multiple traits and pleiotropy JBW lecture Flashcards
define pleiotropy
a single gene/locus affects the expression of multiple traits; traits are not independent
define linkage disequilibrium
alleles at different loci affecting different traits are co-inherited
define linkage group
alleles for 2 traits are close and co-inherited
covariances
covariances measure associations between traits; indicates amount of shared variation between 2 factors; indicates degree and direction of shared deviations from the mean
positive covariance (for trait x and trait y) ie cov>0
positive deviations in x tend to be found with positive deviations in y (ellipse pointing forwards)
negative covariance (for trait x and trait y) ie cov<0
positive deviations in x tend to be found with negative deviations in y (ellipse pointing backwards)
correlations have no units
correlations are covariances but in arbitrary scales - measured on arbitrary scales (variances use units/scale and are given in squared units)
what do covariances depend on?
on the scale of the two traits being examined eg cov between body size and arm length in kgcm units
when cov=0 what does graph look like?
graph is cluster of data in vague circle around bivariate mean - no association between trait 1 and 2; no surplus of individuals with particular ratio of traits
sources of covariation
partition into different sources: genetic covariance - genetically-based association between traits; additive genetic covariance - heritable association between traits eg long-armed tall children; environmental covariance - envtl. effects jointly influ 2 traits
linkage disequilibrium as a source and what breaks down LD?
a source of gen correlation; LD creates a gen assoc. betw. traits but v. hard to build into models; LD is broken down by recombination - unless selection keeps recreating an association, recombination will remove one - need v. strong selection.
how many generations gets rid of LD if no selection
if no selection, 5 generations all lost; with average selection, in 25 generations most LD lost
Pleiotropy
one locus affects more than one trait
Causes of pleiotropy
same gene expressed in different tissues; shared developmental basis; physical interactions (muscle on bone, tissue fields); causal, where trait 1 causes trait 2 eg high resting metabolic rate and body weight; autocorrelation eg height at yr1 and height at yr2
what’s the difference between phenotypic covariance and genetic covariance (or correlation?)
phenotypic covariance involves the values of the two traits you would measure on an individual; genetic covariance represents the average phenotypic value of trait 1 and trait 2 for a given genotype (requires replicating the genotype multiple times)
what do the points on an additive genetic covariance graph show?
points represent the breeding value for trait 1 and trait 2 for a given genotype/individual (it requires information about offspring/progeny)
what do the points on an environmental covariance/correlation show? eg fecundity (x) v. longevity (y)
points represent the mean value for trait 1 and trait 2 of multiple individuals experiencing the same environment; low fecundity and longevity represents a hard envt., high fecundity and longevity represents a benign envt. [this is why genetic relationship between traits in a population is so hard to understand]
why do we care?
genetic correlations determine the multivariate response to selection; correlations and patterns of pleiotropy can tell us whether there are ‘evolutionary constraints’; correlations and patterns of pleiotropy can reveal trade-offs; genetic correlations can bias patterns of macro-evolution
what is the multivariate response to selection?
the evolution of multiple traits are tied together through genetic correlations; selection acting on one trait shifts the distribution of other traits;
break multivariate response to selection down to univariate response to selection
univariate response of a trait depends on: additive genetic variance of trait; phenotypic variance of trait; where these two determine the heritability; strength of selection on trait measure selection using a selection gradient - the linear relationship between the trait and relative fitness.
two traits response to selection
additive; additive and dominance effect on two traits; overdominance on two traits; patterns can be complex and consequences vary
what is correlated response to selection?
selection on one trait leads to evolution of another trait when there is genetic covariance; the size of covariance affects how much the other trait responds/is dragged along
what is multivariate response to selection
total response to selection on a trait is the sum of the direct (univariate) and correlated responses nb there can be many traits. nb might see constraints if some selection is pulling in opposite directions.
evolutionary constraints
genetic correlations can limit the response to selection for some trait combinations ie in some directions; if correlations are large enough, some combinations may be impossible ie the constraint is absolute;
what happens to the population if selection is aligned with pattern of co-variation?
See a responsive population - both selections positive; nb remember there can be many traits.
evolution can be seen as a process of optimisation, but with many traits can an organism optimise all traits?
Trade-offs…: can be ‘absolute’ constraints when they arise from some underlying factor; sharing the pie eg resource trade-offs (if resources are used for one structure, they are not available for another); longevity/fecundity; seed size/seed number
antagonistic pleiotropy (where traits have a genetic basis)
Antagonistic pleiotropy is where traits show a trade-off or are otherwise negatively related to each other: alleles that increase value of one trait, decrease value of the other; cannot increase both traits at the same time; ‘deleterious’ alleles (eg those causing disease) might show antag. pleiotropy - positive and negative selection effects
why can phenotypic correlations be misleading? Or, why do we need to look at genetic correlations?
environments vary in quality in ways that are not observed so what looks like a positive association eg longevity/fecundity, can easily be a change in environmental quality. Need to look at genetic correlations ie the same genotypes many times and in a wide range of envt. and then take the mean value.
Describe evolution on a fitness surface
populations evolve to the peak; correlations bias direction of evolution; direction of high variation is Gmax; evolutionary lines of least resistance;
Brain Body Size Allometry;
Define Gmax
multivariate direction of greatest genetic variation;
bias in evolution towards Gmax can lead to genetic constraints;
Gmax shows the longest skewer through all variation;
it’s a directional combination of traits with most variation;
is Gmax a snapshot view of variation present or is it a property of the population?
Genetic architecture events combine to provide new opportunities within that environment because of its new trait combinations
So the outliers, eg human cranium size are successful because of events that happened in the past.
How can Gmax be calculated?
Bottleneck a population and measure G;
Look back at ancestors - if all G similar - average them to get a representative Gmax
What is G?
The genetic variance-covariance matrix;
