Individual Variation and Quantitative Genetics

1
Lecture 12 Individual Variation and Quantitative
Genetics
2
Part 1 Individual Variation
3
The biological importance of individual variation
has recently been reemphasized in the fields of
comparative physiology, functional morphology,
and animal behavior, as well as eco/evo
physiology. It is also coming into medicine,
spurred by the availability of genetic data
(e.g., SNPs) that clearly demonstrate differences
among individuals (and among "racial" groups,
although that is a political hot
potato!) Childs, B., C. Wiener, and D. Valle.
2005. A science of the individual Implications
for a medical school curriculum. Annual Review
of Genomics and Human Genetics 6313-330.
4
However, an appreciation of the fundamental
importance of individual variation is not new to
these and related fields. For example, Tryon
(1942) outlined four main problems that should be
addressed when studying individual variation
5
Tryon, R. C. 1942. Individual differences.
Pages 330-365 in F. A. Moss, ed. Comparative
psychology. Prentice-Hall, New York. 1. How
consistently do individuals differ and how
large are those differences? 2. Does individual
variation in one (behavioral) domain
correlate with variation in other types of
(behavior)? 3. Does individual variation in
behavior relate to underlying differences in
physiology and morphology? 4. Are
individual differences caused by genetic
differences among individuals?
6
To these we can add the more field-oriented
question 5. Does individual variation in
morphology, physiology or behavior correlate
with Darwinian fitness in nature. In other
words, is selection acting on the individual
variation that presently exists for a
trait? This is one direct way to study
"adaptation" in the genetic, evolutionary sense.
7
Definitions Natural Selection individual
variation in fitness that is correlated with
variation in one or more phenotypic
traits. Fitness (simply) number of offspring
left to the next generation by a given
individual (measure at same stage,
zygote-to-zygote best, difficult in
practice) Components of Fitness, e.g.,
survivorship, fecundity, of mates
8
Definitions Phenotype any measurable trait
(height, metabolic rate, I.Q.) Genotype genetic
material, actual set of genes on
chromosomes Genotype (zygote) is translated into
phenotype (adult) through development during an
organism's ontogeny, and subject to many
environmental effects
9
Thus, adaptive phenotypic evolution consists of
two parts r h2 s r response to selection
evolution from one generation to the
next change in the phenotypic mean of
a population from one generation to the next
10
r h2 s h2 narrow-sense heritability
how much of the phenotypic variation in a
population is caused by genetic effects
that can be passed on from parents
to their offspring additive genetic
variance --------------------------------
------ total phenotypic variance
11
r h2 s s directional selection
differential difference in mean phenotype
between the original whole
population before selection and the mean
of the individuals who actually breed to
produce the next generation
12
Population Before Selection
Consider the distribution of a trait within a
population ...
MeanBefore
13
Population Before Selection
MeanBefore
Now imagine that a selective event occurs, e.g.,
a winter ice storm that kills most of the
individuals. s MeanAfter - MeanBefore
Population After Selection
14
Population Before Selection
MeanBefore
Population After Selection
15
Population Before Selection
MeanBefore
Population After Selection
Next Generation
MeanNext Generation
16
Population Before Selection
r h2 s
MeanBefore
s MeanAfter - MeanBefore r MeanNext
Generation - MeanBefore h2 r/s

Population After Selection
Next Generation
MeanNext Generation
17
We will discuss all of these components in more
detail, and how they can actually be
measured. For now, we need to know how to
quantify and study individual variation. Individua
l differences can be documented by measuring each
of a series of individuals multiple times and
testing for significant "repeatability" of the
differences among individuals.
18
"Repeatability" has a specific definition in
quantitative genetics, but for now just think of
it as any statistic that takes on a high value
when individuals are consistently different and a
low value when they are not. For example, is the
rank order of individual values consistent across
trial days?
19
The Western Fence Lizard, Sceloporus occidentalis
20
The rank order of sprint speed, measured on a
photocell-timed racetrack, is rather consistent
(and statistically significant) across 5
successive trial days.
Bennett, A. F. 1987. Inter-individual
variability an underutilized resource. Pages
147-169 in M. E. Feder, A. F. Bennett, W. W.
Burggren, and R. B. Huey, eds. New directions in
ecological physiology. Cambridge Univ. Press,
Cambridge, U.K.
21
Was this a measurement error?
How dowe treat "outliers" (unusual
observa-tions)?
Having a bad day?
22
Figure 1. Mass-adjusted values of VO2rest and
VO2act in 10 males of Scinax sp. 1. Bars are
standard errors of three measurements. To
facilitate visual comparisons, we present graphs
of residuals of a regression analysis of the
logarithm of body mass that have been adjusted to
the overall mean body mass (3.20 g) and then
reconverted to the original arithmetic scale.
10 Individual Frogs
Gomes, F. R., J. G. Chauí-Berlinck, J. E. P. W.
Bicudo, C. A. Navas. 2004. Intraspecific
relationships between resting and activity
metabolism in anuran amphibians influence of
ecology and behavior. Physiological and
Biochemical Zoology 77197-208.
23
Problems with and/or Objections toStudies of
Individual Variation 1. Extreme values are
atypical or abnormal and do not reflect the true
response of most individuals. Essentially a
restatement of the typological species
concept. Real populations do have highly unusual
individuals! These "abnormal" individuals do
exist and hence must be considered. Selection and
inheritance in the real world applies to them.
24
Example Garland, T., Jr. 1988. Genetic basis of
activitymetabolism. I. Inheritance of speed,
stamina,and antipredator displays in the garter
snakeThamnophis sirtalis. Evolution
42335-350. The 46 dams gave birth
between 2 August and 6 September 1984. Litter
size ranged from one to 26(mean 12.2, SD
5.15). Of the 562 offspring produced,23 were
born dead (frequencies were 10, 5, 2, 2, 1, 1, 1,
and 1 individuals born dead in each of eight
different families) another 13 were born with
obvious deformities (eight with kinked tails,
bodies, or necks one with deformed ventral
scales four with one or both eyes small or
absent) there were no more than two deformed
individuals in any one family.
25
Example 249 of the offspring were tested for
maximal sprint speed on a photocell-lined
racetrack and for endurance on a motorized
treadmill moving at 0.4 km/h. Both tests were
done twice, on consecutive days. This is a
histogram for the higher endurance
Range 1-23 min
But this does not account for the individuals
born dead or severely deformed!
Endurance (min) log10
Endurance (min)
26
2. Extreme points are attributable to
instrumentation or procedural errordo not
result from real biological differences. Typical
physiological procedure might be to do5-10
preparations (e.g., isolated muscles in
ergometer), then throw out the 1 or 2 that were
fairly different from the rest. Unusual points
assumed to represent preparation errors, e.g., a
damaged muscle. Difficult to know if the measure
cannot be repeated on a single individual. If the
"preparation" can be repeated, e.g., most
measures of whole-animal physiological traits,
then we can rule out this possibility.
27
3. The variation is real (errors of measurement
apparatus are not very large) but reflects random
and unrepeatable responses of individuals intra-i
ndividual variability is so high that there is no
significant inter-individual component to total
variance. Significant repeatability across trials
answers this.
28
Most typically, repeatability is quantified by
measuring individuals on each of two days. A
scatterplot is made. The Pearson product-moment
correlation coefficient, r, indicates
repeatability.
29
No Repeatability
N 50
R 0.066
30
High Repeatability
N 50
R 0.894
often found for morphometric traits
31
Typical Repeatability
N 50
R 0.517
for behavioral and physiological traits
32
Example
Harris, M. A., and K. Steudel. 2002. The
relationship between maximum jumping performance
and hind limb morphology/physiology in domestic
cats (Felis silvestris catus). J. Exp. Biol.
2053877-3889.
33
Example Garland, T., Jr. 1988. Genetic basis of
activitymetabolism. I. Inheritance of speed,
stamina,and antipredator displays in the garter
snakeThamnophis sirtalis. Evolution 42335-350.
N 231, r 0.696
log10 Endurance (min) Day 2
log10 Endurance (min) Day 1
34
Lacerta vivipara from southern France
35
Lacerta vivipara from southern France
Note double log transform!
Treadmill endurance is more repeatable than
sprint speed, possibly because the latter is
determined more by inherent physiological
capacities for exercise, whereas the former is
more determined by motivation?
36
Long-term Field Repeatability
37
Long-term Field Repeatability
38
Long-term Field Repeatability
39
Angilletta, M. J., Jr., P. H. Niewiarowski, and
C. A. Navas. 2002. The evolution of thermal
physiology in ectotherms. Journal of Thermal
Biology 27249-268.
40
Why is repeatability important? If a trait is not
consistent, but varies wildly from day-to-day,
then selection has no clear "target." Teleological
ly and anthropomorphically, selection cannot
penalize the bad individuals and reward the good
ones, because these get mixed up every time
selection "looks." Similarly, if the "quality" of
parents varies wildly from day-to-day, then they
cannot pass it on to their offspring. In
general, repeatability sets an upper limit to
heritability.
41
Analytical Uses of Individual Variation 1.
Testing or generating hypotheses about
functional relationships 2. Measurement of
selective importance Field studies - later
lectures correlational experimental Lab
studies - useful because ... correlational exp
erimental 3. Determining heritabilities of
organismal or physiological characters
42
Natural Sexual Selection
Act On
Behavior
Constrain
Morphology, Physiology, Biochemistry
Organismal Performance Abilities
Deter- mine
43
Comparative physiologists have routinely looked
at differences among species. Example urine
concentrating ability and length of loops of
Henle in kidneys a significant correlation
among species (or among populations) suggests
that the latter causes the former. But
interspecific comparisons are fraught with
difficulties.
44
For example, we may be comparing apples and
oranges (e.g., burrowing owl vs. bobwhite
quail). Also, species do not represent
independent data points, and so we need
independent phylogenetic information to perform
proper statistical analyses this may not be
available for the organisms that we study. One
way to avoid such problems is to stick within a
single species. If relationships really do exist,
then we should be able to demonstrate them.
45
... assuming that the "signal" is large enough to
be detected over the "noise" of short-term
variability within individuals! Hence the
importance of first demonstrating that a
physiological or behavioral measurement is
repeatable.
46
What predicts locomotor performance in the
Spiny-tailed Iguana (Ctenosaura similis) from
Costa Rica?
Garland, T., Jr. 1984. Physiological correlates
of locomotory performance in a lizard an
allometric approach. Am. J. Physiol. 247
(Regulatory Integrative Comp. Physiol.
16)R806-R815.
47
Ctenosaura similis
Calculate residuals to remove effects of body
size
48
(No Transcript)
49
(No Transcript)
50
Path Analysis
51
Analytical Uses of Individual Variation 1.
Testing or generating hypotheses about
functional relationships 2. Measurement of
selective importance Field studies - later
lectures correlational experimental Lab
studies - useful because ... correlational exp
erimental 3. Determining heritabilities of
organismal or physiological characters
52
Natural Sexual Selection
Act On
Behavior
Constrain
Morphology, Physiology, Biochemistry
Organismal Performance Abilities
Deter- mine
53
Natural Sexual Selection
Act On
Behavior
Constrain
Morphology, Physiology, Biochemistry
Organismal Performance Abilities
Deter- mine
54
Laboratory Studies to get atRelations of
Performance,Behavior, and Selection A
correlational study Do locomotor abilities
correlate with social dominance? Male Sceloporus
occidentalis Size matched because size
differences are known to have major effect on
dominance interactions Measured social dominance
in the lab by letting pairs compete for access to
basking site Measured locomotor performance in
the lab Compared winners-losers by paired t-test
55
The Western Fence Lizard, Sceloporus occidentalis
No difference in treadmill endurance
56
The Western Fence Lizard, Sceloporus occidentalis
Significant difference in sprint speedon
photocell- timed racetrack
57
But the lizards did not sprint at high speeds
during behavioral interactions, so why would
speed predict social dominance? Perhaps both
traits are correlated with a third variable
One candidate would be testosterone levels
58
A possible experimental study Does testosterone
(T) manipulation affect locomotor abilities
and/or social dominance? Use silastic implants
of T to increase circulating hormone
levels. Compare performance and behavior with
animals that received control implant (no T).
59
Analytical Uses of Individual Variation 1.
Testing or generating hypotheses about
functional relationships 2. Measurement of
selective importance Field studies - later
lectures correlational experimental Lab
studies - useful because ... correlational exp
erimental 3. Determining heritabilities of
organismal or physiological characters
60
Part 2 Quantitative Genetics
61
Variation in most phenotypic traits is continuous
or quantitative, not discrete like Mendel's
peas. Among individuals within a population, some
of the variation is caused by 1. environmental
differences experienced since fertilization of
egg (or even before) 2. genetic differences at
multiple loci Phenotypic variance of a population
is calculated as ? (Xi - Xmean)2
------------------------------- N -
1 Standard deviation is square root of variance
62
Phenotypic variance of a population can be
partitioned into various components. Simplest
partitioning VP VG VE May have
Genotype-Environment interaction a. which
genotype has the higher phenotype depends
on the environment in which rearing
occurs b. "reaction norm" or "norm of
reaction" set of phenotypes produced by
a given genotype across a range of
environments (phenotypic plasticity) So, can add
term VP VG VE VG X E
63
Also may have genotype-environment correlation,
i.e., different genotypes tend to occur in
different environments or microenvironments a.
genotypes might differentially
select microenvironments b. strongest
individuals obtain best territories c. farmer
takes better care of the best calves If so, then
need VP VG VE VG X E VGEcorr Thus,
easiest to make measurements in controlled
environments, with little or no variation among
where individuals live, and hence eliminate these
variance components.
64
However, resulting numbers may or may not be
relevant to natural populations. Some workers do
make estimates in nature. Especially for sessile
organisms, such as plants, or animals that use
nests. Many bird populations have been studied
with artificial nest boxes. In practice, for
practical reasons, workers often just lump VG X E
and VGEcorr into VE,i.e., they do not try to
estimate them separately.
65
Broad-sense heritability VG/VP But not all
genetic variance can be passed on to offspring,
only the additive genetic variance. So, a more
useful partitioning for some purposes is VP
VA VD VI VE VG X E VGEcorr VD
variance caused by dominance deviations,
i.e., non-additive interactions between alleles
at a single locus VI variance caused by
epistatic deviations, i.e., non-additive
interactions between alleles at
different loci
66
Narrow-sense heritability VA/VP This is the
heritability in the breeder's equation r h2
s Narrow-sense heritability is viewed as the
single most important descriptive statistic about
the quantitative genetics of a given trait in a
given population. It indicates the evolutionary
potential of the trait. How do we estimate
narrow-sense heritability? Resemblance of
relatives ...
67
Least-squares linear regression slope
narrow-sense heritability
N 50
If have no maternal effects or common-family
environ-mental effects
68
Least-squares linear regression slope
narrow-sense heritability
N 50
Low repeatability would add a lot of "jitter" to
the data points and reduce the slope
69
In principle, can use any relatives, just need to
know the expected causes of resemblance. For
example, in an organism that had no paternal
care, might measure offspring and only the
fathers. Double the regression slope to estimate
narrow-sense heritability.
70
Another common "breeding design" is to mate each
father (sire) with multiple mothers (dams) and
measure trait of interest in the offspring
only. This half-sib, full-sib design allows
estimation of narrow-sense heritability. In
particular, the among-sire component of variance
is proportional to additive genetic effects (if
have no non-genetic paternal effects). Can also
estimate realized narrow-sense heritability from
a selective breeding experiment because r/s
h2