Transcript quantitative traits II, NEW - Cal State LA

Modes of selection on quantitative traits
Directional selection
The population responds to selection
when the mean value changes in one
direction
Here, the mean increased from 7.0 to
7.5 after one generation
In directional selection, fitness increases
along with the value of a trait
Changes the mean, and slightly lowers
the variance (the standard deviation of
the value)
Stabilizing selection: survival of the average
Survivorship in gall-making flies
Parasitism was highest for small galls
Predation was highest for large galls
- stabilizing selection favored average-sized galls
Stabilizing selection: survival of the average
Stabilizing selection removes the
extreme values of a trait
Fitness is highest for individuals with
intermediate values
Mean value of a trait stays the same
as before selection
Trims the edges off the distribution, reducing the variance
Disruptive selection on bill size
in seed-crackers
Juveniles with either fairly large (wide + long)
or fairly small beaks survived to adulthood
Disruptive selection acts against birds with
average-sized bills; extremes have the
advantage
Disruptive selection tends to split a normal
(bell-shaped) distribution into two distinct,
non-overlapping bells curves
Does not change the mean, but increases the
variance
Identifying Quantitative Trait Loci (QTLs)
A major goal of quantitative geneticists is to identify the genes
that contribute to a given phenotype of interest
We typically want to know 3 things:
- How many genes contribute to a phenotype?
- How much does each contribute? (i.e., large or small effect)
- What are the genes, anyway?
Are there 5 genes that each contribute about 20% to your height,
or 1,000 genes that each contribute 0.1%?
Identifying Quantitative Trait Loci (QTLs)
Standard approach:
1) find “mappable” genetic variation, meaning places on each
chromosome where there is genetic variation among individuals
2) determine if genetic variation at each locus is associated with
the phenotype you are interested in
3) if yes, then you have identified a Quantitative Trait Locus or
QTL
Note: the “variation” you map is not necessarily the same as the
allelic variation that contributes to phenotypic differences, it may
just be linked to the alleles that affect phenotype
Identifying Quantitative Traits: Monkeyflowers
hummingbird pollinated
bee pollinated
Ancestral traits:
large flowers
little nectar
bee pollinated
Derived traits:
red, narrow flowers
much nectar
bird pollinated
Identifying Quantitative Traits: Monkeyflowers
hummingbird pollinated
bee pollinated
Research questions:
How many genes are
involved in flower
appearance?
How much does any
one allele contribute
to flower phenotype?
are there a few genes of
big influence, or lots of genes that each matter a little?
Answers from QTL analysis:
- up to 6 loci influenced each of 12 floral traits, but for 9 of 12
traits, one locus determined >25% of variation in flower
phenotype
Bees visit large flowers, avoid yellow pigment
- allele that increased yellow pigment lowered bee visits 80%
Birds like purple, nectar-rich flowers
- allele that increased nectar yield doubled hummingbird visits
Individual alleles can play a critical role in driving
the evolution of quantitative traits
“alleles of large effect” may thus be key
Identifying Quantitative Traits: Human Height
Advances in genomics have made it possible to determine how
many genes contribute to complex human traits, and to map
the approximate location of those genes
Determine the # of single nucleotide polymorphisms (SNP’s)
that are associated with height of the people who carry them
The assumption is that a SNP associated with greater height is
in linkage disequilibrium with an allele of some gene (typically
not identified) that makes a contribution to how tall you are
Basically: How many spots throughout the genome have genetic
variation that is associated in some way with the height of the
genotyped people?
Single Nucleotide Polymorphisms (SNP’s)
sequence alignments can tell us if people have different
nucleotides at, say, position 739,012 of chromosome 14
ATCGTGTGGAACTAATCGGCGCCGAAACTACGA
ATCGTGTGGAACGAATCGGCGCCGAAACTACGA
ATCGTGTGGAACGAATCGGCGCCGAAACTACGA
This is a SNP – it is also a QTL if people who have a T are, on
average, 1” taller than people who have a G here
i.e., if this SNP is associated with height
Note: This does NOT mean the “T” is causing more height…
it may be linked to an allele of a nearby gene that is
contributing to additive genetic variance for height
non-synonymous substitution
in coding region of gene for
human growth hormone
substitution in the
non-coding region
(“junk DNA”) just
upstream of GH1
G
A
T
“A” allele
Which SNP(s) are likely to be
associated with human height?
A
G
substitution in non-coding DNA
1,000,000 bases upstream of GH1
C
“a” allele
GH1 gene
QTL’s and Human Height
Adult height is a classic quantitative trait, and is highly heritable
- 80% of the variation in height in a population is due to
additive genetic variance
However, over 40 identified QTL’s explain less than 5% of
heritable variation in height
Are we missing a few major genes that make large contributions
to height?
Or are we missing hundreds of minor players?
QTL’s and Human Height
Recent study examined 183,727 individuals, comparing each
person’s height to their genotype at 2.8 million SNPs
Identified 180 loci influencing adult height
- loci include many genes connected in cellular pathways,
or that contribute to skeletal growth defects
- many SNPs alter amino-acid structure of proteins or
expression levels of nearby genes, meaning the SNP really
is directly affecting the phenotype
- some were also associated with diseases (arthritis, diabetes)
suggesting pleiotropic effects of certain alleles on phenotype
However, these 180 loci explain only 13% of variance in height
Allen et al. 2010, Nature 467, 832–838
QTL’s and Human Height
Estimated there are ~700 loci that together would explain 20%
of genetic variance in height, but would need to sample
500,000 people to identify the additional loci
So where is the missing variation?
3,925 people were genotyped for 294,831 SNPs, but instead of
looking for SNPs associated with a certain amount of height,
Yang et al. used all SNP data to model height distribution
- explained 45% of height variance
Thus, most heritability is not “missing” but escapes detection
because most effects of individual loci are too small to pass
significance tests
Yang et al., Nature Genetics 42, 565–569