Adaptation and

Units of Selection

Lecture 05

Email me

Assignments Page

Back to:

Natural Selection explains Adaptation

• Is adaptation gradual?
• How good is adaptation?
• Genetic Constraints on fitness
• Developmental Asymmetry
  • Morphospace
  • Allometry
  • Historical Constraints on Fitness
• Functional Constraints on Fitness
• Adaptation as Design and as Fitness

The Units of Selection

• Genic Selection
• Cellular Selection
• Organismic (Individual) Selection
• Group Selection
• Kin Selection

Natural Selection explains Adaptation

• Ridley argues that directed change has no evidentiary basis but that evidence does exist for natural selection as the cause of adaptation
  • orthogenesis is ruled out
  • no other mechanistic explanation
• Not all evolution is driven by natural selection so not all evolution need result in adaptation
• Lewontin and Gould point out that a common mistake is to view all features of an organism as adaptations
  • drift, hitch-hiking, developmental constraints, chance, complex fitness landscapes are all possible reasons for a feature not being optimized

Is adaptation gradual?

• Rates of molecular evolution vary over time and within a single gene
  • Problem here is that gradual is not an exact term
  • Rate variation can come as new functions are added or as new interactions between existing structures are found
• Fisher's model of an adaptive landscape with one hump leads to the conclusion that large change is more likely to be maladaptive than a small one
  • depends on simple, unchanging landscape
• Gradual evolution must overcome the criticism of no value for intermediate stages
  • example of eye has had most attention
• Preadaptation is also possible
  • a structure serving one function is co-opted for a second function (for which it is said to be pre-adapted)
  • wings and feathers (especially feathers) are supposed to be pre-adaptations for flight
  • exon shuffling to make new enzymes

How good is adaptation?

• Often, adaptation and perfect are in the same sentence but is that realistic?
  • The environment changes but the evolutionary response can be delayed
  • produces a time-lag in evolution
  • will be larger where variation is lacking in population and mutation is the rate-limiting step
• time lags can lead to extinction

Genetic Constraints on fitness

• Heterosis leads to continual production of less fit homozygotes
• Developmental Constraints
  • Pleiotropy - when a gene has effects on more than one phenotypic character
  • if the most fit version of the gene for one character is much less fit for the other, it may not be favored.
  • Developmental Asymmetry
    • a measure of developmental disruption
    • high for new mutations advantageous at one locus but somewhat less fit at secondary loci
    • if new mutation becomes common, DA often drops as alleles at other loci that affect the same secondary traits are selected to reduce the disruption and produce the best overall phenotype
    • secondary alleles are often called Modifier Genes and the process of returning to the best developmental pathway is called Canalization
    • Canalizing Selection is selection for those modifier alleles which are best coordinated with the advatageous allele at the primary locus of interest
  • Morphospace
    • the region of possible morphologies
    • realized morphospace is the region of morphospace represented by actual organisms
    • the remaining morphospace is unoccupied
      • may be due to selective constraint
      • may be due to other constraint (genetic, developmental, mutational, etc.)
• Allometry
  • correlated rates of change between different characters
  • by altering these relationships, you can produce great phenotypic change with very little genetic change
  • look at Homo versus Pan, and consider the small genetic distances involved
  • selection of stalk-eyed flies shows very rapid changes to phenotype but little genetic change

Historical Constraints on Fitness

• This is linked to developmental constraint
• when change takes place, intermediate forms may alter the character in a way that is less fit in the current phenotype than would be if the phenotype had been optimized for current function
• if there are several different, but equally  fit phenotypes possible, then different populations may differ only because of historical accident and not because the selective topology leads to different peaks in different places as is often assumed

Functional Constraints on Fitness

• some changes may be linked to other aspects of the organism's function (e.g. larger bones are stronger and will resist breakage but also weigh much more and will cost more resource to carry around and maintain)
• This is an example of a Phenotypic Trade-Off and trade-offs may constrain evolution of a trait

Adaptation as Design and as Fitness

• Adaptations are often discussed as advances in design
• position of eyes in raptors versus waterfowl
• raptors eyes face forward and allow two eyes to focus on all objects of interest (often called binocular vision)
• good for spatial perception and raptors can snatch a prey item while at full speed flight
• ducks do not focus both eyes on food items but they do have much wider fields of vision
• they are more often prey than raptors and the wider visual field gives them advantages over raptors in detecting approaching predators
• This is an example of adaptation as design as the advantage is a direct result of changes in design
• Fitness is still the coin but it would be very hard to test the fitness differences as all raptors have binocular vision throughout their visual field and waterfowl do not, so how to compare as many other differences also exist between the two groups
• Some of the "Adaptationist Agenda" from the Spandrels article results from assumption that designs are always perfect and that differences must therefore be fitter and so the design is better, which is circular

The Units of Selection

What is favored by selection?

• The gene?
• The organism?
• The population?
• The species?

All have been argued to be the beneficiaries of natural selection but can they all be?

• What happen when a benefit to one level is a cost at another (when selection at different levels is opposed)?
• Answering this question has provided evolution with one of its most long-standing controversies

Genic Selection

• selection that increases the number of copies of a gene in succeeding generations, no matter what the effect on other levels
• Example of genic selection is the phenomenon of meiotic drive genes (segregation distorter genes)
  • t-allele of mice (simplified)
  • t-allele (actually more than one closely linked loci), when homozygous, produces sterile males
  • when heterozygous in male mice it produces a short-tailed mouse that have the t-allele in 90% of their sperm (not the expected 50% so this is the meiotic drive or segregation distortion)
• genic level selection favors the distorter gene but organismic selection favors the wild type allele (haplotype) and it is also not favored at the group level as, in small populations, if all of the males become t-allele homozygous there is reproductive failure (the groups have higher than average probability of extinction).
  • The actual frequency of the t-allele is a balance between the three levels.
• meiotic drive alleles have been discovered in lots of insects, some fungi, and some plants so it is a widely spread phenomenon
• the logic behind a 1 to 1 sex ratio being the stable strategy involves selection for genes that over-produce the rare sex

Cellular Selection

• This is the outcome of different cell lines reproducing independently
• many animals have a distinction between the germ line and the somatic cells
  • only mutations to the germ line cells have any chance of being inherited in this system
• however, most other organisms have no such distinction and a mitotic mutation may lead to a part of an individual that gives rise to some reproductive tissue, which will carry the mutation, although other reproductive tissues on the same individual will not
  • thus, cell lines can be selected independently of the individuals within which they reside

Organismic (Individual) Selection is what we have been studying up until now

• This can get complicated when you consider what an individual organism is
• What is an organism when parts can become separated and reproduce independently (plants connected by rhizomes can be separated and each part is a new plant)
  • some colonial animals are similar (cnidarians are a good example)

Group Selection

• When group and individual and genic selection all promote the same alleles, joy reigns in mudville
  • When group and individual selection are opposed, many believe group selection will always yield to individual selection
  • the trait which harms the individual but helps the group is the Altruistic Trait

Altruism can lead groups with individuals that have it to found new groups faster than groups lacking individuals with the trait

• This is group selection
• however, those individuals in the group containing individuals with the altruistic trait but which lack the trait themselves will be favored within the group
• they do not pay the cost of the altruism but they reap the benefit
• so altruistic groups are vulnerable to invasion (by immigrant individuals or through mutation to the genes for altruism) and are, eventually, taken over by the slackers
• Group selection among family groups

Kin Selection

• helping a relative costs an individual some or all of its reproductive output
• the relative boosts its reproductive output through the help (the benefit)
• because the individual is helping a relative, it can be sure that some of the genes in the relative will be identical to its own genes
• when the relative reproduces, it is reproducing the helper individual's genes, which means the helper gets benefits
• how sure it can be of this depends on how close a relative is being helped, so there must be a discounting of the benefit for more distant relatives

A different model of Group Selection (from Sober and Sloan Wilson, Unto Others, 1998)

• There is a locus which determines altruism (A) or selfishness (S)
  • the gene frequency of A is p, of S is 1 - p
  • in a group, there are n individuals, np of which are altruists and n(1 - p) which are selfish
  • Without altruism in a group, all individuals, no matter what genotype, average x offspring
  • altruists cost themselves c offspring but help every other member of the group to have b additional offspring
  • the fitness of each genotype is

W_A  =  x - c + (b(np -1) / (n - 1))

W_S  =  x + (bnp / (n - 1))

• Altruists lose the cost of altruism but benefit from the presence of other altruists, each of which provides a benefit, b, so the total for the altruist is b(np - 1) total benefit spread over n - 1 individuals in the population
• Selfish individuals have no costs and greater benefits, as they benefit from all altruists (each altruist can not benefit itself, or it wouldn't be an altruist)
• We can use a numerical example to show that this model can select for altruists when there are more than one populations in the metapopulation

x = 10
c = 1
b = 5
	Group 1	Group 2	Total
Population Size	100	100	200
p =	0.2	0.8	0.5
1 - p =	0.8	0.2	0.5
Altruist fitness	9.96	12.99
Selfish fitness	11.01	14.04
Next Generation
n' =	1080	1320	2400
p' =	0.184	0.787	0.516

• From the above, you can see that:
  1. the fitness of altruists is less than that of the selfish individuals in each subpopulation, as expected
  2. the frequency of the altruist decreases in each subpopulation, as expected
  3. the frequency of altruists in the total population INCREASES, which is definitely not expected
• the global increase is due to the relative grow rates of the subpopulations
  • the "altruist" population grows much faster, due to the effort of the altruists
  • where selfish individuals predominate, the populations grow less
  • the overall effect is to boost the overall frequency of altruists (globally only)
• Sober and Wilson are propose that this effect (known as Simpson's Paradox) may be behind the evolution of altruistic behavior

Last updated February 27, 2008