Lec 7 Flashcards
The Neutral Theory of molecular evolution
In the 1960s enzyme electrophoresis provided evidence of cryptic molecular variation: differences in amino acid sequences that do not manifest in phenotypic differences
- Allowed us to visualize differences in amino acids
- Synonymous mutations make changes at nondegenerate/redundant bases = no phenotypic change
At the time it was thought that natural selection was responsible for the maintenance of genetic variation, but these variations appeared to be neutral
The prevalence of apparently neutral variation led to debate about whether natural selection is really an important driver of evolutionary change
Why are there so many neutral variants in the genome?
a) Most of the genome is non-coding DNA, so mutations accumulate that don’t have fitness consequences
b) Mutations with large phenotypic effects are typically deleterious and are weeded out by natural selection
c) Mutations in coding regions can be synonymous and not have phenotypic effects
d) All of the above
d) All of the above
The Neutral Theory
Motoo Kinura compared amino acid sequence differences in well-studied vertebrate proteins
By extrapolating backward to the common ancestor of the two, using fossil evidence Kimura estimated that a new mutation must have been formed and achieved fixation every 2 years on average
Motoo Kimura (1968):
- The vast majority of molecular variation within species and DNA differences between species is selectively neutral
- Genetic drift is the major driving force of evolution on the molecular level
Tomoko Ohta (1973): -"Nearly neutral" theory: The vast majority of molecular variation within species and DNA differences between species is the regime
Synonymous substitutions
2 sea urchin species have several substitutions int he Histone 3 gene but all are silent
NO amino acid change
Selection and effectively neutral alleles
In infinite populations, selection cannot operate effectively on mutations that have extremely small fitness consequences
The random change in allele frequencies due to drift overwhelms any effects due to selection
Even alleles with positive or deleterious fitness effects can be effectively neutral
The neutralist-selectionist debate
Fraction of mutations that have that selection coefficient on y-axis
The vast majority have LOW selection coefficients
Neutralists will point out that many have strong selection coefficients and that most are NEUTRAL
Positive selection = favors new mutations
Negative selection = acts against new mutations
The neutral theory
The neutral theory has been very helpful for understanding how stochastic (random) process can contribute to evolutionary dynamics
With the advent of population genomic data it has become increasingly clear that natural selection does often play an important role in molecular evolution
Neutral theory is a useful null model for molecular evolution
Allows us to make predictions about how much variation and allele frequency change we expect if selection is NOT operating
Deviations from these predictions are then further support for selection
In particular, we use ratios of synonymous to non-synonymous mutations to detect selection
More non-synonymous changes than expected is often evidence of positive selection
Drift is a more important driver of allele frequency change than selection when:
The population is small and selection is weak
Population genetics at multiple loci
Thus far we have been working with a single locus with two alleles that determines phenotypes with major fitness effects
Is that realistic? NO
Quantitative vs. Qualitative
Qualitative (discrete) traits:
Traits with discrete distributions
Quantitative (continuous) traits:
Traits with continuously distributed phenotypes - they are determined by the genotype at many loci and environmental factors. These are much more common
Evolution of polygenic traits
Darwin believed that natural selection acted on continuous variation, but Mendel showed that variation was discrete
The first step in reconciling those two ideas was to recognize that some traits are polygenic - they are coded by multiple genes
Polygenic traits show almost continuous variation, but each individual allele is inherited in a particulate way
Reconciling Darin and Mendel
Experiments with kernel colors showed continuous variation
Color have additive genetic effects - the phenotype of one individual is the result of the sum of the effects of each allele it carries
Although more complicated, phenotypes were still predictable under Mendelian inheritance
Small, graded effects are critically important to genetic variation and natural selection, and consistent with Mendel’s laws of inheritance
Look at this figure. Which is not true about polygenic traits?
The phenotypes of offspring are completely unpredictable based on the genotypes of the parents
Latent Variation
Experimental data showed that polygenic traits produced the necessary variation for natural selection to operate
But in some cases new phenotypes not seen in the parental population could appear in the offspring
How can we explain the emergence of novel phenotypes?
Ten loci with 2 alleles each could produce 60,000 different phenotypes
Some populations may be too small (or we have observed too few individuals) for all of these phenotypes to be detected
When a new phenotype is observed is not necessarily the result of mutation - it could be a new assortment of previously occurring Mendelian variation
Populations thus contain latent variation - undetected phenotypes that can appear from existing genetic variation
Epistasis (gene interactions)
In some cases the interaction of multiple loci do not produce additive effects but result in epistatic effects
The phenotypic effects of one allele at a locus are determined by the allele present at the other locus
Phenotypic effects are context dependent
Mendel’s Laws of Derived from experiments
The Law of Independent Assortment: Which allele is passed down to the next generation at one locus is INDEPENDENT of which allele is passed down at another locus
- Mendel also experimented with other traits like seed shape
- The allele passed down at one locus (e.g. flower color) is INDEPENDENT (not influenced by) the allele at another locus (e.g. seed shape)
- Today we know that story is more complex
Allele and Haplotype frequencies
We have been discussing how we measure allele frequency change - how common a particular allele at a particular locus is in a population
Now we extend these ideas to multiple loci - the multilocus genotype of a chromosome or gamete
This is called the HAPLOTYPE: A set of alleles, one at each locus we are interested in
Here we have two chromosomes, one with the haplotype AB and one with the haplotype ab
This diploid organism would have the genotype AaBb
When we talk about haplotypes, we often mean a set of alleles that are inherited together because they are on the same chromosome
To understand the population genetics of multiple loci, we need to keep track of halplotypes - which alleles occur together - to help us predict genotypes
We need to remember that loci that are __________________ on chromosomes are more likely to be associated with each other - loci across the genome are NOT independent
Closer together
For example, if I tell you that in this classroom: 50% of the students have blonde hair 50% of the students have brown hair and 50% of the students have blue eyes 50% of the students have brown eyes
How many students have brown hair and blue eyes?
Not enough information to determine
We could have:
Perfect equilibrium: Alleles are inherited independently as Mendel predicted (and like his pea plants)
- 25% blonde and blue eyes (A1B1)
- 25% blonde and brown eyes (A1B2)
- 25% brown hair and blue eyes (A2B1)
- 25% brown hair and brown eyes (A2B2)
Perfect disequilibrium: Alleles are always inherited together
- 50% blonde and blue eyes (A1B1)
- 50% brown hair and brown eyes (A2B2)
Physical linkage
When the A and B loci are on separate chromosomes, the alleles segregrate independently. A double heterozygote produces four different gamete types, each with equal frequency
When the A and B loci are CLOSE together on the SAME chromosome, the alleles segregate together, In the absence of recombination, a double heterozygote produces only two gamete types
Recombination breaks up associations between alleles
Therefore generating new haplotypes
Alleles that always appear close together are often close together on chromosomes
A haplotype is
The combination of alleles present across multiple loci on a single chromosome or gamete
Alleles that always appear together in HAPLOTYPES are often close together on chromosomes
Because these loci are close together, we are UNLIKELY to get a mix/recombination
Linkage disequilibrium
Two loci are in linkage EQUILIBRIUM when the genotype of a chromosome at one locus is INDEPENDENT of its genotype at the other locus
Two loci are in linkage DISEQUILIBRIUM when there is a non-random association between the genotypes at both loci
The non-random association of alleles at different loci
Loca are said to be in linkage disequilibrium when the frequency of association of thier different alleles is HIGHER or LOWER than what would be EXPECTED if the loci were independent and associated randomly
How do we quantify linkage disequilibrium and clarify our expectations about independent vs non-independent assortment?
Build a model
Linkage disequilibrium is how we describe ______ that are associated with each other
Alleles
How many possible combinations of alleles (haplotypes) are there?
4
AB, Ab, aB, ab
Each combination of alleles = 1 haptlotype
Is each haplotype (combination of alleles) equally likely?
First calculate the frequency of each haplotype in our population (n = number of alleles/haplotypes, f = frequency)
Next, calculate the frequency of each allele
-Add up any haplotype that has the allele of interest (in this case, add up all with A)
Now calculate probabilities of two alleles occurring together assuming loci are independent:
fA = fAB + fAb fB = faB + fAB
What is the change that A and B occur together?
fAB = fA x fB
If these are INDEPENDENT, the probability of getting AB haplotype is the same as flipping a coin
But A and B are NOT independent - they occur on the same chromosome
Coefficient of linkage disequilibrium
D
DAB = fAB - fAfB
DAB = coefficient of linkage disequilibrium between alleles A and B
fAB = observed haplotype frequency
fAfB = Haplotype frequency calculated assuming independence
The difference between the OBSERVED frequency of the AB haplotype and the EXPECTED frequency under independence
If we see a big difference between these 2 values, we know that there IS linkage disequilibrium
If there is NO difference, there is NO linkage disequilibrium (expected = observed)
There is strong linkage disequilibrium when
The observed frequency the AB haplotype is very different from the expected frequency
Linkage disequilibrium would continue if ______ are inherited together
Haplotypes
What reduces linkage disequilibrium?
Genetic recombination
Randomizes allelic combinations, reducing linkage disequilibrium
When recombination occurs in an individual that is heterozygous at only one locus, no new haplotypes are created
When recombination occurs in a double heterozygote, new haplotypes are produced
Sexual reproduction
DECREASES linkage disequilibrium because it allows for genetic recombination of loci
Asexually reproducing organisms don’t have any recombination; recombination rate, r, it zero
What creates linkage disequilibrium?
Several evolutionary forces can create linkage disequilibrium
- Mutation
- Natural selection
- Migration
- Drift
To understand and predict patterns of evolutionary change - and particularly patterns of adaptation - we need to understand how loci are linked together
Mutation
Mutation has produced the haplotype ab, but the corresponding haplotype Ab is not yet present in the population
Therefore ab are in linkage disequilibrium
If loci are far apart, recombination will break up these associations over time
Migration
a and b loci are fixed on the mainland, while A and B are fixed on the island
When ab migrants reach the island, there will be a statistical association between alleles on the island
This will break down over time, depending on how close alleles are on chromosomes
Drift
On linkage disequilibrium
Imagine a small population with four haplotypes: AB, Ab, aB , and ab
There are low recombinations rates between A and B
Drift can lead to the loss of one allele, for example B
What would then happen to A?
Linked alleles can also be lost to drift
-Results in linkage disequilibrium between remaining alleles
Selection and linkage disequilibrium
Selection has distinct and measurable effects on linkage disequilibrium
Identifying LD is therefore one of the main ways that we identify genes under selection
A population of mice has brown or white fur, and blue or yellow eyes. White fur is under positive selection. If white fur and yellow eyes are in linkage disequilibrium, what pattern might you expect to find?
The frequency of yellow eyes increases
Since allele for yellow eyes is next to allele for white fur, it will also increase since white fur is favored in the populations
Selection and haplotype frequencies
A or B, but not both, is needed to produce an essential molecular product
Selection disfavors ab, but only in homozygous aabb indiviudals
The ab haplotype will be less common than expected among surviving adults given the allele frequencies int he popualtion
Genetic Hitchhiking and selective sweeps
Genetic hitchhiking = selective sweep
Neutral (or even weakly deleterious) alleles closely linked to alleles under selection can increase in frequency because of proximity to selected allele
In a scenario with recombination, what effect do selective sweeps have on genetic variation (heterozygosity) around a selected locus?
Decrease genetic variation
Effects of selective sweeps and recombination on genetic variation along chromosomes
Over time allele (and entire haplotype) will increase
Lose some of alleles due to recombination
Results in REDUCTION of variation the closer you get to the selected locus due to less recombination
REDUCE heterozygosity
Periodic selection
Each time a new beneficial mutation arises, it sweeps to fixation
-Everything close to favored allele will also sweep to fixation
In the absence of recombination, it also fixes the particular haplotype on which it arose
This is what happens with antibiotic resistance
You are examining genetic variation within a population and discover several regions of a chromosome with very low heterozygostiy. What might you hypothesize is the cause?
Positive selection led to selective sweeps in these regions
Evolution of dog-like behavior in Russian foxes
60 years ago, Russian scientists started breeding foxes for “tameness” behavior
They select foxes to breed exclusively based on a simple behavioral measure: whether the fox approaches humans or not
After 4 generations, foxes began wagging tails. After 6, lick human’s hands and roll over
Exhibit high frequencies of floppy ears, mottled coats, and curly tails
Why do new traits emerge in foxes? (curly tail, floppy ears, mottled coats, etc.)
Traits located close together on chromosome
Selectively breeding for “friendly” behavior, gene may be attached to many other traits
If a fox with curly tail and is friendly, more foxes with curly tails due to selective breeding
Genetic hitchhiking and background selection
HIV virus sequences from one patient (i.e. a population of viruses)
At start of experiment, all virus strains are susceptible to antivirals
How many haplotypes are there? How much genetic variation?
8 haplotypes, lots of genetic variation
After 84 generations, a mutation conferring resistance has arisen
What happened to the probability of fixation of neutral mutations around the selected locus?
Increased probability of fixation
Genetic variation has DECREASED
Background selection
Genetic variation is also reduced around DELETERIOUS loci
Just as positive selection can increase the frequency of neutral alleles and decrease genetic variation, so can selection reduce diversity by removing alleles linked to deleterious alleles
In both positive and negative selection, you WILL get reduced heterozygosity around deleterious loci
Recent work suggests linkage is a crucial process affecting the amount of neutral molecular variation
Not clear that the second pillar of the neutral theory, random genetic drift, is always the dominant stochastic
Genetic hitchhiking and background selection (and clonal interference and genetic draft) are ALL crucial processes that also contribute to random neutral variation
Quantitative Genetics
The branch of evolutionary biology that deals with the analysis of evolution at multilocus traits
-Traits controlled by multiple different genes
Quantifies relative influences of genes and environment on continuously distributed phenotypes (i.e. phenotypic plasticity)
- Sometimes variability is due to environment (phenotypic plasticity)
- Only traits with a genetic component can evolve/respond to natural selection
The central goal of quantitative genetics is to predict how continuously varying traits will respond to selection (in other words, how much of the variation we observe in phenotype is genetic)
To figure out how a continuous trait will respond to selection, we need to figure out how much of the variance in that trait is due to genes vs. environment
In other words, are 2 indiviudals, 1 with trait value of 2.5 and 1 with trait value of 1, different becuase they have different genes or because they live in different environments
Measuring Variation
Variance is the statistical measure of the amount of variation in a sample
Different members of a population have different trait values, and the variance tell us how different these indiviudals are from each other
Greater variance = greater differences among individuals
Which population has grater variance?
tan is flatter and more spread out, blue has stronger trend in one place
Tan
Which population has greater variance? (blue and tan roughly equal in height and spread)
Same variance
Means are different, but distributions similar
Quantifying Heritability (amount of variation due to genes)
To do this, we separate variation in a trait, VP, into the genetic variance (VG) and the environmental variance (VE)
Total variance:
VP = VG + VE
Broa-sense heritability is proportion of total variation due to genes
H^2 = VG/(VG + VE) = VG/VP
How do we measure heritability?
We need to figure out a way to measure genetic and environmental variance
For model organisms, we can compare the amount of phenotypic variance among inbread lines of genetically identical individuals
Each line is genetically identical, so any differences among indiviudals = VE
We then look at how much variation there is among multiple inbred lines - any difference among lines is due to genes
We can’t make inbred lines of people, so researchers estimate heritability of continuous traits in humans by comparing identical vs. fraternal twins
Twins are raised in the same environment, but identical twins are 100% genetically related to each other while fraternal twins are only 50% related
If identical twins are very similar in a particular trait, but fraternal twins are less similar, that suggest the trait has a genetic basis
Here we see the correlation in phenotype (in this case, presence of clinical depression) in identical (MZ) vs fraternal (DZ) twins. A larger correlation coefficient means twins are more similar. What can we infer?
Strong genetic component to depression
Correlation coefficients are much higher in identical twins = other sibling is more likely to share depression when identical
Heritability is further divided into ____________ and _______________. From now on, we focus on _____________
Broad-sense, narrow sense, narrow sense
Narrow-sense heritability
The amount of phenotypic variance attributable to additive effects of genes. It removes dominance and epistasis. We define it as:
h^2 = VA/(VA+VE+VD+VI)
Measures the extent to which offspring resembe thier parents (a key component of natural selection)
There are 3 classic experimental designs that allow you to measure narrow-sense heritability:
- Truncation selection experiments
- Parent offspring regression
- Cross-fostering
A parent-offspring regression examines the correlation between the average trait value the parents to the average trait value in the offspring. What type of relationship would you expect to see if the trait is heritable?
Positive correlation between parent trait and offspring trait
Parent-offspring regression
Compare average trait value in parents to average trait value in offspring
Slope of regression line = h^2
Steeper slope = larger heritability = bigger role for genetics in trait variation
Truncation selection experiments allow you to measure the phenotypic response to selection
1) Measure the trait in a bunch of individuals
2) The trait we will look at is how long it takes ravens to approach a novel object (a measure of BOLDNESS)
3) We calculate a mean boldness for the population
4) Then we decide on a “truncation line” - the boldness level that is allowed to breed
Truncation Selection Experiments
x0 = population mean = 60s
Truncation line = 80s = only individuals 80s or slower are allowed to breed (these are less bold individuals)
x1= mean approach score of breeders
Calculate SELECtion DIFFERENTIAL = maximum potential change in trait value in one generation if trait is 100% genetic
- x1 - x0 = S
- S = selection differential
Calculate the RESPONSE to selection observed in generation 2 = average of how long the next generation takes to approach
- x2 = mean for generation 2
- x0 = mean for generation 1
R = x2 - x0
Heritability = the proportion of phenotypic variance due to additive genetic variation h^2 = R/S = observed response/max possible expected response
In example, h^2 is 0.33 (10/30)
- 1/3 of variance in approach speed due to genetic varaince
- Extent to which we can predict an individual’s trait value based on the trait value of it’s parents is 33%
You are curious about whether the number of leaves a plant produces is heritable. You have a population of plants with an average of 4 leaves per plant. You do an experiment where you only allow plants with 8 or more leaves (average = 9) to reproduce. In the next generation, the average number of leaves in the population is 5. Which is correct about this population:
S = 5, R = 1, h^2 = 1/5, 20% of the variation in leaf number is due to genes
S = x1 - x0 = 9-4 = 5 R = x2 - x0 = 5-4 = 1 h^2 = R/S = 1/5
Cross-fostering experiments
Remove young from their parent(s) and have them raised by unrelated adults
Offspring traits that resemble foster parents are environmental
Traits that resemble genetic parents are genetic
Can be used to determine which traits are environmental vs heritable
Average h^2 of different trait types varies
It is important to understand that heritability is a property of a population, not a particular trait
Heritability for the same trait can vary among populations of the same species, depending on environmental context and strength of selection
What will happen to genetic variation if there is strong directional selection for a long time? What effect will that have on heritability?
Genetic variation declines, heritability declines
