Introduction to Quantitative Genetics

Visit the link below to download the full version of this book:
https://cheaptodownload.com/product/introduction-to-quantitative-genetics-full-p
df-download/
Introduction to Quantitative Genetics
Fourth Edition

D. S. Falconer
Form erly with Institute o f Cell, A nim al and Population Biology
University o f E dinburgh

and

Trudy F. C. Mackay
D epartm ent o f G enetics
N orth Carolina State U niversity

LONGMAN
Addison Wesley Longman Limited
Edinburgh Gate, Harlow
Essex CM20 2JE, England
and Associated Companies throughout the world.

© D. S. Falconer 1960, 1981,1989

This edition © Longman Group Ltd 1996

All rights reserved; no part of this publication

may be reproduced, stored in a retrieval
system, or transmitted in any form or by any
means, electronic, mechanical, photocopying,
recording or otherwise, without the prior
written permission of the Publishers.

First published by Oliver & Boyd 1960

Fifth reprint 1972
Reprinted by Longman Group Ltd 1975,1976
Second edition 1981
Reprinted 1982, 1983 (with amendments), 1985 (twice)
Reprinted by Longman Scientific & Technical 1986
Third edition 1989
Reprinted 1990,1993
Fourth edition 1996
Reprinted 1996
Reprinted 1997

British Library Cataloguing in Publication Data

A catalogue entry for this title is available from the British Library.

ISBN 0582-24302-5

Library of Congress Cataloging-in-Publication Data

A catalog entry for this title is available from the Library of Congress.

Set by 8 in 10 on 12pt Times

Produced through Longman Malaysia, PP
 Contents

PREFACE TO THE THIRD EDITION ix

PREFACE TO THE FOURTH EDITION x
ACKNOWLEDGEMENTS xi
INTRODUCTION xiii

1 GENETIC CONSTITUTION OF A POPULATION 1

Frequencies of genes and genotypes 1
Mendelian variation in natural populations Causes o f change
Hardy-Weinberg equilibrium 5
The Hardy-Weinberg law Applications o f the Hardy-Weinberg law
Mating frequencies and another proof o f the Hardy-Weinberg law
Multiple alleles Sex-linked genes More than one locus
Non-random mating 19
Assortative mating
Problems 20

2 CHANGES OF GENE FREQUENCY 23

Migration 23
Mutation 24
Non-recurrent mutation Recurrent mutation
Selection 25
Change o f gene frequency under selection Effectiveness o f selection
Number o f generations required Average fitness and load
Equilibria 34
Balance between mutation and selection Changes o f equilibrium
Selection favouring heterozygotes
Polymorphism 42
Problems 45

3 SMALL POPULATIONS: I. CHANGES OF GENE

FREQUENCY UNDER SIMPLIFIED CONDITIONS 48
The idealized population 49
Sampling 51
Variance o f gene frequency Fixation Genotype frequencies
Inbreeding 57
Inbreeding in the idealized population Variance o f gene frequency
Genotype frequencies
Problems 63
iv Contents

4 SMALL POPULATIONS: II. LESS SIMPLIFIED

CONDITIONS 65
Effective population size 65
Exclusion o f closely related matings Different numbers of males and
females Unequal numbers in successive generations Non-random
distribution o f family size Minimal inbreeding Overlapping generations
Mutation, migration, and selection 72
Non-recurrent neutral mutation Recurrent mutation and migration
Selection
Random drift in natural populations 76
Polymorphism 78
Neutral theory
Problems 81

5 SMALL POPULATIONS: III. PEDIGREED

POPULATIONS AND CLOSE INBREEDING 82
Pedigreed populations 82
The inbreeding coefficient o f an individual Coancestry or kinship
Regular systems of inbreeding 88
Close inbreeding Fixation Repeated backcrosses Crosses and
subsequent generations Mixed inbreeding and crossing Change o f
base: structured population Mutation Selection favouring
heterozygotes
Problems 98

6 CONTINUOUS VARIATION 100

Metric characters 102
Properties of metric characters 104
Problems 106

7 VALUES AND MEANS 108

Population mean 109
Average effect 112
Breeding value 114
Dominance deviation 116
Interaction deviation 119
Problems 120

8 VARIANCE 122
Components of variance 122
Components as proportions o f the total Estimation o f the degree o f
genetic determination, VGIVp
Genetic components of variance 125
Additive and dominance variance Total genetic variance Interaction
variance Variance due to disequilibrium
Correlation and interaction between genotype and environment 131
Correlation Interaction
 Contents v

Environmental variance 134

Multiple measurements: repeatability
Summary of variance partitioning 143
Problems 143

9 RESEMBLANCE BETWEEN RELATIVES 145

Genetic covariance 146
Offspring and one parent Offspring and mid-parent Half sibs Full sibs
Twins General Epistatic interaction
Environmental covariance 155
Phenotypic resemblance 157
Problems 158

10 HERITABILITY 160
Estimation of heritability 163
Offspring-parent regression Sib analysis Intra-sire regression o f
offspring on dam Combined estimates
Twins and human data 171
Assortative mating 174
Precision of estimates and design of experiments 177
Offspring-parent regression Sib analyses Selection o f parents
Problems 181

11 SELECTION: I. THE RESPONSE AND ITS PREDICTION 184

Response to selection 185
Prediction o f response Selection differential and intensity o f selection
Improvement o f response
Measurement of response 194
Variability o f generation means Weighting the selection differential
Realized heritability Maternal effects
Change of gene frequency under artificial selection 199
Effects of selection on variance 201
Problems 204

12 SELECTION: II. THE RESULTS OF EXPERIMENTS 208

Short-term results 208
Repeatability o f response Sampling variance Asymmetry o f response
Long-term results 215
Selection limits Mutation Causes o f selection limits
Number o f loci (effective factors) and standardized effects
Problem 226

13 SELECTION: III. INFORMATION FROM RELATIVES 228

Criteria for selection 229
Simple methods Prediction o f response Combined selection
Relative merits o f the methods
Contents

Index selection 240

Construction o f an index Accuracy Response to selection
Actual achievements
Problems 245

14 INBREEDING AND CROSSBREEDING: I. CHANGES

OF MEAN VALUE 247
Inbreeding depression 247
The effect o f selection
Heterosis 253
Single crosses
Problems 261

15 INBREEDING AND CROSSBREEDING: II. CHANGES

OF VARIANCE 263
Inbreeding 264
Redistribution o f genetic variance Environmental variance Uniformity
o f inbred strains
Mutation 269
Subline divergence
Crossing 272
Variance between crosses Combining ability
Problems 279

16 INBREEDING AND CROSSBREEDING: III. APPLICATIONS 281

Selection fo r combining ability Three-way and four-way crosses;
backcrosses Reciprocal recurrent selection Overdominance
Naturally self-fertilizing plants
Problems 288

17 SCALE 290
Distribution and variance Interactions Conclusions
Problems 297

18 THRESHOLD CHARACTERS 299

Liability and threshold Two classes, one threshold Adequacy o f the
liability model Scale relationships Three classes, two thresholds
Selection fo r threshold characters
Problems 310

19 CORRELATED CHARACTERS 312

Genetic and environmental correlations 312
Estimation o f the genetic correlation
Correlated response to selection 317
Indirect selection
Genotype-environment interaction 321
 Contents

Index selection 325

Construction o f the index Response Effect o f selection on genetic
correlations
Problems 332

20 METRIC CHARACTERS UNDER NATURAL SELECTION 335

Natural selection 335
Fitness and its components
Relationships between metric characters and fitness 337
‘Fitness profiles'
Responses to natural selection 339
Fitness Correlated responses Strength o f selection
Equilibrium populations 342
Fitness Major components Characters with intermediate optima
Characters with minimum fitness o f intermediates Neutral characters
Origin of variation by mutation 348
Mutational variance
Maintenance of genetic variation 351
Balance between neutral mutation and random drift
Mutation-selection balance
Problems 354

21 QUANTITATIVE TRAIT LOCI 356

Major genes 356
Methods o f detection
Methods for mapping QTLs 359
Marker loci QTL genotypes Single marker analysis Interval
mapping analysis
Genetical and statistical considerations 366
Experimental design Multiple tests Maximum likelihood
estimation Multiple QTLs
Experimental results 370
Number o f loci Gene effects Consistency
From QTL to gene 375
Problem 377

APPENDIX TABLES 379

GLOSSARY OF SYMBOLS 381
Equivalence of symbols used by Mather and Jinks 383
SOLUTIONS OF PROBLEMS 385
REFERENCES 437
INDEX 459
 Preface to the third edition

This book was written with the intention of providing an introductory textbook,
with the emphasis on general principles rather than on practical applications. I tried
to make the book useful to as wide a range of readers as possible, particularly
biologists who, like myself, have no more than ordinary mathematical ability. The
mathematics does not go beyond simple algebra; neither calculus nor matrix
methods are used. Some knowledge of statistics, however, is assumed, particularly
of the analysis of variance and of correlation and regression.
The second edition kept the same structure but was somewhat enlarged by the
inclusion of developments in the intervening twenty years, and by more attention
being given to plants. In consequence the book came to contain a good deal more
material than is needed by those for whom the subject is part of a course on general
genetics. The section headings, however, should facilitate the selection of what is
relevant. My main regret then, as it is now, was the impossibility of mentioning
more than a very few of the experimental studies that have illuminated the subject
since the book first appeared.
The revisions made in this new edition are less extensive. The desire not to
increase the length of the book has meant that many of the recent developments are
noted by little more than references to the sources. The demonstration that mutation
is not negligible for quantitative genetics has, however, necessitated more substan­
tial revision of Chapter 12 and to a lesser extent Chapters 15 and 20.
The Problems, which were hitherto published separately, are now put together
with the text, following the chapters to which they refer. They are of varying diffi­
culty and I hope that all students will find some that they can solve immediately
and some also that will tax their ingenuity to the full. Some of the problems are
based on the data and solutions of earlier ones. Students are therefore advised to
keep their workings for later use; this will save the repetition of calculations. I have
based the problems on real data wherever I could, to make them more interesting
and realistic. In consequence, however, the arithmetic seldom works out simply,
and a pocket calculator will be needed for most of them. A few of the problems
have been revised for this edition. The solutions are at the end of the book,
arranged in a different order from the problems so as to avoid the risk of inadver­
tently seeing the solution of the next problem. The solutions are not simply answers
but give fairly full explanations of how the problems are solved.

Acknowledgements It is not exaggeration to say that this book could not originally
have been written without the help of Professor Alan Robertson. My understanding
of the subject grew from my frequent discussions with him. I owe the same debt of
X Preface

gratitude to Professor W. G. Hill for his guidance on the preparation of the second,
and now this, edition. Without his advice many of the revisions could not have
been attempted. Dr R. C. Roberts read the manuscripts of the first and second
editions and his suggestions led to many improvements being made. Dr Paul M.
Sharp checked the solutions of all the problems and made many valuable sugges­
tions. I have had help also from many other colleagues who have advised me on
particular matters. To all of these, and to my wife who helped me in many ways, I
am deeply grateful. The mistakes and misunderstandings that remain are entirely
my own. I should be grateful to be told of these.

Department of Genetics D. S. Falconer

West Mains Road February 1988
Edinburgh, EH9 3JN
Scotland

Preface to the fourth edition

Quantitative genetics is now merging with molecular genetics and this very active
area of the subject needs more consideration than it was given in the previous
edition. Accordingly, a new chapter has been added, on quantitative trait loci
(QTLs) - the location and characterization of the genes causing quantitative varia­
tion. Chapter 20, on natural selection, has been largely rewritten, with fuller
treatment of mutation and the maintenance of genetic variation; we hope these
additions will make the book more useful to students of evolutionary quantitative
genetics. In the earlier chapters, the treatment of polymorphism and of neutral
mutation has been expanded, and some sections in the chapters on inbreeding have
been shortened.
We gratefully acknowledge advice from Dr James D. Fry, Professor W. G. Hill,
Dr Peter D. Keightley, Dr Mark Kirkpatrick and Dr Michael Turelli. We are
indebted also to Dr Richard Lyman for producing Figures 21.3 and 21.4, and to Dr
Hartwig H. Geiger for pointing out an error in equation [15.8], which has now been
corrected. Finally, the first author is most grateful to Professor Hill for the hospital­
ity provided in his laboratory.

D. S, Falconer
T. F. C. Mackay
March 1995
 Acknowledgements

We are grateful to the following for permission to reproduce copyright material:

The American Mathematical Society for Fiq. 4.2 adapted from the article
‘Statistical genetics and evolution’ by S. Wright pp. 223-46 Bulletin o f the
Am erican M athem atical Society (1942) Vol. 48; Cambridge University Press for
Fig. 4.1 adapted from the article ‘The genetic structure of populations’ by S. Wright
pp. 323-54 A nnals o f H um an G enetics (1951) Vol. 15; Professor J.W. Dudley for
Fig. 12.3(a) from his paper ‘76 generations of selection for oil and protein percent­
age in maize’ in P roceedings o f the International Conference on Quantitative
G enetics: August 16-21, 1976 by E. Poliak, O. Kempthome and T. Bailey Jr
© 1977 by Iowa State University Press; the Genetics Society of America for Fig.
15.2 from the article ‘The effect of inbreeding on the variation due to recessive
genes’ by A. Robertson pp. 189-207 G enetics (1952) Vol. 37, Fig. 17.2 adapted
from the article ‘Selection for small'and large body size in the house mouse’ by
J.W. MacArthur pp. 194-209 G enetics (1949) Vol. 34, Fig. 21.3 from the article
‘The isolation of polygenic factors controlling bristle score in D rosophila
m elanogaster. II. Distribution of third chromosome effects with chromosome sec­
tions’ by A.E. Shrimpton and A. Robertson pp. 445-59 G enetics (1988) Vol. 118
and Fig. 21.4 from the article ‘QTL analysis of transgressive segregation in an
interspecific tomato cross’ by M.C. de Vincente and S.D. Tanksley pp. 585-96
G enetics (1993) Vol. 134; the International Union of Biological Sciences for Figs
2.2 and 2.3 from the article ‘Asymmetrical responses in selection experiments’ by
D.S. Falconer pp. 16-41 Sym posium on G enetics o f Population Structure Series B
No. 15; Pergamon Journals Ltd for Fig. 15.1 from the article ‘Variation in the
bristle number of D rosophila m elanogaster ’ by M. Rasmuson A cta Zoolog. No. 33;
Professor A. Robertson and Cold Spring Harbor Laboratory for Fig. 12.3(c) from
the article ‘Selection in animals: synthesis’ pp. 225-9 C old Spring H arbor
Sym posium on Q uantitative Biology Vol. 20; University of Chicago Press for Fig.
2.4 adapted from the article ‘The elimination of an autosomal lethal from an experi­
mental population of D rosophila m elanogaster ’ by B. Wallace pp. 65-6 Am erican
N aturalist No. 97 (1963); B.H. Yoo and Cambridge University Press for Fig.
12.3(b) from the paper ‘Long-term selection for a quantitative character in large
replicate populations of D rosophila m elanogaster ’ pp. 1-17 G enetical Research
No. 35.

While every effort has been made to trace the owners of copyright material, in a few
cases this has proved impossible and we take the opportunity to offer our apologies
to any copyright holders whose rights we may have unwittingly infringed.
 Introduction

Quantitative genetics is concerned with the inheritance of those differences

between individuals that are of degree rather than of kind, quantitative rather than
qualitative. These are the individual differences which, as Darwin wrote, ‘afford
materials for natural selection to act on and accumulate, in the same manner as man
accumulates in any given direction individual differences in his domestic produc­
tions’. An understanding of the inheritance of these differences is thus of
fundamental significance in the study of evolution and in the application of genetics
to animal and plant breeding; and it is from these two fields of enquiry that the sub­
ject has received the chief impetus to its growth.
Virtually every organ and function of any species shows individual differences
of this nature, the difference of size among ourselves or our domestic animals being
an example familiar to all. Individuals form a continuously graded series from one
extreme to the other and do not fall naturally into sharply demarcated types.
Qualitative differences, in contrast, divide individuals into distinct types with little
or no connexion by intermediates. Examples are the differences between blue-eyed
and brown-eyed individuals, between the blood groups, or between normally
coloured and albino individuals. The familiar Mendelian ratios, which display the
mechanism of inheritance, can be seen only when a gene difference at a single
locus gives rise to a readily detectable difference in some such property of the
organism. Quantitative differences, in so far as they are inherited, depend on genes
whose effects are small in relation to the variation arising from other causes.
Furthermore, quantitative differences are usually, though not necessarily always,
influenced by gene differences at many loci. Consequently the individual genes,
whether few or many, cannot be identified by their segregation; the Mendelian
ratios are not displayed, and the methods of Mendelian analysis cannot be applied.
It is, nevertheless, a basic premiss of quantitative genetics that the inheritance of
quantitative differences depends on genes subject to the same laws of transmission
and having the same general properties as the genes whose transmission and
properties are displayed by qualitative differences. Quantitative genetics is there­
fore an extension of Mendelian genetics, resting squarely on Mendelian principles
as its foundation.
The methods of study in quantitative genetics differ from those employed in
Mendelian genetics in two respects. In the first place, since ratios cannot be
observed, single progenies are uninformative, and the unit of study must be
extended to ‘populations’, that is, larger groups of individuals comprising many
progenies. And, in the second place, the nature of the quantitative differences to be
studied requires the measurement, and not just the classification, of the individuals.
Introduction

The extension of Mendelian genetics into quantitative genetics may thus be made
in two stages, the first introducing new concepts connected with the genetic proper­
ties of ‘populations’ and the second introducing concepts connected with the
inheritance of measurements. This is how the subject is presented in this book. In
the first part, which occupies Chapters 1 to 5, the genetic properties of populations
are described by reference to genes causing easily identifiable, and therefore qual­
itative, differences. Quantitative differences are not discussed until the second part,
which starts in Chapter 6. These two parts of the subject are often distinguished by
different names, the first being referred to as ‘population genetics’ and the second
as ‘quantitative genetics’ or ‘biometrical genetics’.
The theoretical basis of quantitative genetics was established round about 1920
by the work of Fisher (1918), Haldane (summarized 1932) and Wright (1921). The
development of the subject over the succeeding years, by these and many other
geneticists and statisticians, has been mainly by elaboration, clarification, and the
filling in of details, so that today we have a substantial body of theory accepted by
the majority as valid.
The theory consists of the deduction of the consequences of Mendelian inherit­
ance when extended to the properties of populations and to the simultaneous
segregation of genes at many loci. The premiss from which the deductions are
made is that the inheritance of quantitative differences is by means of genes, and
that these genes are subject to the Mendelian laws of transmission and may have
any of the properties known from Mendelian genetics. The property of ‘variable
expression’ assumes great importance and might be raised to the status of another
premiss: that the expression of the genotype in the phenotype is modifiable by non-
genetic causes. Other properties whose consequences are taken into account include
dominance, epistasis, pleiotropy, linkage, and mutation. The theory then allows us
to deduce what will be the genetic properties of a population if the genes have the
properties postulated. It allows us also to predict the consequences of any specified
breeding plan, including those of natural selection. It therefore forms the basis for
understanding evolutionary change. The main practical use of the theory is in com­
paring the merits of alternative procedures for animal and plant improvement.
The experimental side of quantitative genetics has three roles, complementary to
the theoretical side. First, experimental study of populations allows us to deduce
the properties of the genes associated with quantitative variation. Second, experi­
mental breeding allows us to test the validity of the theory. And third, there are
some consequences of breeding procedures that cannot be predicted from the
theory, and questions about these can be answered only by experiment. There is
now a large body of experimental data which substantiates the theory in consider­
able detail, showing that the genes concerned with quantitative variation do have
the properties known from Mendelian genetics, and that the outcome of most
breeding procedures can be predicted with some confidence. The aim is to describe
all that is reasonably firmly established and, for the sake of clarity, to simplify as
far as is possible without being misleading. Consequently, the emphasis is on the
theoretical side. Though conclusions will often be drawn directly from experimen­
tal data, the experimental side of the subject is presented chiefly in the form of
examples, chosen with the purpose of illustrating the theoretical conclusions. These
 Introduction

examples, however, cannot always be taken as substantiating the postulates that

underlie the conclusions they illustrate. Too often the results of experiments are
open to more than one interpretation. The experimental work mentioned is only a
very small, and far from random, sample of what has been done. In particular, a
great deal more experimentation has been done with plants and farm animals than
would appear from its representation among the work cited.
No attempt has been made to give exhaustive references to published work in
any part of the subject; or to indicate the origins, or trace the history of the ideas.
To have done this would have required a much longer book, and a considerable
sacrifice of clarity. Most of the material in the book is covered more fully in one or
other of the sources listed below. These sources are not regularly cited in the text.
References are given in the text when any conclusion is stated without full explana­
tion of its derivation. These references are not always to the original papers, but
rather to the more recent papers where the reader will find a convenient point of
entry to the topic under discussion. A selection of the original papers that have
most influenced the development of the subject is reprinted with extensive com­
mentaries by Hill (1984) in the Benchmark Papers in Genetics series (Vol. 15).

Chief sources
(For full bibliographical details see list o f References)

Becker (1984) M anual o f Q uantitative G enetics.

Bulmer (1985) The M athem atical Theory o f Q uantitative G enetics.
Crow (1986) Basic Concepts in P opulation, Q uantitative, and Evolutionary
Genetics.
Crow and Kimura (1970) An Introduction to Population G enetics Theory.
Hartl and C lark (1989) Principles o f Population Genetics.
Hedrick (1985) G enetics o f Populations.
Jacquard (1974) The G enetic Structure o f Populations.
Kempthorne (1957) An Introduction to G enetic Statistics.
Kimura ( 1983) The N eutral Theory o f M olecular Evolution.
Li (1976) First Course in Population Genetics.
M ather and Jinks ( 1977) Introduction to B iom etrical Genetics.
(1982) Biom etrical G enetics.
Mayo ( 1987) The Theory o f Plant Breeding.
Weir ( 1990) G enetic Data Analysis.
W right (1968-78) Evolution a nd the G enetics o f Populations, Vols 1-4.
 1 Genetic constitution of a population

Frequencies of genes and genotypes

To describe the genetic constitution of a group of individuals we should have to

specify their genotypes and say how many of each genotype there were. This would
be a complete description, provided the nature of the phenotypic differences
between the genotypes did not concern us. Suppose for simplicity that we were
concerned with a certain autosomal locus, A, and that two different alleles at this
locus, Aj and A2 were present among the individuals. Then there would be three
possible genotypes, AjA p A jA2, and A2A2. (We are concerned here, as throughout
the book, exclusively with diploid organisms.) The genetic constitution of the
group would be fully described by the proportion, or percentage, of individuals that
belonged to each genotype, or in other words by the frequencies of the three geno­
types among the individuals. These proportions or frequencies are called genotype
frequencies, the frequency of a particular genotype being its proportion or percent­
age among the individuals. If, for example, we found one-quarter of the individuals
in the group to be A jA j, the frequency of this genotype would be 0.25, or 25 per
cent. Naturally, the frequencies of all the genotypes together must add up to unity,
or 100 percent.

Example 1.1
The M-N blood groups in man are determined by two alleles at a locus, and the three
genotypes correspond with the three blood groups, M, MN, and N. The following
figures, taken from the tabulation of Mourant (1954), show the blood group frequen­
cies among Eskimos of East Greenland and among Icelanders as follows:

Number o f
Blood group individuals

M MN N

Frequency, % Greenland 83.5 15.6 0.9 569

Iceland 31.2 51.5 17.3 747

Clearly the two populations differ in these genotype frequencies, the N blood group
being rare in Greenland and relatively common in Iceland. Not only is this locus a
source of variation within each of the two populations, but it is also a source of
genetic difference between the populations.
2 1 Genetic constitution of a population

A population, in the genetic sense, is not just a group of individuals, but a breed­
ing group; and the genetics of a population is concerned not only with the genetic
constitution of the individuals but also with the transmission of the genes from one
generation to the next. In the transmission the genotypes of the parents are broken
down and a new set of genotypes is constituted in the progeny, from the genes
transmitted in the gametes. The genes carried by the population thus have continu­
ity from generation to generation, but the genotypes in which they appear do not.
The genetic constitution of a population, referring to the genes it carries, is
described by the array of gene frequencies', that is, by specification of the alleles
present at every locus and the numbers or proportions of the different alleles at each
locus. If, for example, Aj is an allele at the A locus, then the frequency of Aj
genes, or the gene frequency of A p is the proportion or percentage of all genes at
this locus that are the Aj allele. The frequencies of all the alleles at any one locus
must add up to unity, or 100 per cent.
The gene frequencies at a particular locus among a group of individuals can be
determined from a knowledge of the genotype frequencies. To take a hypothetical
example, suppose there are two alleles, Aj and A2, and we classify 100 individuals
and count the numbers in each genotype as follows:

A jA j a ,a 2 A2A2 Total

Number of individuals 30 60 10 100

60 60 0
Number of genes
(¾ 0 60 20 ■ s} ~

Each individual contains two genes, so we have counted 200 representatives of the
genes at this locus. Each A jA j individual contains two Aj genes and each A jA2
contains one Aj gene. So there are 120 Aj genes in the sample, and 80 A2 genes.
The frequency of A { is therefore 60 per cent or 0.6, and the frequency of A2 is 40
per cent or 0.4. To express the relationship in a more general form, let the frequen­
cies of genes and of genotypes be as follows:

Genes Genotypes

A1 A2 ■ A,A, A,A2 A2A2

Frequencies p q P H Q

so that p + q = 1 and P + H + Q = 1. Since each individual contains two

genes, the frequency of genes is \ {IP + //), and the relationship between
gene frequency and genotype frequency among the individuals counted is as
follows:

p=P+hH
q = Q + hH ..[1.1]
 Frequencies of genes and genotypes

Example 1.2
To illustrate the calculation of gene frequencies from genotype frequencies we may
take the M -N blood group frequencies given in Example 1.1. The M and N blood
groups represent the two homozygous genotypes and the MN group the heterozy­
gote. The frequency of the M gene in Greenland is, from equation [1.1], 0.835 +
i(0.156) = 0.913, and the frequency of the N gene is 0.009 + 4(0.156) = 0.087, the
sum of the frequencies being 1.000 as it should be. Doing the same for the Iceland
sample, we find the following gene frequencies in the two populations, expressed
now as percentages:

Gene
M N
Greenland 91.3 8.7
Iceland 57.0 43.0

Thus the two populations differ in gene frequency as well as in genotype frequen­
cies.

Mendelian variation in natural populations

There are many different levels at which we can observe genetic variation in
natural populations for discrete traits that segregate as Mendelian units. At one end
of the spectrum are visible variants with large effects on the phenotype, such as
plant flower colour, shell colours and patterns in snails, or major mutations such as
dwarfism. Much variation between individuals does not give rise to obvious differ­
ences in phenotype, however. This cryptic variation is revealed by techniques that
study differences in proteins and in the DNA itself. The MN blood group variation
in Examples 1.1 and 1.2 illustrates one kind of cryptic variation, detectable by an
antibody reaction. Protein electrophoresis is a technique that detects differences in
mobility of soluble proteins on a gel, in the presence of an electric field. Variants
detected in this manner are inherited as co-dominant alleles, called allozymes. They
are caused by amino acid substitutions that give rise to a change in the electric
charge of the protein. Protein electrophoresis thus detects about 25% of the amino
acid differences between proteins, since 5 of the 20 amino acids are charged.
Variation in DNA sequences can be detected using restriction enzymes. These
enzymes recognize specific 4- or 6-base DNA sequences and cut the DNA when­
ever these sequences occur. The DNA pieces are then separated by size by
electrophoresis, and visualized by hybridization to a labelled probe DNA by a pro­
cess called Southern blotting. If there is variation in restriction sites between
individuals for the stretch of DNA recognized by the probe, this is revealed on the
blot as a change in size of the restriction fragment, called restriction fragment
length polymorphisms, or RFLPs. Unlike protein electrophoresis, which only
detects changes in functional proteins, RFLP variation can be in non-coding as well
as coding regions of the genome. Although cloned DNA is necessary to detect this
variation, it is not necessary to know the function or chromosomal location of the
cloned probe. Variation in restriction fragment lengths can also be caused by inser­
tions and deletions of DNA sequences between two restriction sites. Large
4 1 Genetic constitution of a population

insertions are usually transposable elements: DNA sequences that are present in
multiple, dispersed copies in the genome and that are able to move from location to
location. Other length variation is caused by variation in numbers of tandemly
repeated DNA sequences at ‘minisatellite’ or ‘microsatellite’ loci. The former, also
called VNTR (for variable number of tandem repeat) loci, consist of repeating units
10-60 base pairs long. Microsatellite (or simple sequence repeat, SSR) loci consist
of shorter repeating units of 1-6 base pairs, such as (CA)n or (AGC)^, where n, the
number of repeat units, is variable. Finally, the ultimate level of resolution of varia­
tion between individuals is to compare their actual DNA sequences obtained by
direct sequencing.
Allelic variation for discrete traits, whether phenotypically visible or cryptic, is
known as polymorphism, about which more will be said in Chapters 2 and 4.
Polymorphic loci give rise to the variation in quantitative characters, which is the
subject of this book.

Causes of change
Several agencies affect gene and genotype frequencies in the process of transmis­
sion of genes from one generation to the next. To understand quantitative genetic
variation fully we need to know how these factors, separately and together, influ­
ence genetic variation in populations over time, and what is their relative
importance as agencies of gene frequency change. These agencies form the chief
subject-matter of the next four chapters, but we may briefly review them here in
order to have some idea of what factors are being left out of consideration in this
chapter. The agencies through which the genetic properties of a population may be
changed are these:

Population size The genes passed from one generation to the next are a sample of
the genes in the parent generation. Therefore the gene frequencies are subject to
sampling variation between successive generations, and the smaller the number of
parents the greater is the sampling variation. The effects of sampling variation will
be considered in Chapters 3-5, and meantime we shall exclude it from the discus­
sion by supposing always that we are dealing with a ‘large population’, which
means simply one in which sampling variation is so small as to be negligible. For
practical purposes a ‘large population’ is one in which the number of adult indivi­
duals is in the hundreds rather than in the tens.

Differences of fertility and viability Though we are not at present concerned with
the phenotypic effects of the genes under discussion, we cannot ignore their effects
on fertility and viability, because these influence the genetic constitution of the suc­
ceeding generation. The different genotypes among the parents may have different
fertilities, and if they do they will contribute unequally to the gametes out of which
the next generation is formed. In this way the gene frequency may be changed in
the transmission. Further, the genotypes among the newly formed zygotes may
have different survival rates, and so the gene frequencies in the new generation may
be changed by the time the individuals are adult and themselves become parents.
These processes are called selection, and will be described in Chapter 2. Meanwhile
we shall suppose they are not operating. Human blood-group genes may be taken
 Hardy-W einberg equilibrium

frequencies show two important features. First the frequency of the heterozygotes
cannot be greater than 50 per cent, and this maximum occurs when the gene fre­
quencies are p = q = 0.5. Second, when the gene frequency of an allele is low, the
rare allele occurs predominantly in heterozygotes and there are very few homozy­
gotes. This has important consequences for the effectiveness of selection, as will be
seen in the next chapter.

Applications of the Hardy-Weinberg law

There are three ways in which the Hardy-Weinberg law is particularly useful,
which will now be illustrated.

Gene frequency of recessive allele At the beginning of the chapter we saw, in

equation [1.1], how the gene frequencies among a group of individuals can be
determined from their genotype frequencies; but for this it was necessary to know
the frequencies of all three genotypes. Consequently, the relationship in equation
[1.1] cannot be applied to the case of a recessive allele, when the heterozygote is
indistinguishable from the dominant homozygote. If the genotypes are in Hardy-
Weinberg proportions, however, we do not need to know the frequencies of all
three genotypes. Let a , for example, be a recessive gene with a frequency of q\ then
the frequency of aa homozygotes is q2, and the gene frequency is the square-root of
the homozygote frequency. Example 1.3 illustrates the calculation. For this way of
estimating the gene frequency to be a valid one, it is obviously essential that there
should be no selective elimination of homozygotes before they are counted. It
should be noted also that the estimation of gene frequency in this way is rather sen­
sitive to the effects of non-random mating.

Example 1.3

Phenylketonuria (PKU) is a human metabolic disease due to a single recessive gene.

Homozygotes can be detected a few days after birth, and selective elimination
before then will be assumed to be negligible. Tests of babies bom in Birmingham,
UK, over a 3-year period detected 5 cases in 55,715 babies (Raine et al., 1972). The
frequency of homozygotes in the sample is 90 X 10-6 or about 1/11,000. The
Hardy-Weinberg frequency of homozygotes is q1, so the gene frequency is q =
V(90 X 10"6) = 9.5 X 1(T3 = 0.0095.
The frequency of heterozygotes in the whole population is 2qr(l — q), and among
normal individuals is 2ql{\ + q). Both work out to be 0.019, approximately. Thus
about 2 per cent of normal people, or 1 in 50, are carriers of PKU. It comes as a sur­
prise to most people to discover how common heterozygotes of a rare recessive
abnormality are. The point has already been noted as a conclusion drawn from Fig. 1.1.

Frequency o f 'carriers' It is often of interest to know the frequency of heterozy­

gotes, or ‘carriers’, of recessive abnormalities, and this can be calculated if the
gene frequency is known. If Hardy-Weinberg equilibrium can be assumed, the
frequency of heterozygotes among all individuals, including homozygotes, is
given by 2q(\ - q ). It is, however, often more relevant to know the frequency
among normal individuals, though this will not be very different if homozygotes
1 Genetic constitution of a population

are rare. The frequency of heterozygotes among normal individuals, denoted by

H \ is the ratio of genotype frequencies Aa/(AA + Aa), where a is the recessive
allele. So, when q is the frequency of a ,

^ / _ ) _ 2q . . . [ 1. 3]
(1 - q)2 + 2q(l - q) 1+ q

Test of Hardy-Weinberg equilibrium If data are available for a locus where all the
genotypes are recognizable, the observed frequencies of the genotypes can be tested
for agreement with a population in Hardy-Weinberg equilibrium. According to the
Hardy-Weinberg law, the genotype frequencies of progeny are determined by the
gene frequency in their parents. If the population is in equilibrium, the gene fre­
quency is the same in parents and progeny, so the gene frequency observed in the
progeny can be used as if it were the parental gene frequency to calculate the geno­
type frequencies expected by the Hardy-Weinberg law. The procedure is illustrated
in Example 1.4.

Example 1.4

The M-N blood group frequencies in Iceland were given in Example 1.1. The
observed numbers in the sample were as in the following table. The gene frequen­
cies in the sample are first calculated from the observed numbers by equation [1.1].
Then the Hardy-Weinberg genotype frequencies p 2, 2pq and q2 are calculated from
the gene frequencies by equation [1.2], and each is multiplied by the total number to
get the numbers expected. For example, the expectation for MM is (0.5696)2 X 747.
Comparing the observed with expected numbers shows a deficiency of both homo­
zygotes and an excess of heterozygotes. The x2 tests how well, or how badly, the
observed numbers agree with the expected. The discrepancy is not significant and
could easily have arisen by chance in the sampling. Note that this x2 has only 1
degree of freedom because the gene frequency has been estimated from the data, so
that the observed and expected numbers must agree in their gene frequencies as well
as in their totals.

Genotypes Gene frequencies

MM MN NN Total M N
Numbers observed 233 385 129 747 0.5696 0.4304
Numbers expected 242.36 366.26 138.38 747
x \ = 1-96 P - 0 .2

The test for agreement with an equilibrium population is a test of whether the
conditions for the production of Hardy-Weinberg genotype frequencies have been
fulfilled. The conclusions that can be drawn from the test, however, are limited.
When good agreement is found, the test gives no reason to doubt the fulfilment of all
the conditions. Tests made with blood-group genes nearly always show very good
agreement, as in Example 1.4. But there is one condition whose non-fulfilment will
not lead to a discrepancy, and that is equal fertility among the parents. The reason