Understanding
`Ani■al Breeding
`
`SECOND EDITION
`
`Richard M. Bourdon
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`Understanding Animal Breeding
`
`Second Edition
`
`Richard M. Bourdon
`Colorado State University
`
`.,,
`
`Prentice Hall
`Upper Saddle River, NJ 07 458
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`Library of Congress Cataloging-in-Publication Data
`
`Bourdon, Richard M.
`Understanding animal breeding / Richard M. Bomdon. - 2nd ed.
`p. cm.
`Includes bibliographical references.
`ISBN 0-13-096449-2
`1. Animal breeding. I. Title.
`SF105 .B67 2000
`636.08'2-dc21
`
`99-34849
`CIP
`
`Publisher: Charles E. Stewart, Jr.
`Associate Editor: Kate Linsner
`Managing Editor: Mary Carnis
`Production Liaison: Eileen O'Sullivan
`Production Editor: Lori Harvey, Carlisle Publishers Services
`Director of Manufacturing & Production: Bruce Johnson
`Production Manager: Marc Bove
`Marketing Manager: Ben Leonard
`Cover Design: Wanda Espaifa
`Formatting/Page make-up: Carlisle Communications, Ltd.
`Printer/Binder: LSC Communications
`Cover Photos:© William Mullins I Alamy (Top), Tyler Olson I Shutterstock (Bottom)
`
`© 2000, 1997 by Prentice-Hall, Inc.
`Upper Saddle Rive1~ New Jersey 07458
`
`All rights reserved. No part of this book may be
`reproduced, in any form or by any means,
`without permission in writing from the publisher.
`
`Printed in the United States of America
`
`ISBN □ -13-096449-2
`
`Prentice-Hall International (UK) Limited, London
`Prentice-Hall of Australia Pty. Limited, Sydney
`Prentice-Hall Canada Inc., Toronto
`Prentice-Hall Hispanoamericana, S.A., Mexico
`Prentice-Hall of India Private Limited, New Delhi
`Prentice-Hall of Japan, Inc., Tokyo
`Prentice-Hall (Singapore), Pte. Ltd.
`Editora Prentice-Hall do Brasil, Ltda., Rio de Janeiro
`
`7 2021
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`Contents
`
`A NIMAL BREEDING FROM THE TOP D OWN
`
`CHAPTER 1
`
`WHAT ls 1HE "BEST" A NIMAL? 3
`
`Traits, Phenotypes and Genotypes 3
`Analyzing the System 5
`Genotype by Environment Interactions 6
`Interactions and Breeding Objectives 6
`Common Misconceptions about Interactions 8
`Other Interactions Involving Genotype 10
`Breeding Objectives and Industry Structure 11
`Traditional Livestock Species 11
`Recreational and Companion Animal Species 12
`Factors that Distort Breeding Objectives 13
`Change Versus Stasis 13
`Exercises 14
`Study Questions 14
`Problems 15
`
`CHAPTER 2
`
`How ARE ANIMAL POPULATIONS IMPROVED? 17
`
`Selection 17
`Phenotypic Selection 19
`Measuring Performance 20
`
`V
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`vi
`
`Contents
`
`Selection Using Information on Relatives 20
`Selection for Simply-Inherited Traits 23
`Between-Breed Selection 24
`Mating and Mating Systems 24
`Mating Systems and Industry Structure 26
`Selection and Mating Together 27
`Exercises 28
`Study Questions 28
`
`I'. I in 1 I
`
`ANIMAL BREEDING FROM THE BOTTOM UP
`
`CHAPTER 3
`
`MENDELIAN INHERITANCE 33
`
`Genes, Chromosomes, and Genotypes 33
`Germ Cells and Their Formation 34
`Formation of the Embryo 37
`The Randomness of Inheritance 39
`Dominance and Epistasis 42
`Complete Dominance 43
`Partial Dominance 45
`No Dominance 47
`Overdominance 47
`Common Misconceptions about Dominance 48
`Epistasis 49
`Sex-Related Inheritance 51
`Sex-Linked Inheritance 52
`Sex-Limited Inheritance 52
`Sex-Influenced Inheritance 52
`Exercises 53
`Study Questions 53
`Problems 54
`
`CHAPTER 4
`
`GENES IN POPULATIONS 56
`
`Gene and Genotypic Frequencies 56
`The Effect of Selection on Gene and Genotypic Frequencies 57
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`Contents
`
`vii
`
`The Effect of Mating Systems on Gene and Genotypic Frequencies 60
`Inbreeding 60
`Outbreeding (Crossbreeding) 61
`Exercises 69
`Study Questions 69
`Problems 70
`
`CHAPTER 5
`
`SIMPLY-INHERITED AND POLYGENIC TRAITS 71
`
`Simply-Inherited and Polygenic Traits 71
`Common Characteristics of Simply-Inherited and Polygenic Traits 73
`Different Breeding Appmaches for Simply-Inherited Versus Polygenic
`Traits 74
`Exercises 76
`Study Questions 76
`
`I' I /{ t I I I
`
`SELECTION
`
`CHAPTER 6
`
`SELECTION FOR SIMPLY-INHERITED TRAITS 79
`
`Simple One-Locus Case 79
`Proving Parental Genotypes-Test Matings 80
`Probabilities of Outcomes of Matings 82
`Factors Influencing the Effectiveness of Selection 91
`Exercises 99
`Study Questions 99
`Problems 100
`
`CHAPTER 7
`
`THE GENETIC MODEL FOR QUANTITATIVE TRAITS 102
`
`The Basic Model 102
`Breeding Value 105
`Progeny Difference 107
`Additive Properties of Breeding Values and Progeny Differences 108
`Breeding Value and Selection 110
`Gene Combination Value 110
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`viii
`
`Contents
`
`A New Model 112
`Producing Ability 114
`Permanent and Temporary Environmental Effects 115
`The Genetic Model for Repeated Traits 115
`The Importance of Producing Ability 118
`The Genetic Model and Threshold Traits 118
`Exercises 119
`Study Questions 119
`Problems 120
`
`CHAPTER 8
`
`STATISTICS AND THEIR APPLICATION
`TO QUANTITATIVE TRAITS 123
`Individual Values and Population Measures 123
`The Normal Distribution 124
`TheMean 127
`Variation 129
`The Importance of Variation 129
`Measures of Variation 131
`Variation and the Normal Distribution 132
`Covariation 137
`The Importance of Covariation 139
`Measures of Covariation 140
`Prediction 154
`True and Predicted Values 154
`Prediction Equations 154
`A Summary Example 156
`Exercises 159
`Study Questions 159
`Problems 160
`
`GHAPTER 9
`
`HERITABILITY AND REPEATABILITY 161
`
`Heritability 161
`Common Misconceptions about Heritability 165
`Heritability and Resemblance Among Relatives 166
`The Importance of Heritability 169
`Repeatability 172
`Common Misconceptions about Repeatability 175
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`Contents
`
`ix
`
`The Importance of Repeatability 177
`Ways to Improve Heritability and Repeatability 179
`Environmental Uniformity 180
`Accurate Measurement 181
`Mathematical Adjustments for Known Environmental Effects 182
`Contemporary Groups 186
`Exercises 194
`Study Questions 194
`Problems 195
`
`CH APTER O -
`
`.,
`FACTORS AFFECTING THE RATE OF GENETIC CHANGE 198
`
`. :,
`
`Elements of the Key Equation for Genetic Change 198
`Accuracy of Selection 199
`Selection Intensity 199
`Genetic Variation 199
`Generation Interval 200
`The Key Equation in More Precise Terms 200
`The Rate of Genetic Change 201
`Accuracy of Selection 201
`Selection Intensity 201
`Genetic Variation 206
`Generation Interval 208
`Realized Response to Selection 208
`The Key Equation with Phenotypic Selection 209
`Partitioning the Key Equation 211
`Trade-Offs Among Elements of the Key Equation 213
`Accuracy Versus Generation Interval 213
`Accuracy Versus Intensity 213
`Intensity Versus Generation Interval 214
`Intensity Versus Risk 216
`Comparing Selection Strategies Using the Key Equation: An Example 216
`The Key Equation in Perspective 220
`Male versus Female Selection 221
`Genetic Change when Sires are Purchased 222
`Exercises 223
`Study Questions 223
`Problems· 224
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`x
`
`Contents
`
`CHA PT ER 11
`
`GENETIC PREDICTION 227
`
`Comparing Animals Using Data from Genetically Similar Groups-The Selection
`Index 227
`Prediction Using Regression: A Review 229
`Prediction Using a Single Source of Information 230
`Prediction Using Multiple Sources of Information 242
`Comparing Animals Using Data from Genetically Diverse Groups-Best Li.near
`Unbiased Prediction 245
`Types of BLUP Models 246
`Capabilities of Advanced BLUP Models 246
`Exercises 254
`Study Questions 254
`Problems 256
`
`CHAPTER 12
`
`LARGE-SCALE GENETIC EVALUATION 258
`
`A History of Across-Herd and Across-Flock Comparisons 259
`Sire Summaries 260
`Predictions 261
`Accuracy Measures 262
`Importance of Sire Summaries 264
`Genetic Evaluation for Nonsires 265
`Types of EPDs 265
`Interpreting Genetic Evaluation Information 266
`Predictions Are for Comparing Animals 266
`The Meaning of Zero 266
`Using Accuracy Measures 267
`Evaluation or Characterization? 269
`Pitfalls in Large-Scale Genetic Evaluation 269
`Faulty Data 269
`Lack of Relationship Among Contemporary Groups 270
`Genotype by Environment Interaction 271
`Alternatives to Large-Scale Genetic Evaluation 271
`Nucleus Breeding Schemes 271
`Exercises 273
`Study Questions 273
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`CHAPTE'R 13
`
`CORRELATED RESPONSE TO SELECTION 275
`
`Contents
`
`xi
`
`What Causes Correlated Response? 275
`Genetic, Phenotypic, and Environmental Correlations 276
`Keeping Genetic, Phenotypic, and Environmental Correlations Straight 279
`More Perspective on Correlations between Traits 282
`Factors Affecting Correlated Response 282
`Correlated Response to Phenotypic Selection 285
`Selecting for Correlated Traits 287
`The Good News/Bad News about Genetic Correlations and Correlated
`Response 290
`Exercises 291
`Study Questions 291
`Problems 292
`
`CHAPTER 14
`
`MULTIPLE-TRAIT SELECTION 294
`
`Methods of Multiple-Trait Selection 294
`Tandem Selection 295
`Independent Culling Levels 295
`Economic Selection Indexes 297
`Combination Methods 304
`Selection Intensity and Multiple-Trait Selection 305
`Loss of Selection Intensity and Correlations between Traits 307
`Loss of Selection Intensity in Perspective 308
`Exercises 309
`Study Questions 309
`Problems 310
`
`I' \ I~ I I \I
`
`MATING SYSTEMS
`
`CHAPTER 15
`
`SELECTION FOR SIMPLY-INHERITED TRAITS 315
`
`Mating to Produce Particular Gene Combinations 315
`Repeated Backcrossing to Import an Allele 319
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`xii
`
`Contents
`
`Exercises 322
`Study Questions 322
`Problems 323
`
`CHAPTER 16
`
`MATING STRATEGIES BASED ON ANIMAL PERFORMANCE: RANooM
`AND ASSORTATIVE MATING 325
`
`Strategies for Making Individual Matings 325
`Random Mating 325
`Assortative Mating 326
`Combination Strategies 329
`Strategies for Crossing Breeds or Lines 330
`Exercises 331
`Study Questions 331
`
`CHAP1 £R 17
`
`MATING STRATEGIES BASED ON PEDIGREE RELATIONSHIP:
`INBREEDING AND OUTBREEDING 333
`
`Inbreeding 333
`Effects of Inbreeding 333
`Measuring Inbreeding and Relationship 341
`Calculating Inbreeding and Relationship Coefficients 346
`Linebreeding 359
`Reasons to Inbreed 359
`Inbreeding and Industry Structure 361
`Outbreeding 362
`Effects of Out breeding 362
`Reasons to Outbreed 365
`Outbreeding and Industry Structure 367
`Exercises 367
`Study Questions 367
`Problems 369
`
`CHAPTER18
`
`HYBRID VIGOR 371
`
`A Population Model for Hybrid Vigor 371
`Measuring Hybrid Vigor 374
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`Contents
`
`xiii
`
`Individual, Maternal, and Paternal Hybrid Vigor 377
`Loss of Hybrid Vigor 379
`Predicting Hybrid Vigor 384
`Exercises 390
`Study Questions 390
`Problems 391
`
`CHAPTER 19
`
`CROSSBREEDING SYSTEMS 393
`
`394
`
`Evaluating Crossbreeding Systems
`Merit of Component Breeds 394
`Hybrid Vigor 395
`Breed Complementarity 395
`Consistency of Performance 396
`Replacement Considerations 396
`Simplicity 396
`Accuracy of Genetic Prediction 396
`Rotational Systems 397
`Spatial Rotations Using Purebred Sires 397
`Spatial Rotations Using Crossbred Sires 401
`Rotations in Time 405
`Terminal Sire Systems 406
`Static Terminal Systems 407
`Rotational/Terminal Systems 410
`Composite Animals 411
`Pure Composite Systems 413
`Composite/Terminal Systems 420
`Breeding Composite Seedstock 421
`Breeding for Uniformity 423
`UniformihJ within a Herd or Flock 423
`Uniformity within an Industn; 425
`Exercises 426
`Study Questions 426
`Problems 427
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`/' ,\/{I V
`
`NEW TECHNIQUES, OLD STRATEGIES
`
`CHAPTER 20
`
`BIOTECHNOLOGY AND ANIMAL BREEDING 431
`
`Reproductive Technologies 431
`Artificial Insemination 432
`Embryo Transfer and Related Technologies 432
`Sex Control 434
`Cloning 436
`Same-Sex Mating 438
`Conservation Genetics 439
`Molecular Technologies 440
`DNA Fingerprinting for Animal Identification 440
`Other Applications of DNA Fingerprinting 441
`Marker Assisted Selection for Simply-Inherited Traits 442
`Marker Assisted Selection for Polygenic Traits 445
`Gene Transfer 446
`Exercises 448
`Study Questions 448
`
`CHJ.lPIER 21
`
`COMMONSENSE ANIMAL BREEDING 450
`
`Be Knowledgeable 450
`Use Good Information 451
`Take Time to Think 452
`Be Consistent 452
`Keep It Simple 452
`Be Patient 453
`Exercises 453
`Study Questions 453
`
`GLOSSARY 455
`
`APPENDIX 467
`The Algebra of Variances and Covariances 467
`A Sampling of Proofs Using the Algebra of Variances and Covariances 468
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`Contents
`
`xv
`
`Repeatability as a Ratio of Variances 469
`Predicting Producing Ability from an Individual's Own Records 469
`Heritability as a Ratio of Variances 469
`Heritability as a Regression Coefficient 470
`Predicting Breeding Value from Progeny Records 474
`
`ANSWERS TO ODD-NUMBERED PROBLEMS 479
`
`Answers to Chapter 1 Problems 479
`Answers to Chapter 3 Problems 481
`Answers to Chapter 4 Problems 484
`Answers to Chapter 6 Problems 486
`Answers to Chapter 7 Problems 490
`Answers to Chapter 8 Problems 494
`Answers to Chapter 9 Problems 496
`Answers to Chapter 10 Problems 499
`Answers to Chapter 11 Problems 502
`Answers to Chapter 13 Problems 507
`Answers to Chapter 14 Problems 510
`Answers to Chapter 15 Problems 512
`Answers to Chapter 17 Problems 516
`Answers to Chapter 18 Problems 524
`Answers to Chapter 19 Problems 527
`Quick Key 529
`
`INDEX 531
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`CHAPTER 12
`
`Large-Scale
`Genetic EvaluatiQn
`
`The last chapter was about genetic prediction. It dealt with the methodologies of
`genetic prediction, including the most advanced methodology-best linear unbi(cid:173)
`ased prediction (BLUP). This chapter is also about genetic prediction, but it deals
`with the application of BLUP techniques in developed countries. In other words, it
`deals with large-scale genetic evaluation. From this chapter you can learn what
`a sire summary looks like and where sire summary information comes from, what
`information is available on nonsires, how data from genetic evaluations should be
`interpreted, what pitfalls to be aware of, and what alternatives there are to con(cid:173)
`ventional large-scale evaluation programs.
`Large-scale genetic evaluation refers to the genetic evaluation of large popu(cid:173)
`lations. Typically, these populations are entire breeds within a country or within an
`even larger geographical area. Because the data used for large-scale genetic evalu(cid:173)
`ation come from many breeders and are processed centrally, large-scale genetic
`evaluation is a cooperative effort involving breeders, breed associations, and pro(cid:173)
`fessionals in animal breeding technology.
`The purpose of large-scale genetic evaluation is simple--to allow genetic
`comparison of animals in different herds or flocks. Why is this important? Suppose
`that you own a sire that you think is outstanding. His own performance and prog(cid:173)
`eny records in your herd or flock are excellent, and you are convinced he is the best
`you have ever bred. But without some mechanism for comparing him with sires
`owned by other breeders, you have no objective way of knowing how good he is
`in the breed as a whole. Large-scale genetic evaluation provides that mechanism.
`Large-scale genetic evaluation speeds the rate of genetic change in a popula(cid:173)
`tion. By allowing direct comparison of animals in different herds or flocks, it effec(cid:173)
`tively enables breeders to select individuals from a larger pool of candidates. In(cid:173)
`stead of being limited to the animals they themselves own, breeders can select from
`a much larger population-an entire breed. And just as it is easier to field quality
`athletic teams at a big school than at a small school because the big school has more
`athletes to choose from, so it is easier to find truly outstanding breeding animals in
`a large population than in a small one. In terms of the key equation for genetic
`change, large-scale genetic evaluation allows increased selection intensity.
`Large-scale genetic evaluation also speeds the rate of genetic change by in(cid:173)
`creasing accuracy of prediction. Breed databases contain enormous amounts of in(cid:173)
`formation, many times the amount of information available from any one herd or
`flock. When records from an entire breed are used for prediction, accuracy of pre(cid:173)
`diction increases by virtue of the sheer volume of information available.
`
`Large-Scale
`Genetic
`Evaluation
`The genetic evaluation
`of large populations-(cid:173)
`typically entire breeds.
`
`258
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`A HISTORY OF ACROSS-HERD AND
`ACROSS-FLOCK COMPARISONS
`
`Chap. 12 Large-Scale Genetic Evaluation
`
`259
`
`Central Test
`A test designed to
`compare the
`performance of animals
`(usually young males>
`from different herds or
`flocks for growth rate
`and feed conversion by
`feeding them at a
`central location.
`
`Not countirlg horse races, which have been used to compare arumals from differ(cid:173)
`ent herds for millennia, the earliest across-herd comparisons were probably those
`developed for European dairy cattle near the begirlnirlg of the twentieth century.
`These irlvolved progeny tests for dairy sires. Similar progeny tests were initiated
`in the U.S. dairy industry in the 1930s.
`Some of the first across-herd and across-flock comparisons for swirle, sheep,
`and beef cattle were provided by central tests. In a central test young boars, rams,
`or bulls from different breedirlg operations are brought 'to one location where they
`are fed together for a period of time. They are then compared for daily weight gairl,
`feed conversion (sometimes), and physical measures. Central tests allow breeding
`arumals from different herds or flocks to directly compete agairlst each other, and
`have historically been used as a marketing tool for seedstockbreeders and as a fo(cid:173)
`rum for promoting the use of performance information. But the ability of central tests
`to compare the genetic merit of animals is limited. Comparisons are restricted to the
`few traits measured at the test station and are based on individuals' own perform(cid:173)
`ance data only. And pretest environment can affect performance irl the test, givirlg an
`advantage to animals from some locations and a disadvantage to animals from other
`locations. Smee the advent of BLUP-with its ability to account for genetic and en(cid:173)
`vironmental differences among contemporary groups-and large-scale genetic eval(cid:173)
`uation using BLUP, central tests have lost much of their genetic justification.
`The first true sire evaluations for beef cattle appeared irl the 1970s. With the
`advent of high-speed computers, analyses usirlg BLUP and BLUP-like procedures
`became the norm for cattle breeds in the 1980s. Change is occurrirlg rapidly, but a t
`this writirlg (1995), large-scale genetic evaluation is common in irldustrialized
`
`FIGURE 12.1 Rams on central test.
`
`lologies of
`near unbi(cid:173)
`ut it deals
`r words, it
`earn what
`rom, what
`should be
`ire to con-
`
`.rge popu(cid:173)
`withinan
`~tic evalu(cid:173)
`le genetic
`, and pro-
`
`w genetic
`?Suppose
`andprog(cid:173)
`is the best
`with sires
`~ood he is
`chanism.
`a popula(cid:173)
`:s, it effec(cid:173)
`ida tes. In(cid:173)
`electfrom
`ld quality
`has more
`mimalsin
`)r genetic
`y.
`1ge by in(cid:173)
`mts of in-
`1e herd or
`cy of pre-
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`countries for beef and dairy breeds, and programs for breeds of sheep and swine
`are in the development stages or beyond. Evaluations of performance traits in
`horses are also available--more technically sophisticated ones in Europe, less so(cid:173)
`phisticated ones in the United States.
`The earliest evaluations were designed tests, carefully monitored progeny
`tests free of sources of bias like nonrandom mating and culling for poor perform(cid:173)
`ance. Designed tests were expensive and necessarily limited in size. With the
`power of modern statistical procedures to account for a number of biases, the gen(cid:173)
`eral trend in large-scale evaluation has been toward the use of field data, data that
`are regularly reported by individual breeders to breed associations or government
`agencies. Field data provide huge amounts of information-much more informa(cid:173)
`tion than designed tests can supply.
`Large-scale genetic evaluations using field data require advanced technical
`expertise, complex software, and powerful computers. They are conducted once a
`year or more frequently by specialists at breed associations, in government, at uni(cid:173)
`versities, or (less commonly) in private companies.
`
`The most visible product of large-scale genetic evaluation is the sire summary.
`Sire summaries are lists of genetic predictions, accuracy values, and other useful
`information about sires in a breed. Summaries vary in format from species to
`species and from breed to breed. Typically, however, they are comprised of an in(cid:173)
`troductory section followed by a list of sire data. The introductory section of a sire
`summary is very informative. It may contain an explanation of the data in the list,
`including a glossary of terms, the qualifications necessary for a sire to be listed in
`
`260
`
`Part Ill Selection
`
`Designed Test
`A carefully monitored
`progeny test designed to
`eliminate sources of bias
`like nonrandom mating
`and culling for poor
`performa~
`
`Field Data
`Data that are regularty
` reported by indMdual
`breeders to breed
`associations or
`government agencies.
`
`1,
`
`SIRE SUMMARIES
`
`Sire Summary
`A 11st of genetic
`predidions, acwracy
`values, and other useful
`information about sires
`in a breed.
`
`98
`ElBVIEH
`RE
`Mt1,V,R'<
`
`FIGURE 12.2 Beef sire summaries.
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`:> and swine
`1ce traits in
`)pe, less 80_
`
`·ed progeny
`or perform(cid:173)
`e. With the
`,es, the gen(cid:173)
`:a, data that
`;overnment
`re informa-
`
`d technical
`cted once a
`tent, at uni-
`
`summary.
`ther useful
`species to
`d of an in(cid:173)
`m of a sire
`in the list,
`>e listed in
`
`Chap. 12 Large-Scale Genetic Evaluation
`
`261
`
`the summary, a table of genetic parameter estimates (heritabilities and correlations
`used in calculating predictions), distributions of predictions within the breed (in
`the form of graphs that usually look like normal distributions and(or) tables of per(cid:173)
`centiles), a table converting accuracies to confidence ranges or possible change val(cid:173)
`ues (defined later in this chapter), and graphs of genetic trend.
`The sire list itself typically includes three types of data: animal identification;
`miscellaneous information about the sire including simply-inherited characteris(cid:173)
`tics such as coat color, genetic defects, etc.; and predictions and accuracy measures.
`A fabricated sample segment of a beef sire list is shown in Table 12.1.
`In the example shown in the table, bulls are identified by name, name of sire
`and maternal grandsire, registration number, and owher(s). Other items of in(cid:173)
`formation include the animal's birth date and a code or suffix indicating simply(cid:173)
`inherited characteristics. The names of the second, third, and fourth bulls in the
`list are followed by the suffix "R" connoting red coat color. The trait leader column
`indicates whether the bull ranks particularly high in the breed for a specific trait.
`RCN Crescendo 538 has an especially high EPD for yearling weight and is a trait
`leader for that trait. RCN Prelude 732 has an especially low EPD for birth weight
`and is a trait leader in that category. An actual beef sire summary w6uld present
`considerably more information-including more EPDs-than is shown in this sim(cid:173)
`plified example. Typical dairy summaries present much more information.
`
`Predictions
`
`The genetic predictions published in modem, state-of-the-art sire summaries are ex(cid:173)
`pected progeny differences or EPDs. Dairy summaries use different terms: predicted
`differences (PDs) and estimated transmitting abilities (ETAs), but EPD, PD, and ETA
`all mean the same thing. (I use EPD in this book.) Expected progeny differences (as
`
`TABLE 12,1 Sample Segment of a Beef Sire List
`
`Name (Suffix)
`
`RAB George Washington
`Sire: RAB Sam Adams
`MGS: PBC 737 D2020
`
`RCN Crescendo 538 (R)
`Sire: RCN Intonation 338
`MGS: RCN Sonata 008
`
`RCN Intonation 338 (R)
`Sire: Copper Kettle
`MGS: Gini's Chief 105
`
`RCN Prelude 732 (R)
`Sire: RCN Ensemble 614
`MGS: Copper Kettle
`
`RD Madison Ave 6X
`Sire: PCH Sun Valley 1141
`MGS: BJR Fireworks 416
`
`Reg.#
`
`Birth
`Date
`
`Owner, State
`or Province
`
`129755 3/16/81 N. Maclean, MT
`
`181650 3/11/85 T. Morrison, NY
`
`y
`
`153082
`
`3/9/83 W. Stegner, VT
`
`274698 3/11/87 T. Williams, UT
`M. Golden, CO
`
`B
`
`-5.6
`.87
`
`329877
`
`3/8/90
`
`J. Salinger, NH
`E. Hemingway, ID
`
`1.1
`.67
`
`Tot.
`Birth Wean. ~ Milk
`EPD Mat.
`EPD
`EPD
`EPD
`Trt.
`ACC ACC EPD
`Ldr. ACC ACC
`
`8.4
`.93
`
`0.9
`.87
`
`1.3
`.86
`
`31
`.93
`
`39
`.86
`
`24
`.85
`
`5
`.86
`
`39
`.63
`
`53
`.92
`
`71
`.85
`
`43
`.84
`
`10
`.85
`
`54
`.51
`
`16
`.91
`
`-12
`.80
`
`11
`.81
`
`20
`.78
`
`8
`.40
`
`31
`
`7
`
`23
`
`22
`
`27
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`262
`
`Part III Selection
`
`Accuracy Measures
`
`Repeatability
`On dairy publications):
`Accuracy of prediction.
`
`Confidence Range
`A range of values within
`which we expect-with a
`given probability, a
`given degl'ff of
`confidence-that a true
`value of intemt lies.
`
`opposed to estimated breeding values) are listed because they are comparatively
`easy to interpret. They represent, in relative terms, the expected performance of a
`sire's progeny.
`The sire list shown in Table 12.1 provides examples of EPDs for different
`components of traits: direct, maternal, and total maternal. The EPDs for birth
`weaning, and yearling weight are all measures of direct components. They predic;
`progeny performance that is attributable to genes inherited from the sire. Birth,
`weaning, and yearling weight EPDs predict the growth potential of the sire's calves
`at different ages. The milk EPD is a measure of the maternal component of wean(cid:173)
`ing weight. It predicts the milking and mothering ability (measured in pounds of
`weaning weight) of the sire's daughters. The total maternal EPD measures the
`combination of direct and maternal components of weaning weight known as to(cid:173)
`tal maternal weaning weight. It predicts the overall ability of the sire's daughters
`to produce calf weaning weight.
`
`Accuracy is defined as a measure of the strength of the relationship between true
`values and their predictions and, in mathematical terms, is the correlation between
`true values and their predictions. We might call accuracy defined in this way clas(cid:173)
`sical accuracy. (The term for accuracy that commonly appears in dairy publications
`is repeatability-not to be confused with the concept of repeatability described in
`Chapter 9.)
`In sire summaries, classical accuracy would be the correlation between true
`progeny differences and EPDs (rpo,Pb)- The accuracy values published in sire sum(cid:173)
`maries rarely represent this correlation, however. Instead they are Junctions of clas(cid:173)
`sical accuracies. The reason for this is that classical accuracies approach 1.0 rather
`easily (i.e., with relatively little information) and they do a poor job of differentiat(cid:173)
`ing between sires that have a great deal of data and those that have just a moder(cid:173)
`ate amount.
`Listed in Table 12.2 are the EPDs and published accuracies of the same sires
`cataloged in Table 12.1. Also shown are several alternative accuracy measures in(cid:173)
`cluding classical accuracies. Note how high the classical accuracies are for the first
`four sires in the list. They are so high, in fact, that it is hard to see much difference
`in accuracy between the first bull, whose EPDs are derived from a truly large vol(cid:173)
`ume of data, and the next three bulls, whose EPDs are derived from moderate
`amounts of data. The published accuracies are much better in this respect. Like
`classical accuracies, they range from zero to one, so you can interpret them in much
`the same way as you would classical accuracies.
`Another way of expressing accuracy is with confidence ranges. As explained
`in Chapter 11, a confidence range is a range of values within which we expect-with
`a given probability, a given degree of confidence--that a true value of interest lies.
`Sixty-eight percent confidence ranges for the yearling weight EPDs of two of the sires
`from Table 12.2 are illustrated in Figure 12.3. RAB George Washington is well evalu·
`ated, having a published accuracy for yearling weight EPD of .92. The 68% confi(cid:173)
`dence range for this EPD is narrow-from 51.3 lb to 54.7 lb. In other words, the
`chance of his true progeny difference being between 51.3 and 54.7 lb is a little better
`than two out of three. RAB George Washington's true progeny difference might be
`outside this range, but even if it is, it is unlikely to be far outside it. RD Madison Ave
`
`-
`
`--
`
`-
`
`-
`
`-
`
`-----
`
`'
`
`Exhibit 1030
`Select Sires, et al. v. ABS Global
`
`

`

`nparatively
`mance of a
`
`TABLE 12,2 Alternative Accuracy Measures for the Sires Listed in Table 12.1
`Weaning
`Birth
`
`Sire Name
`
`Yearling
`
`Milk
`
`Chap.12 Large-Scale Genetic Evaluation
`
`263
`
`•
`
`I
`
`I
`
`>r different
`s for birth
`lley predic;
`sire. Birth
`rre s calves
`lt of wean(cid:173)
`pounds of
`iasures the
`own as to(cid:173)
`daughters
`
`tween true
`nbetween
`sway c/as-
`1blications
`~scribed in
`
`ween true
`lSire sum(cid:173)
`ms of clas-
`1.0 rather
`.fferentiat(cid:173)
`: a moder-
`
`;ame sires
`:asures in(cid:173)
`)r the first
`difference
`large vol(cid:173)
`moderate
`pect. Like
`ninmuch
`
`explained
`ect-with
`terest lies.
`>f the sires
`rell evalu(cid:173)
`,8% confi(cid:173)
`rords, the
`ttle better
`might be
`:lisonAve
`
`RAB George
`Washington
`
`RCN Crescendo
`538
`
`RCN Intonation
`338
`
`RCN Prelude
`732
`
`RD Madison Ave
`6X
`
`EPD
`Published accuracy
`Accuracy (rpo,lo)
`68% confidence range
`Possible change
`EPD
`Published accuracy
`Accuracy (rpo,p·o)
`68% confidence range
`Possible change
`EPD
`Published accuracy
`Accuracy (rpo,r"b)
`68% confidence range
`Possible change
`EPD
`Published accuracy
`Accuracy (rpo,P·o)
`68% confidence range
`Possible change
`EPD
`Published accuracy
`Accuracy (rpo,P·o)
`68% confidence range
`Possible change
`
`8.4
`.93
`.998
`8.2 to 8.6
`:t.20
`
`.9
`.87
`.992
`.54 to 1.26
`:!:.36
`
`1.3
`.86
`.990
`.91 to 1.69
`:t.39
`-5.6
`.87
`.992
`-5.96 to -5.24
`:!:.36
`
`1.1
`.67
`.944
`-.77 to 2.97
`:!:l.87
`
`31
`.93
`.998
`30.0 to 32.0
`±1.0
`39
`.86
`.990
`37.0 to 41.0
`±2.0
`
`24
`.85
`.989
`21.9 to 26.1
`±2.1
`
`5
`.86
`.990
`3.0 to 7.0
`±2.0
`
`39
`.63
`.929
`33.8 to 44.2
`±5.2
`
`53
`.92
`.997
`51.3 to 54.7
`:!:1.7
`
`71
`.85
`.989
`67.8 to 74.2
`:!:3.2
`
`43
`.84
`.987
`39.6 to 46.4
`:!:3.4
`
`10
`.85
`.989
`6.8 to 13.2
`:!:3.2
`
`54
`.51
`.872
`43.7 to 64.3
`:!:10.3
`
`16
`.91
`.996
`15.4 to 16.6
`:t.6
`
`-12
`.80
`.980
`-13.4 to -10.6
`±1.4
`
`11
`.81
`.982
`9.6 to 12.4
`±1.4
`'
`· 20
`.78
`.975
`18.4 to 21.6
`±1.6
`
`8
`.40
`.800
`3.7 to 12.3
`:!:4.3
`
`possible change
`lPC) or standard
`error of
`predlc:tion
`A measure of accuracy
`Indicating the pcm!lltial
`amount of future change
`in a prediction.
`
`6X has relatively little information as evidenced by his published yearling
`weight accuracy of .51. The 68% confidence range for his yearling weight EPD
`is much broader-from 43.7 lb to 64.3 lb. Clearly there is more potential error as(cid:173)
`sociated with RD Madison Ave 6X's yearling weight EPD than with RAB George
`Washington's.
`Genetic predictions for individual animals change over time as more and
`more data are included in the calculations used to produce successive evaluations.
`A measure of accuracy indicating the potential amount of future change in a pre(cid:173)
`diction is called possible change (PC). The statistical term for possible change is
`standard error of prediction. A prediction ± possible change is simply an alter(cid:173)
`native way of representing a 68% confidence range. For example, RAB George
`Washington's EPD ± PC for yearling weight is 53 ± 1.7 lb, which corresponds to
`his 68% confidence range of 51.3 to 54.7 lb (see Figure 12.3). Likewise, RD Madison
`Ave 6X's EPD ±