Quantitative Genetics Biology 140

Quantitative Genetics

Complex traits Three categories of traits found to have complex

inheritance:
 influenced by the alleles of 2 or more genes
and the environment  Continuous traits or quantitative traits

 aka multifactorial traits because of multiple  Meristic traits

genetic and environmental factors implicated
in their causation  Threshold traits

 show complex inheritance

Estimating the number of genes involved in a

Four sources contribute to phenotypic variation: polygenic trait
…

 Genotypic variation A. Determining the fraction of the F2 (resulting from

selfing the F1 hybrid between 2 pure varieties) that is
 Environmental variation as extreme in its phenotype as that of one of the pure
parental strains
 Variation due to genotype-environment
interaction No. of gene loci 1 2 3 … n
Fraction of F2 as
 14
n
1 1 1
 Variation due to genotype-environment extreme as one parent 4 16 64
association

Two pure breeding strains of a theoretical plant have heights

Example:
of 22 cm and 10 cm, respectively. When they were crossed
the F1 plants were 16 cm tall. Self-fertilization of the F1’s
In a plant, height varies from 6 cm to 36 cm. When produced F2 plants of 7 different heights with the following
6- and 36 cm plants are crossed, all F1 plants are proportions:
21 cm. In the F2 generation, a continuous range of
heights is observed. Most are around 21 cm, and 3 1/64 (22 cm); 6/64 (20 cm); 15/64 (18 cm); 20/64 (16 cm);
15/64 (14 cm); 6/64 (12 cm); and, 1/64 (10 cm).
of 200 are as short as 6-cm P1 parent.
How many gene pairs are involved and how much does each
(a) How many gene pairs are involved? gene contribute? What is the probability of obtaining a height
(b) How much does each additive allele contribute of 18 cm?
to height?
(c) List all genotypes that give rise to 31-cm plants.

JPQ/BSS/AMAM/AUL 1
Quantitative Genetics Biology 140

P1 t1t1 t2t2t3t3 x T1T1T2T2T3T3

10 cm 22 cm
25

F1 T1t1T2t2T3t3 x T1t1T2t2T3t3 20
20
16 cm 16 cm

Frequency
15 15
15

F2 t6 + 6t5T +15t4T2 + 20t3T3 + 15t2T4+ 6tT5 + T6 10

10 cm 12 cm 14 cm 16cm 18 cm 20cm 22 cm 6 6
5
1 1
0
10 cm 12 cm 14 cm 16 cm 18 cm 20 cm 22 cm

Height

Height distribution among plants with three pairs of contributing

genes

B. Variance method
aabbccdd
AABBCCDD

  
2 2 2 2 All variation is

GF2 PF2 PF1  2

GF2 a N
2
environmental
(VE)

where a is the contribution of each additive allele and N is the

number of pairs of genes involved in the quantitative trait.
An estimate of a is obtained by the formula AaBbCcDd

aD
2N All variation is environmental (V E)

where D is the numerical difference between the two parental

means
2 2
 PF
2
2
  PF
2
  GF
2
a N D
1 2 8N2
from which
Variation is the result of both
environmental factors (VE) and
2
ND genetic segregation (VG)


8  PF
2
2
  PF
2
1


Assumptions:
Example:

 All generations are reared in the same The Flemish breed of rabbits has an average body
environment weight of 3,000 grams. The Himalayan breed has a
 The alleles of each gene are additive mean of 1,875 g. Matings between these two breeds
produce an intermediate F1 with a standard deviation of
 The genes contribute equally to the trait 162 g. The variability of the F2 is greater, as indicated by
 The genes are unlinked a standard deviation of 230 g.

 The original parental strains are homozygous (a) Estimate the number of pair of factors contributing to
for alternative alleles mature body weight in rabbits.
 No dominance
(b) Estimate the average quantitative contribution of
 No interaction each active allele.

JPQ/BSS/AMAM/AUL 2
Quantitative Genetics Biology 140

0.5
Note 0.4
0.4

0.3
0.3
0.2
 If there are n pairs of genes, then the kinds and 0.2

0.1
0.1

proportions of phenotypes of offspring from 0

0
4C 3C 2C 1C 0C
2C 1C 0C
crosses of individuals heterozygous for the n 1 gene pair 2 gene pairs
genes would be given by the terms of the
0.4 0.3
expansion of the binomial (C+c)2n where C is 0.3
0.25
0.2
0.3
0.25
0.2
contributory gene and c is non-contributory. 0.2 0.15
0.1
0.15
0.1
0.1
0.05 0.05
0 0 0

 The number of phenotypes increases with the 6C 5C 4C 3C 2C 1C 0C 8C 7C 6C 5C 4C 3C 2C 1C 0C 10C 9C 8C 7C 6C 5C 4C 3C 2C 1C 0C

3 gene pairs 4 gene pairs 5 gene pairs

number of pairs involved and the binomial
distribution (C+c)2n approaches normal Relative frequencies (y-axis) of genotypes (given as number of
distribution when n is large. contributory alleles in the x-axis) produced from crosses between
individuals heterozygous for various numbers of independently
segregating gene pairs.

Example
Modifications of the Additive Effects
The minimum weight of a fruit is 10 g and there are two pairs
of genes involved with each gene contribution equal to 3 g.
 Non-additivity and interactions of various kinds will
cause the distribution curve to appear asymmetric (W + w)4 = W 4 + 4W 3w + 6W 2w2 + 4Ww3 + w4
22g 19 g 16 g 13 g 10 g
 Dominance between alleles
If there is dominance, W1W1 and W1w1 will have the same
effect and will contribute only 3 g. The ratio will now be
9 (16 g): 6 (13 g): 1 (10g)

Skewed Distribution in Traits Involving Dominance in Two pairs of Polygenic

Genotypes and Phenotypes Inheritance

10
1 W1W1W2W2 2 W1w1W2W2 1 W1W1w2w2 2 W1w1w2w2 1w1w1w2w2 9
1 (16 g) 2 (16 g) 1 (13 g) 2 (13 g) 1 (10 g) 9
8
2 W1W1W2w2 1 w1w1W2W2 2 w1w1W2w2
2 (16 g) 1 (13 g) 2 (13 g) 7
Frequency

6 6
6
4 W1w1W2w2
5 No Dominance
4 (16 g) 4 4
4 With Dominance

3
2
1 1 1
1
0
10 13 16 19 22
Weight (grams)

JPQ/BSS/AMAM/AUL 3
Quantitative Genetics Biology 140

No Dominance W/ Dominance
W1 W2 W1 W1 W2 W2 4C 16 g
W1 w2 W1 W1 W2 w2 3C 16 g
Discontinuous Traits and Polygenes
W1 W2
w1 W2 W1 w1 W2 W2 3C 16 g
w1 w2  e.g. number of vertebrae, number of finger or toes,
W1 w1 W2 w2 2C 16 g
presence or absence of resistance to diseases, presence or
W1 W2 W1 W1 W2 w2 3C 16 g
absence of affliction
W1 w2 W1 W1 w2 w2 2C 13 g
W1 w2
w1 W2 W1 w1 W2 w2 2C 16 g
 those individuals below the “threshold” number of
w1 w2 W1 w1 w2 w2 1C 13 g
contributory genes will have one phenotype and those at
W1 W2 W1 w1 W2 W2 3C 16 g threshold value or above it will have another phenotype
W1 w2 W1 w1 W2 w2 2C 16 g
w1 W2
w1 W2 w1 w1 W2 W2 2C 13 g
w1 w2 w1 w1 W2 w2 1C 13 g
W1 W2 W1w1 W2 w2 2C 16 g
w1 w2 W1 w2 W1w1 w2 w2 1C 13 g
w1 W2 w1 w1 W2w2 1C 13 g
w1 w2 w1 w1w2 w2 0C 10 g

Example Review of Statistical Tools for

A certain affliction may be controlled by three pairs of genes and
Characterization of Continuous Traits
presence of less than three contributory genes would lead to the A. Mean
manifestation of the disease. Thus, a marriage between two 𝑛
𝑖=1 𝑥𝑖
Sample Mean: 𝑋=
normal individuals but heterozygous for the three pairs would 𝑛
produce afflicted and normal children with the following
probabilities: B. Variance
𝑁 2
𝑖=1 𝑋𝑖−𝜇
Population Variance: 𝜎2 =
(C+c)6 = C6 + 6C5c + 15C4c2 + 20C3c3 + 15C2c4 + 6Cc5 + c6 𝑁
N N N N A A A 𝑛 2
𝑖=1 𝑋𝑖−𝑋
Sample Variance: 𝑠2 =
𝑛−1
N = Normal = 1/64 + 6/64 + 15/64 + 20/64 = 42/64 or 21/32
𝑛 2 𝑛 2
𝑛 𝑖=1 𝑥𝑖 −( 𝑖=1 𝑋𝑖 )
A = Afflicted = 15/64 + 6/64 + 1/64 = 22/64 or 11/32 =
𝑛(𝑛−1)

Review of Statistical Tools for Standard error versus Standard

Characterization of Continuous Traits deviation
C. Standard deviation
𝑁 2
• SE describes the uncertainty due to sampling error in
𝑖=1 𝑋𝑖 − 𝜇 the mean of the data
Population standard deviation: 𝜎 =
𝑁
𝑛 𝑛 2
𝑖=1 𝑋𝑖 −
𝑛 𝑋 2
𝑖=1 𝑖 • SD describes the dispersion of the data
Sample standard deviation: 𝑠 = 𝑛(𝑛−1)

• As sample size (n) increases, the sample mean and SD

D. Standard Error of the Mean (SEM) tend to approach more closely the population mean
and SD
𝑠
𝑆𝐸𝑀 = • As n increases, SE tends to decrease
𝑛

JPQ/BSS/AMAM/AUL 4
Quantitative Genetics Biology 140

Review of Statistical Tools for

The Simple Linear Regression Equation
Characterization of Continuous Traits
E. Simple Linear Regression Population data are unlikely to be exactly on a
– a dependent relationship between two variables straight line
– X – Independent variable/ “predictor” or “regressor”
variable Y may be said to be related to X by
– Y – Dependent variable/ “response” or “criterion”
variable
𝑌𝑖 = 𝛼 + 𝛽𝑥𝑖 + ϵ𝑖
𝑌 = 𝛼 + 𝛽𝑥

Where 𝛼 and 𝛽 are population parameters ϵ𝑖 − 𝑒𝑟𝑟𝑜𝑟 𝑜𝑟 𝑟𝑒𝑠𝑖𝑑𝑢𝑎𝑙

𝛼 – Y-intercept departure of an actual 𝑌𝑖 from a predicted 𝑌𝑖
𝛽 – slope the sum of ϵ𝑖 is zero.

Steps needed to describe the experimental regression

relationship between Y and X: Scatterplots

Graph the data Test if the best

Find the best fit line explains a
Check for fitting straight significant
apparent linear line portion of the
relationship variability in Y

Fitting the Best Straight Line The Regression Coefficient (𝛽)

• the criterion for the “best fit” line through the • regression coefficient
data utilizes the concept of least squares • slope of the best fit regression line

• (𝑌𝑖 − 𝑌)2 • The best estimate of 𝛽 is

𝑆𝑆𝑋𝑌
𝛽=
𝑆𝑆𝑋

JPQ/BSS/AMAM/AUL 5
Quantitative Genetics Biology 140

The Y Intercept (𝛼) Example

• Fit a linear regression equation for the
following data on the wing lengths of thirteen
sparrows of various ages:
From 𝑌= 𝛼 + 𝛽𝑋, 𝛼 = 𝑌 − 𝛽𝑋

The best estimate of 𝛼 is

𝑎 = 𝑌 − 𝑏𝑋

Statistical Tools for Characterizing

Calculations Continuous Traits
n = 13 𝑌 = 44.4
𝑋 = 130.0 𝑌 = 3.415
𝑋= 10.0 SSXY = 514.80 -
130.0 (44.4) F. Simple Linear Correlation
13
SSX2 = 1562.00
2
= 514.80 – 444.00
130.0
SSX2 =1562.00 −
13 =70.80 – measures the strength of linear relationship
= 1562.00–1300.00
between two variables
= 262.00

𝛽=
𝑆𝑆𝑋𝑌
=
70.80
= 0.270
𝑐𝑚 The simple linear
𝑆𝑆𝑋2 262.00 𝑑𝑎𝑦
regression equation is – neither of the variables is assumed to be
𝑎 = 𝑌 − 𝑏𝑋 =3.415 cm – (
0.270 𝑐𝑚
)(10.0 𝑑𝑎𝑦𝑠) 𝑌 = 0.715 + 270𝑋 functionally dependent upon the other
𝑑𝑎𝑦
= 3,415 cm – 2.700 cm
=0.715 cm

The Pearson Correlation Coefficient (r) The Coefficient of Determination (r2)

• it is the measure of the goodness of fit of a linear

model

• the square of the Pearson correlation r and is

interpreted as the proportion of the total
variation in Y that is accounted for by the
variation in X

• ranges from 0 to 1

JPQ/BSS/AMAM/AUL 6
Quantitative Genetics Biology 140

Example
The Coefficient of Determination (R2)
• Calculation of the simple correlation
coefficient and coefficient of determination.
The data are wing and tail lengths among
• calculated by squaring the correlation birds of a particular species.
coefficient r

• may be described as the amount of variability

in one of the variables accounted for by
correlating that variable with the second
variable

Review of Statistical Tools for

Characterization of Continuous Traits
G. The Normal Distribution
– the most important continuous probability distribution
– also called Gaussian distribution or bell-shaped curve
– many of the continuous variables of interest in biology
form a bell-shaped curve or can be transformed to form
such a curve

Definition
The probability density function for a normal
𝑥−𝜇
2random
1
variable has the form 𝑓 𝑥 = 𝑒 − 2𝜎 where 𝜎 is the
2

𝜎 2𝜋
standard deviation of the random variable and 𝜇 is its mean.

H. Tests Concerning Means

• There are two constants in the equation: , which approximately equals

1
3.14159, making equal 0.39894, and e, the base of the Naperian or
2𝜋
natural logarithms, whose value approximates 2.71828.

• There are two parameters,  and 𝜎 which determine the location and
shape of the distribution, respectively

JPQ/BSS/AMAM/AUL 7
Quantitative Genetics Biology 140

Estimating Heritabilities of
A. Variance Components
Quantitative Traits
The total phenotypic variance of a quantitative trait can be partitioned in the following
manner:

• Heritability
VP = total phenotypic variation of the segregating population
VG = genetic variation
– proportion of population’s phenotypic variation is VE = environmental variation
attributable to genetic factors VGE = variation associated with the genetic and environmental factor interactions

A. Variance Components A. Variance Components

The total phenotypic variance of a quantitative trait can be partitioned in the following The total phenotypic variance of a quantitative trait can be partitioned in the following
manner: manner:

VP = total phenotypic variation of the segregating population VP = total phenotypic variation of the segregating population
VG = genetic variation VG = genetic variation
VE = environmental variation VE = environmental variation
VGE = variation associated with the genetic and environmental factor interactions VGE = variation associated with the genetic and environmental factor interactions

VA = additive genetic variance VA = additive genetic variance

VD = dominance genetic variance VD = dominance genetic variance
VI = interaction genetic variance VI = interaction genetic variance

Two types of heritability

1. Broad-sense heritability (H2)

2. Narrow-sense heritability (h2)

JPQ/BSS/AMAM/AUL 8
Quantitative Genetics Biology 140

B. Genetic Similarity of Relatives

B. Genetic Similarity of Relatives 1. Correlation Analysis
1. Regression Analysis The correlation coefficient (r)
Regression coefficient (b)
– Regression and correlation coefficients are related
– Since daughters receive only a sample half of their genes from each parent,
𝑆
the daughter-dam regression estimates only one-half of the narrow by 𝑏 = 𝑟 ( 𝑦)
heritability of a trait. If the variances in the two populations are equal, then 𝑆𝑥

– The regression of offspring on the average of their parents (midparent) is also

an estimate of heritability

– Full sibs (having the same parents) are expected to share 50 % of their genes
in common; half-sibs share 25 % of their genes, therefore

C. Response to Selection
In evaluating progress through individual selection
Example 1. If all the variation between offspring 3 distinct phenotypic means are important:
and one parent (e.g. their sires) is genetic, then r
should be equal to 0.5; if r = 0.2, then h2 is ?
1. M is the mean phenotype of the entire population in the parental
generation, including both the selected and the nonselected individuals.
Example 2. What is h2
if litter mates are 2. M* is the mean phenotype among those individuals selected as parents
phenotypically correlated for a trait by r = 0.2? (those with a phenotype above the truncation point)
3. M’ is the mean phenotype among the progeny of selected parents
∗
𝑀′ = 𝑀 + ℎ2(𝑀 −𝑀)
Example 3. What is h2 if the correlation coefficient where h2 is the narrow-sense heritability
for half-sibs is 0.08? 𝑀′ − 𝑀
ℎ2 =
𝑀 ∗ −𝑀

Example

The average yearly milk production of a herd of

cows is 18,000 lb. The average milk production
of the individuals selected to be parents of the
next generation is 20,000 lb. The average milk
production of the offspring generation is 18,440
lb. Estimate the heritability of milk production in
this population.

JPQ/BSS/AMAM/AUL 9
Quantitative Genetics Biology 140

Quantitative Trait Loci (QTL)

• Most quantitative traits are not highly
heritable: • recent advances combining molecular genotyping
and statistical tools allow geneticists to make
connection between quantitative traits and the
– High heritability: h2 > 0.5 specific QTLs that control them
– Medium heritability: h2 = 0.2 to 0.5
– Low heritability: h2 < 0.2 • QTL identification is most powerful when
analyzing a population with detailed linkage map,
substantial phenotypic variation and a large
number of individuals

JPQ/BSS/AMAM/AUL 10