Transcript 3000_2013_2d+e

independent
units of
contrasts
inheritance
association
homology drift
P=G+E,
and selection/ tests
memorizing
pop size
fitness random Mendelian
equations
v
evolutionary
Punnett
expectations
stochastic,
quant
mechanisms
squares
under
deterministic
HWE
impact
types of mutations
of
Luriaand the
Delbruck
picture was
Alfred Russell Wallace
degrees of
freedom
mechanisms of
evolution
mutation
selection
drift
non-random mating
migration
• different take on
how many genes
lead to normal
distribution of trait
• just count “big”
alleles, assume
additive
contribution
• more loci leads to
more potential
variation
phenotypic variance is caused by genotypic variance
AND variance caused by environment
Var(P)=Var(G)+Var(E)
heritability
• proportion of total phenotypic variance
caused by genotypic variance
•H
2
= Var(G)/Var(P) =
Var(G)/(Var(G)+Var(E))
• broad-sense heritability, all forms of
genotypic variance are included
• but not all genotypic variance
contributes to phenotype of next
generation equally
additive v dominance,
again...
• when alleles make
an additive contribution to
phenotype, a single allele makes same contribution
regardless of other allele
•
•
•
•
aa: white flower; Aa: some red pigment made, PINK flower; AA: more red made, RED
flower
when alleles are not additive (dominant/recessive),
their contribution depends on the other allele in
genotype
a / a: white flower; a / A: red pigment made, RED flower; A / A: red made, RED flower
that context disappears each generation with
sexual recombination; alleles are heritable, the
genotype isn’t
• slows effects of selection: recessive alleles ‘hide’ in
heterozygotes (no effect) so requires drift to increase in
frequency enough to make homozygotes (if good effect), or
difficult to purge if negative
“dominance” and “recessiveness” is a variable trait
e.g. if selection coefficient is s
we can score relative fitness as 1-hs
h is the level of dominance of that allele
0.5 is additive
but h can vary from 0 to 1 (in this plot, it is 100% dominance/recessive)
maintaining diversity
• standing variation: created by mutation
• stochastically changing in frequency via
drift
• even with selection, dominance can
maintain some allelic variation
• at level of an entire gene, there is a
balance between mutation (µ) and
selection (s)
• diversity ITSELF can be selected for!
selection and polymorphism
•
flowers and pollinators:
typical reward is nectar
•
not all orchids produce
nectar, must deceive
•
elderflowers do this with
color polymorphism
yellow flowers actually
have higher relative
fitness, but when they get
too common it disappears
and bees favor purple
actual frequency ~ 70%
frequencydependent
mhc
•
MHC molecules “show”
processed proteins on
cell surface
•
immune system
responds (usually to your
benefit)
•
extreme diversity at this
locus: why?
molecule with benefits
•
diversity allows presentation/recognition of diverse
pathogen/foreign material (helps immune system clear
body of disease)
•
greater diversity, better presumed immune response
•
•
(heterosis, overdominance: two forms of increased
fitness with heterozygosity)
so life (vertebrates) might act to increase diversity
somehow?
old shirts and
mate choice
•
Wedekind study: MHC
dissimilar mates
preferred?
•
T-shirts worn by guys,
presented to women - all
genotyped at MHC loci
•
greatest mismatch at
genotype (different
alleles) = greatest
“attraction”
being unusual and mating
Incongruity of primate species tree and DQA1
MHC-promoter
related region
gene gene tree.
Loisel D A et al. PNAS 2006;103:16331-16336
©2006 by National Academy of Sciences
•
•
•
•
many forms of
variance
remember we are asking about causes for
phenotypic variance in some trait, Var(P)
environmental variance, Var(E) contributes:
amount of sun, amount of nutrient, altitude, etc.
genetic variation Var(G) now includes additive
Var(A), dominance Var (D), epistasis (gene-gene
interactions, Var(I)), and even gene x environment
interactions
selection acts on additive variance Var(A)
h2 is narrow-sense
heritability (only involves
additive variance)
why do we care about
heritability?
QuickTime™ and a
decompressor
are needed to see this picture.
breeders equation
R=
2
hS
•if the environment selects on a
heritable trait, how will the
population respond?
• quantitative trait
loci: where are
the genes
contributing to
such traits?
• generally
requires at least
F2 display of
traits and dense
genotype
“mapping”
Small or Large?
• Question with phylogenetic variation is how
labile is the trait? How conserved?
• Are flower colors and shapes controlled by
many mutations of small effect, or can some
individual mutations have major effect?
• Need to answer experimentally, but first need
candidate loci: the QTLs
•
•
•
linkage group may refer to a chromosome
or region of the genome that are physically
linked, and thus diversity at nearby genes
is linked - not entirely independent
recombination unlinks these regions, but
frequency of recombination depends on
proximity of genes
linkage may be between known genes, or
maybe just anonymous marker and a gene
linkage, haplotype
•
haplotype - the multilocus description of a chromosome
or gamete, or other physically linked set of loci
(mitochondrion)
•
genotype - the multilocus description of an individual,
composed of haploid contributions from parental
gametes
•
an AaBBCc individual genotype could be generated by
gametes ABC+aBc, or aBC+ABc, or ABc+aBC...
linkage disequilibrium
•
loci are in linkage equilibrium when the genotype of one locus is independent
entirely from the genotype at another locus (knowing one does not predict the
other)
•
disequilibrium when there is a nonrandom association
•
3 conditions (for 2 loci) must all be met for equilibrium:
•
frequency of B on haplotypes with A is equal to B on haplotypes with a
•
frequency of any haplotype obtained by mutliplying frequency of alleles
•
for frequency g: gABgab - gAbgaB = D (coefficient of linkage disequilibrium) = 0
back to hardy-weinberg
•
extend Hardy-Weinberg analysis to 2 loci: same
conditions (selection, migration, mutation, nonrandom
mating, drift)
•
follow frequencies of multilocus genotypes created by
possible haplotypes
•
linkage disequilibrium can happen via selection, drift,
and population admixture (pooling 2 genotypically
distinct populations)
gamete/chromosome/haplo 2 ways to make ABAb
type frequencies will stay genotype, so frequency is
2gABgAb
constant in HWE
recombination
•
adults of genotype AB/AB, AB/Ab, AB/aB will always
produce some AB gametes (chromosomes) for next
generation
•
adults of genotype AB/ab will produce AB gametes only
when meiosis involves no crossing-over
(recombination, rate r)
•
adults of genotype Ab/aB CAN produce AB gametes as
long as there IS recombination (r > 0)
SNP, microsat,
whatever
estimating linkage
• given some trait P that an F2 individual can be homozygous or
heterozygous for at single locus (PP, Pp, pp; determined by
phenotype)...
• and some marker M that an F2 can be scored at (MM, Mm, mm)...
• calculate probability of observed F2, e.g. trait suggests PP, and
genotype is MM
• under LD with recombination rate r=0.1, an MP/mp F1 generates
gametes MP (45%), mp (45%), Mp (5%), mP (5%)
• so MP/MP homozygote frequency 0.45x0.45 = 0.2025
• with r=0.5 all 2-locus gametes equally likely, 0.25 x 0.25 =0.0625
45
Log of Odds: 10x=ratio of odds
so
103=1000 times more likely
0.2025/0.0625 = 3.24
LOD 0.511
cases where odds ratio is < 1
produce negative LOD score
sum LOD scores from many
individuals to see result for given
hypothesis test
find your QTLs
• find quantitative trait loci, and then use
experiments to confirm what some of these
loci do
• for example, clear relationship between
pollinator and phenotype (bees like big flowers
with less yellow pigment; hummingbirds deep,
purple ones)
• experiment: breed M. lewisii but with the QTL
for YUP locus (adds yellow)
47
adding YUP allele to M. lewisii made them
much more preferred by hummingbirds, less
so by bees
shows that some mutations/alleles can have
LARGE effect, can be quickly selected on
QuickTime™ and a
decompressor
are needed to see this picture.
“sensitive stigma” in Mimulus guttatus
stigma NOT sensitive in closely related self-fertilizing species
novel quantitative trait, continuous variation in F2
with ≥4 QTLs contributing to sensitivity
•
genes interacting with other genes is
termed epistasis. in Hoekstra’s mice, the
Agouti and Mc1r loci are in epistasis.
QTLs are themselves hypotheses
• if region seems to be involved in trait,
now use genetic crosses/manipulation
to directly test marker and how it affects
trait
phenotypic plasticity
• how does the phenotype, as determined
by genotype, respond to the
environment?
• traits that are plastic - able to be molded
• reaction norm: pattern of phenotypic
expression of a single genotype across
a range of environments
•V
P
= VA + VD + VI + VE +VGxE
1. single genotype in
different
environments
2. multiple genotypes,
but with little
variation for trait
3. multiple genotypes,
with genetic
variation for trait
4. multiple genotypes,
genetic variation for
trait and variation in
how genotype
responds to
environment
C. elegans
(Nematoda)
• some traits have little GxE effect, others
have a very strong GxE effect
plasticity
• remember ‘move, adapt, acclimate, or die’?
• one adaptive mechanism is to allocate
resources only when necessary
• trade-offs between (for example) growth or
reproduction and defense mechanisms
• reaction norm: how the phenotype of a
genotype changes in a different environment
59
Lots of fish...
...or none
phenotypic variation among 10 clones
reaction norms; genotype x environment
potential for adaptive response
• is there genotypic and phenotypic variation?
• some genotypes alter their behavior more than
others in presence/absence of fish
• variation in phenotypic plasticity is a
genotype-by-environment interaction
• when many fish present, greater plasticity
(steeper slopes of reaction norm): plasticity
has evolved!
61