the loci closely linked to the marker
    • By observing individuals that are recombinant with respect to the marker loci, one can infer that the additive phenotypic changes are due to recombination of loci linked to the markers
  • Markers used to be rare mutations that affected the phenotype strongly but now they can be small base-pair sequences for which their position along the gene may be determined
    • This produces a map of the genome and each variable locus detected is detected because it is linked to a different place on the map
    • Accuracy depends on how finely the genome is mapped
      • Loci not closely linked to markers will be missed
      • When more than one locus is linked to a marker, only one will be counted
  • This technique (which is both experimentally and computationally intense) is used to identify the number and location of Quantitative Trait Loci (QTL) for some important phenotypic trait
    • often used by agronomists seeking to improve complex traits of crops (like fruit size or flavor)

Neutral Loci

  • In the book, there is a long discussion of neutrality and phenotypic evolution, which may strike you as odd if you consider neutral mutations to be mutations that do not affect the phenotype but this is not completely accurate
    • neutral mutations are those that do not affect the fitness of the organism in which they occur
    • mutations that do not change the phenotype in any way must be neutral
    • mutations that do change the phenotype in some measurable way but do not affect the organism's fitness are also neutral
    • Also, many researchers in this area have worked with "nearly neutral" mutations, one that have selection coefficients so small and population sizes so small that random chance is a greater influence on their fates  (fixation or elimination) than is selection
  • Genetic drift and mutation will cause trait averages to change slightly from generation to generation in a random manner
    • This means that, each generation change can be in either direction, so that the mean seems to wander
      • The rate of wandering is directly linked to the mutation rate
    • Here, selection can be detected as a departure from the expected rate of neutral change
      • Greater rates of change indicate directional or diversifying selection
      • Lower rates indicate stabilizing selection
    • Studies that have looked at changes in trait values and compared them to the expected rate of change due to neutral mutations and genetic drift have found many more cases of reduced rates of change than of higher than or equal to rates of change
      • Conclusion: stabilizing selection is very common, neutral drift and other modes of selection are less common

Correlated Evolution

  • Change in one phenotypic trait may be linked to change in another trait - this is Correlated Evolution of Phenotypic Traits
  • There are two generally recognized reasons for this correlated response:

Correlated Selection

  • Correlated Selection comes from a functional connection between the two traits
    • If the nature of one trait determines the fitness of the other trait, then their correlation is due to correlated selection
  • Consider a species of beetle that is polymorphic with respect to size and coloration of its elytra (the hard coverings over its wings) and is subject to intense predation
    • The two most successful strategies for avoiding predation are to avoid detection or to be able to startle the predator by rapidly moving the elytra
      • Large beetles with bright elytra are most effective at startling the predator
      • Small beetles with dull elytra are hardest for the predator to detect
      • Small beetles with bright elytra or large beetles with dull elytra are not particularly good at either strategy
    • In this scenario, selection for increased size (for some reason other than predation) will also increase the fitness of bright elytra and the correlation is due to correlated selection

Genetic Correlation

When organisms are observed, one can often measure correlation between two phenotypic traits

  • These Phenotypic Correlations may be due to either
    • Environmental Correlation - the traits are correlated because of some similarity in their environments (both the growth rate and viability of infants will improve as the quality of their diet improves)
    • Genetic Correlation - the traits are correlated due to some aspect of their genotype
    • Linkage disequilibrium - alleles at different loci occur together more often than expected and, when one allele is favored as the phenotype it affects is selected, then the other allele will also increase and so change the value of the phenotype it affects
    • Pleiotropic Effects - one allele affects more than one trait
      • Artificial selection has demonstrated that pleiotropy is common
        • Often occurs when a single locus is selected by crop scientists - the trait of interest is changed as expected but there are other effects (often negative) on other important traits
        • As fruits get larger, trees grow more slowly or are less disease resistant or produce less pollen, etc.
      • Modifier Alleles - these are alleles at loci other than the locus under selection that modify the selected locus' pleiotropic effects
        • The action of modifier alleles can be seen when selection for one trait causes negative effects (on fitness) of other traits
          • Note - the total effect on fitness is the sum of the positive effect due to selection of the first trait minus the negative effects on the other traits and the sum must be positive for selection to drive the changes
        • Any changes due to selection at other loci that affect those traits negatively affected and that ameliorate the negative effects are due to changes at modifier loci
  • Genetic correlation, if strong, may result in neither trait reaching its optimal phenotype if there is a trade-off when increasing fitness
    • Book's example of egg size versus egg number has been demonstrated for many invertebrates
  • Genetic correlation may also speed the rate of adaptation if, through linkage or pleiotropy, when adaptation of one trait enhances the rate of adaptation at a correlated trait

Norms of Reaction

  • Phenotypic Plasticity is the ability of a single genotype to produce multiple phenotypes
    • A single genotype may produce different phenotypes that are linked to changes in the environment
  • The set of phenotypes that a single genotype can produce is it's  Norm of Reaction
    • Different genotypes can have different reaction norms
      • When the differences between genotypes are additive (like the differences between parallel lines are additive because the lines only differ in the value of their intercepts), then there is no interaction between genotype and environment
      • When genotypes change in non-additive ways, then the interaction is important, so reaction norm plots are ways of demonstrating the nature of G x E interactions
  • Development and Reaction Norms
    • The phenotype of an organism may be set by the environment in which it developed
        • Multivoltine insects (those with multiple generations per year) in temperate climates may have different developmental pathways for producing spring/summer phenotypes versus fall/winter phenotypes
      • Developmental Switches are molecular mechanisms that are sensitive to environmental conditions and direct the developmental pathway toward one of the possible phenotypes that constitute the reaction norm of a particular genotype
    • Canalization
      • Many traits are not altered much by environmental effects and some are little changed by genotype
      • The general term for the processes that influence development such that the phenotype is not altered by environment or genotype is Canalization
        • This idea is tied to genetic correlation in that sets of alleles are selected for if they produce the optimal phenotype
        • Phenotypic Integration is the idea that functionally connected phenotypes should be genetically correlated so that canalization is achieved
Last updated February 26, 2009