MASTERS IN ANIMAL SCIENCE – BY THESIS
MASSEY UNIVERSITY
NEW ZEALAND

SUPERVISOR (NEW ZEALAND); DR CHRIS ROGERS
C0-SUPERVISOR (SOUTH AFRICA); DR ANTJE HIGGO BARTKOWIAK

TITLE

Exploring the correlation between inbreeding and performance in the Hanoverian Horse.

CONTENTS

Table of Contents
PERSONAL DETAILS.............................................................................................................. 1
TITLE........................................................................................................................................ 1
CONTENTS............................................................................................................................... 1
INTRODUCTION...................................................................................................................... 3
HYPOTHESIS........................................................................................................................... 3
LITERATURE REVIEW.......................................................................................................... 4
Breeding objectives in sport horses..................................................................................... 4
Genetic gain - Generation interval and it’s influence on results........................................ 4
Artificial insemination.......................................................................................................... 4
Breed vs type – the structure of studbooks........................................................................ 4
Breed vs brand...................................................................................................................... 5
Measurement of performance in sport horses................................................................... 5
Direct and indirect selection - field versus central performance test data............................................. 6
Young horse competitions and progeny testing....................................................................................................... 9
WBFSH Ranking............................................................................................................................................................................. 9
Heritability............................................................................................................................ 9
Heritability of individual traits measured at station performance test................................................... 9
Heritability of field/competition performance...................................................................................................... 10
Heritability of individual competition traits........................................................................................................... 10
Heritability vs Genetic correlation................................................................................................................................ 10
Inbreeding: Advantages and disadvantages.................................................................... 11
Inbreeding in relevant horse populations........................................................................ 12
The Thoroughbred and Standardbred Racing Horses...................................................................................... 12
Arab Horses.................................................................................................................................................................................. 13
The European Warmblood.................................................................................................................................................. 13
Wild Horse Populations........................................................................................................................................................ 14
Calculating Inbreeding....................................................................................................... 14
Influence of inbreeding on Breeding Values.................................................................... 14
A relevant threshold of inbreeding depression............................................................... 17
METHODOLOGY and DATA ANALYSIS............................................................................. 18
INBREEDING........................................................................................................................ 18
INBREEDING COEFFICIENT............................................................................................................................................... 18
BlOOD QUOTA OF DOUBLE ANCESTORS.................................................................................................................... 18
PARENTAL INBREEDING COEFFICIENT.................................................................................................................... 18
PERFORMANCE.................................................................................................................... 18
PERFORMANCE vs INBREEDING.......................................................................................... 19
RESTRICTION OF RANGE..................................................................................................... 19


INTRODUCTION

 

The focus of horse breeding in Europe has evolved from that of the draught horse to that of the Sport Horse {Wallins, 1991 #291}.  The modern riding horse has a  different purpose to that of the foundation breed stock and as such required a different phenotype and thus genotype {Langlois, 1983 #47}.  The transformation was assisted by the use of modern animal breeding  theory, namely objective instruments for measuring important traits, and development of indices as criteria for more effective selection of breeding animals {Falk-R0nne, 1992 #87}.  This process has involved utilizing superior stallions, at times more than once in a pedigree to improve phenotypic aspects important to the performance of the breed type. As such, since the earliest of times, horse breeders have advocated the use of inbreeding; breeding with a high proportion of a superior animal’s blood in a pedigree to increase the chance of replicating the positive qualities of that horse {merriam webster, 2010 #285}. But is this correct? Early authors advocate the use {Tesio, 1947 #282} but it seems that the more we know the more the reverse may be true. This appears specifically true for production traits {Ercanbrack, 1991 #33;Miglior, 1995 #126;Thompson, 2000 #211}. Casual observation suggests that researchers prior to the 1950’s tend to support the ideology of inbreeding {Pearson, 1914 #136;Wright, 1920 #290;Pearson, 1933 #25;Fletcher, 1946 #16;Rhoad, 1946 #145}, whereas most researchers after 1950 have generally reported on the negative impacts of inbreeding {Rozhdestvenskaya, 1972 #150;Bohlin, 1975 #65;Azevedo, 1982 #28;Strom, 1982 #159;Langlois, 1983 #121;Cothran, 1984 #1;Duncan, 1984 #12;Lamberson, 1984 #32;Clutton-Brock, 1989 #24;Ercanbrack, 1991 #33;VanRaden, 1992 #229;Caballero, 1996 #71;Klemetsdal, 1996 #113;Klemetsdal, 1996 #114;Smith, 1998 #156;Sevinga, 2004 #154;Cothran, 2005 #17;Van Eldik, 2006 #21}; usually referred to as inbreeding depression {Azevedo, 1982 #28;Bijma, 2000 #43;Brisbane, 1995 #213;Cothran, 2005 #17; Dolvik, 1994 #20;Frankham, 1993 #89}. Inbreeding depression is a term that refers to an increase in undesirable recessive disorders, a loss in genetic variation and a decrease in performance of the animals {Farlex, 2010 #278}.  Despite this, it is a commonly held belief among modern breeders that inbreeding to superior animals will produce a higher chance of breeding superior individuals {Faversham, 1999 #79; Lyons, 2010 #223}. This philosophy is especially prevalent in the practice of breeding racehorses {Thiruvenkadan, 2009 #216}.  

One of the greatest Thoroughbred breeders of all time was the Italian FedericoTesio (1869-1954). Tesio developed a breeding philosophy that he proposed allowed him to win twenty-four Italian Derby races between 1911 and 1966 {Tesio, 1947 #282}. Tesio’s methods of inbreeding are even today seen as a yardstick for breeding and his philosophy’s are still seen in modern breeding {Poto nik, 2010 #29;Strom, 1982 #159;Klemetsdal, 1989 #112;Klemetsdal, 1996 #114;Sevinga, 2004 #154;Radomska, 1984 #152;Rozhdestvenskaya, 1972 #150;Lamberson, 1984 #32;Smith, 1998 #156;Wiener, 1992 #175;Bohlin, 1975 #65;Sørensen, 2005 #208;Van Eldik, 2006 #21; Cassell, 2003 #72; Norberg, 2007 #209} and the Warmblood horse is no exception {Niemann, 2009 #268}.

HYPOTHESIS

 

The aim of this project is to examine the hypothesis that inbreeding in the world population the competitive Sport Horses is associated with higher performance. 

LITERATURE REVIEW

Breeding objectives in sport horses

Definitions of breeding objectives vary largely across organisations {Koenen E, 2001 #263}.
Most breeding organisations for Sport Horses want to improve performance in dressage and show-jumping competition by genetic selection {Koenen, 2004 #36}.

For a performance focused breeding scheme, a clear and well-accepted breeding objective is an important aspect when optimising selection strategies {Dekkers, 1998 #75}.
Breeding objectives between breed societies vary largely and often include many subjectively defined traits {Koenen, 2004 #36}.
Clear definitions of performance test traits are required to understand and improve horse-breeding programs {Posta J, 2009 #248}. An international review of these traits was reported by {Koenen, 2002 #41}.
Results at advanced levels of competition are emphasised in the breeding goal for most sport horse breeds {Koenen, 2004 #36}.

Genetic gain - Generation interval and it’s influence on results

There is a relatively short interval of time to obtain data from performance tests (+/- 3 to 6 years) which promotes the selection of shorter generation intervals, thus creating a quicker Genetic Gain {Stewart, 2010 #199}. In contrast, the use of competition data generally requires a longer generation interval and will encourage a slower genetic gain.
In the Hanoverian Sport Horse population the generation interval has decreased in the stallion side (from 12,4 (1970-1974) to 9,9 years (2000-2004)) and increased in the mare’s side (from 8,6 (1970-1974) to 9,7 years( 2000-2004) {Niemann, 2009 #268}.

Artificial insemination

The advent of artificial insemination and better marketing techniques have resulted in the majority of stallions producing less progeny.  About 50 % of the available Hanoverian approved stallions brought only 4 or less registered progeny in 2006 while in 1986 the median was 12 progeny per stallion {Niemann, 2009 #268}. This could in the long term dramatically increase the degree of inbreeding in the population as this indicates that there will be a smaller pool of stallions used.

Breed vs type – the structure of studbooks

In a study by {Koenen, 2004 #36}, they found that the percentages of mares covered by a foreign stallion for KWPN, DWB, SWB, SF and IHB are, respectively, 31%, 74%, 62%, 6% and 32%.  This shows a strong tendency toward genetic exchange between the societies and dilutes any argument of distinctive “Breed types”.
The Hanoverian has operated an open stud book and as such accepted breeding stallions from a  variety of Breeds and Breed types {Christmann, 1995 #73}.
The distinguishing factor when considering breed type over breed is the speed at which the aims of the society’s breeding goals are realized. A breed society is restricted in that they have to protect the specific genetics of the breed whereas the breed type have the freedom to utilize relevant “outside sires” to realize an immediate improvement of genotype {Wallins, 2010 #262}.

Breed vs brand

With the advent in recent years of the international use of artificial insemnation in Warmblood horses {ASVH, 2007 #63;Christmann, 2010 #247;Cunningham, 1975 #217;Dubois, 2008 #53;Foote, 2004 #86;SLU, 2001 #155`, Niemann`, 2009 `#268`, E. Thore ́n Hellsten E`, 2006 `#250;Schade, 2010 #284}, breed societies need to defend strongly their identity as distinct “breed types” as we are seeing a greater and great cross pollination of genetic material. In the Hanverian population the use of foreign stallions has increased from 20% to 40% from 1990 to 2009 {Schade, 2010 #284}. In another study, {Niemann, 2009 #268} found that 44 outside stallions were used in 1985 and compared to 174 used in 2005.  Traditionally the Hanoverian society has promoted the use of outside stallions specifically for the purpose of refining the more heavier German type of horse. This would however only constitutes 2% of the stallions used {Schade, 2010 #284}.
With the greater intermingling of foreign blood the question begs; is the “brand” of specific warmblood societies simply a marketing tool? It would suggest that to maintain a distinct brand the society needs to protect and strongly promote the specific breed society’s lines with ability to compete at the top level. Leading from this one would question if by restricting breeders to less foreign stallions would this disadvantage breeders in the pursuit of breeding top quality horses. Surely the wider the gene pool of top stallions, the greater chance of increasing genotype and thus phenotype which will lead to a better type of horse able to perform at the top level.

Measurement of performance in sport horses

One of the best performance measures used in horse breeding relates to that of the Thoroughbred. Here, performance is often a combination of measurements that involve the speed of the race, the weight the horse is carrying and the number of lengths ahead or behind the average horse in the race.  The level of the race is assessed at the end of a season by the mean value of participating horses. This is generally expressed by a handicap value utilized in certain races {Gillespie, 1971 #95;Langlois, 1983 #121;Porter, 1971 #144}. It is assumed that 5-6 races are needed to establish a horses correct handicapping {Langlois, 1983 #207}.
In Sport Horses, testing procedures vary depending on the breeding scheme {Ström, 1978 #55; Thorén Hellsten, 2006 #54} but despite this, {Thorén Hellsten, 2006 #54}  found that the results for the major European horse populations were in good agreement. Sport Horse breed organisations generally use three types of tests to record performance data: station performance tests, field performance tests and competition tests {Bruns E, 2001 #220}. Horses are usually selected early in life on their own performance and to a large extent their pedigree for the discipline to which they are trained. Relatively few horses are trained for more than one discipline {Philipsson, 1990 #37} found that 73% of Swedish Warmbloods were trained for Show Jumping, 43% were trained for dressage, 7% for eventing and 1% in driving, thus making 24% trained for more than one discipline. This selection may cause severe biases in studies of genetic parameters based on the data {Philipsson, 1990 #37}. It should be noted that an inherent inaccuracy of such lists is the fact that not always the best horses will make it to top-level competition. A number of factors may influence the results of genetic worth based on competition data. These may include;

  • Many horses are eliminated early in their career due to injury.
  • Talented horses may not have the best riders available to take them to the top.
  • Access to the correct trainer is imperative.

Some stallions, after they prove their worth in a station performance test are then “retired” as breeding stallions.
The choice of annual earnings as criteria to measure horse performances has often been discussed {Hintz, 1980 #271;Klemetsdal, 1990 #272;Langlois, 1980 #120;Langlois, 1984 #273;Langlois, 1996 #274;Tolley, 1985 #275; Ricard, 2001 #147}. {Bruns, 1978 #67} utilized economic weight for dressage and show jumping horses based on auction data. However, in contrast to other livestock production systems, many horse breeders manage horse breeding rather as a hobby activity, with no or very low requirements on profit level {SLU, 2001 #155}. This then would dramatically influence results of these weightings.

Direct and indirect selection - field versus central performance test data

Direct selection is the use of competition data to establish performance levels {Wallin, 2003 #181}, whereas indirect selection utilises such mechanisms as the central performance test and the prediction of performance based on certain conformation, movement and temperatment “scores” assigned to a predetermined selection criteria {Gerber Olsson, 2000 #206;Koenen E, 1995 #256}.
By nature, the basic aims to breed for “Breeding Horses” or “Competition Horses” will cause inaccuracies in the statistical analysis of performance regarding the suitability of these data for estimating genetic parameters and genetic correlations between results in different disciplines {Philipsson, 1990 #37}.

Indirect

In most Warmblood associations, Sport Horse performance estimates are mainly ascertained via a numerical “score” for each pre-determined category during  a centralised performance test {Dorofejew, 1976 #76; Ducro B, 2007 #254; Gerber Olsson, 2000 #206; Langlois, 1978 #119; Miiller, 1979 #130; Neisser, 1976 #131}. }.
The limitation with this method is that the data is often not normally distributed and judges tend not to use the full scale.  An alternative approach has been the linear assessment system utilised by the KWPN.  Within the linear score system the horse are marked on their positive (or negative deviation) from the population mean in an attempt to provide data with a normal distribution {Koenen, 1995 #115} .
Individual performance testing permits measurement of a large number of traits {Brockmann A, 1997 #266; Koenen, 2004 #36} and can last up to 120 days {Bruns, 2001 #220}. Horses are thus trained and judged under uniform conditions to minimize the effects of rider and pre-training {Huizinga, 1991 #105;Bruns, 1998 #70}. In Germany and Holland, stallions are generally selected in special test stations after the measurement of performance {Bruns, 1998 #70;Huizinga, 1990 #197}. These test stations allow for the direct comparison of different stallions {Dubois, 2008 #53}. This is alligned with the expected higher heritabilities from performance testing at stations for both mares and stallions than competition based estimates.{Brockmann, 2000 #184;Schade, 1996 #182;Von Velsen-Zerweck, 1988 #60;Von Velsen-Zerweck, 1997 #183}. {Ducro B, 2007 #254} reasons that  inspections allow more horses to be tested, reduce testing costs, and lead to higher accuracies of estimated breeding values.
Conformation, Movement and Jumping are the three main attributes generally tested in Station and Field performance tests { Dolvik, 1994 #20; Langlois, 1979 #193}. Conformation is heavily weighted; the premise being that good conformation is a precursor for good performance {Wallins, 2010 #262}. Researchers having explored this relationship between conformation and performance have at best described a moderate correlation { Langlois, 1978 #119;Holmstro ̈m, 1990 #194;Langlois, 1978 #192;A ́rnason, 1993 #190;A ́rnason, 1996 #191}. {A ́rnason, 1993 #190} and {Ducro B, 2007 #254} found that the subjective traits of conformation and movement tested in test stations showed favourable and highly favourable correlations with performance in dressage (about 0.68). Both found a moderate correlation to performance in show-jumping (about 0.26). {Koenen E, 1995 #256} found high correlations of walk and trot traits with dressage. More specifically, {Holmstro ̈m, 1990 #194} found a positive genetic correlation of a long sloping shoulder with dressage while {Langlois, 1978 #192} found a favourable correlation between a long, strong croup and showjumping performance.
Contradictorily, {Uphaus, 1994 #198} found negative, close to zero or zero correlations between most gait traits and show jumping performance. {Dusek, 1970 #78} also concluded that movement constitutes no real measurement of performance. {A ́rnason, 1993 #190} found that correctness of walk, although heritable, apparently has no linear relationship to performance in sport. In broader context, {Koenen, 1995 #115} and {Burczyk, 1989 #269} concluded that the information of linear scored conformation scores is of minor importance as far as indirect selection for performance. {Dubois, 2007 #77} also argues that selection based on breeding values that include only performance values measured during a station test, the genetic response decreases by 53.1%.
It seems that the relationship between conformation, movement and performance is at best argueable.

Direct

Competition performance results are widely used in the genetic evaluation of showjumpers {Aldridge, 2000 #189;Reilly, 1998 #276;Tavernier, 1990 #187; Brockmann, 2000 #184}. In France, selection has mainly focused on jumping horses and is based on success in competition tests {Dubois, 2007 #77}. Recently, in some countries, breeders are transferring to the German model of in-station selection tests for stallions {Dubois, 2008 #53}.
{Ducro B, 2007 #254} suggested that in a breeding scheme, direct selection for performance using only highest results in regular competition is not efficient as results in high-level competitions can be recorded only at older ages (greater than 8 years), which further prolongs the horses already long generation interval. Further, heritabilities of high-level competition traits are generally low, ranging from 0.10 to 0.25 {Ricard, 2000 #146}. This is partly due to influence of the rider {Brockmann, 2000 #184}. A low heritability means that breeding values become reliable only if large numbers of offspring are tested.
{Dubois, 2007 #77} argues that competition tests may have inherent inaccuracies due to the fact that only a small number of horses are in fact in competition (i.e. 22.5% of the birth Cohort). He argues that for jumping, whatever the testing scheme for stallions, extensive performance in competition is neccesary {Dubois, 2008 #53}.
{Tavernier, 1990 #44} reasons that the only constant in the results on competition is the actual ranking of the horses compared to others in the competition, owing to the fact that it is extremely hard to quantify level of jumps, faults and course. Utilizing 18,798 French jumping horses in 1978, Tavernier estimated repeatability between performances to be 0.29.  The correlation between breeding values based on rank and those based on annual earnings reached 75%.  
Several authors suggest the standardized use of “earnings” for establishing performance {Descoqs, 1976 #45;Langlois, 1983 #47;Langlois, 1987 #48;Tavernier, 1987 #46} and it has been used extensively for breding value rankings in France and Germany {Langlois, 1980 #49;Meinardus, 1987 #50;Tavernier, 1987 #51}.

Combination

In recent years it has become accepted that results from a combination of data sources such as competition and performance tests increases the accuracy of the resulting selection {Albertsdóttira E, 2007 #252; Koerhuis, 1994 #59;Olsson, 2006 #61;Von Velsen-Zerweck, 1998 #246; Philipsson, 1990 #37;Brockmann, 2000 #184;Bruns, 1998 #70;Schade, 1996 #182}.
Illuding to the positive effect of a combination of performance criteria, {Philipsson, 1990 #37} listed the following points that should be taken into account for selection:

  • The effect of selection for conformation only at an early age vs. a step-wise selection including performance test at a later stage on traits characterizing conformation and competitive ability in dressage and jumping.
  • The relative importance of performance tests of stallions vs. selection based on competition results at advanced levels at a later age.
  • The relative importance of performance tests of mares.
  • The correlated responses in different traits as a result of various definitions of the breeding objectives i.e. single vs. all-purpose horses.
  • The importance of an early performance test for stallions.

{Dubois, 2008 #53} presents a modern model for assessing performance of jumping sport horses that utilizes selection objectives that included three traits: conformation and gaits (weighted 20%), competition jumping (weighted 60%) and a third trait (weighted 20%) such as sperm quality or orthopaedic status. The model then utilizes four phases of a horse’s life: a pedigree analyses, a performance test, selection of males allowed to breed and then performance of offspring.
It must be noted that the disadvantage of considering a multiple trait-breeding objective is that the genetic improvement per trait in absolute biological units can be considerably lower when compared to single trait breeding goals {Koenen, 2004 #36}.
Most evaluations now use a combination of young horse and adult data {Thorén Hellsten, 2009 #200;Janssens, 2010 #202;Luehrs-Behnke, 2002 #201}.

Young horse competitions and progeny testing

Traits recorded at young horse competitions, have moderate heritabilities and high genetic correlations compared to results at regular competitions {Brockmann, 2000 #184;Lührs-Behnke H, 2006 #264; Thorén-Hellsten E, 2006 #265}.
The performance value of progeny in competition are often used to establish a stallions worth {Langlois, 1980 #49; Hugason, 1987 #57;Huizinga, 1990 #56;Ström, 1978 #55;Tavernier, 1994 #58;Ström, 1978 #55}.
Studies in Warmbloods have shown a high genetic correlation (0.84) between observations at station performance testing of stallions and competition results of their offspring {Bruns, 2004 #68;Skoglund, 2001 #195}. In fact, station performance testing of progeny allows for the estimation of breeding values on stallions three to four years earlier than from progeny competition results {Christmann, 1995 #73;Olsson, 2000 #133}.
When assigning breed values it should be noted that most horse breeding is of an assortative nature. {Field, 1976 #88;Langlois, 1980 #120;Minkema, 1976 #127}. Breed values based on progeny performance may not then be an accurate reflection of the animals worth as the progeny is restricted by the quality of the mares the respective sire is put to. Thus two sires of equal real worth may have dramatically different breed values depending on the access of each to quality mares.

WBFSH Ranking

The WBFSH was setup in 1994 and their ranking system is now commonly accepted as the yard-stick for measuring performance of individual sport horses.
For the jumper rankings, horses are awarded points for each internatioal competition based on their results, and the difficulty of the competition. All points earned in a year are added up to determine it’s position. For dressage and Eventing a similar system is employed, though for dressahe only the top eight results are utilized and for Eventing only the top six results {Pedersen, 2011 #303}.
For the purpose of this study as our primary source of performance we will use the ranking as published by the World Breed federation of Sport Horses (WBFSH). In this case the 2010 list being relevant.

 

Heritability

Lonka and Vainikainen in 1946 were reportedly the first authors to publish heritability coefficients for horses {Varo, 1965 #170}.

Heritability of individual traits measured at station performance test

Station performance testing seems to be an accurate measure of a horse’s performance heritability {Christmann, 1995 #73;Bruns, 1998 #70;Gerber Olsson, 2000 #91;Huizinga, 1990 #56;Koerhuis, 1994 #59; Olsson, 2006 #61;Philipsson, 1990 #37;Ström, 1978 #55;Thorén Hellsten, 2006 #54`, Ricard`, 2000 `#146}.  The traits tested in the Stallion Performance Tests have moderate to high heritabilities and high genetic correlations with results in both dressage and show jumping{Olsson, 2006 #241; Ohlsson, 1992 #134}. {Wallin, 2003 #181}  and {Van Veldhuizen, 1997 #196} found a high correlation between gait performance tests and dressage competition and an even higher correlation between jumping traits and competition {Bruns, 1998 #70;Huizinga, 1990 #56}. {Gerber Olsson, 2000 #206} found free jumping results were highly correlated (0.93) to results in jumping under rider.
Selection indices can be ascertained from individual scores by data from performance tests by calculated genetic parameters that show heritability for such traits, {Bade, 1975 #64;Bruns, 1978 #67;Glodek, 1975 #90}.

Heritability of field/competition performance

Dressage has been found to be moderately heritable {Meinardus, 1988 #123;Huizinga, 1989 #104;Koenen, 1995 #115;Stewart, 2010 #199`,Wallin`, 2003 `#181}, whereas jumping performance has been found to have a good heritability {Meinardus, 1988 #186;Huizinga et al., 1991;Wallin, 2003 #181;Koenen, 1995 #115; Luehrs-Behnke, 2002 #221; Ricard, 2000 #146;  Ricard, 2004 #178`, Ricard A`, 2000 `#146}, though jumping under saddle has a lower heritability than free jumping {Olsson, 2000 #133}. In relation more to dressage, {Olsson, 2000 #133} found that gaits generally had high heritability’s.
Researchers found results sometimes conflicting, due to an increase of environmental factors with increasing age {Wallin, 2003 #181;Philipsson, 1975 #140}. This is partly due to influence of the rider {Brockmann, 2000 #184}.

Heritability of individual competition traits

Individual competition traits in seem to have a moderately high heritability {{A ́rnason, 1988 #3;Aldridge, 2000 #189;Brockmann, 2000 #184;Bruns, 1981 #185;Foran, 1994 #188;Huizinga, 1989 #104;Langlois, 1975 #117;Langlois, 1980 #49;Meinardus, 1988 #123;Tavernier, 1990 #44;Tavernier, 1990 #187;Wallin, 2003 #181; Koenen, 1995 #115}.
{Albertsdóttira E, 2007 #252} found that heritabilities for specific Sport competition traits in the Iclandic horses were fairly high ranging from 0.18 to 0.21.
{Foote, 2004 #86} in his study of Bovine breeding, also suggests that when selection is made for one or more trait, the genetic gain for each other trait is reduced. 
In relation to breeding of sport horses, the breeder is often selecting for multiple phenotypic traits in search of a superior performance.

Heritability vs Genetic correlation

Genetic correlation is often confused with heritability. Genetic correlation is the proportion of variance that two traits share due to genetic causes {Merriam-Webster, 2010 #238}. The genetic correlation of traits is independent of their heritability. The Genetic Correlation tells us how much of the genetic influence on two traits is common to both. If the value is above zero it means that the two traits are influenced by common genes. The Genetic Correlation is calculated by dividing the covariance between the additive genetic effects of two traits by the square root of the product of the variances for the additive genetic effects of the two traits {Genetics, 2010 #240}.
Description: ::genetic correlation.pdf
{Ricard, 2000 #146} found a genetic correlation of 0.70 to 0.90 for station performance testing observations and competition data. They then found the genetic correlations with eventing are 0.45 and 0.58 for show jumping and dressage, respectively ({Ricard, 2001 #147}. {Ducro B, 2007 #254} found the genetic correlations of movement traits with dressage to be moderate to strong, but with show-jumping, weak to moderate. On the other hand, they found the genetic correlation of free jumping traits to show-jumping to be moderate to strong, whereas free jumping traits to dressage are weak to moderate.  So what does this show; that movement is correlated to dressage and free jumping correlated to jumping. It has however been observed that many dressage horses do well with Jumping blood in their pedigree {Schade, 2010 #284}. The common consensus is that this gives strength and power from behind resulting in good extension and impulsion from beneath the horse {Wallins, 2010 #262}.

Inbreeding: Advantages and disadvantages

Interest in scientific research relating to inbreeding and performance in horses has been documented from at least the early part of the 20th century {Pearson, 1933 #25}. Many modern horse breeders advocate the use of inbreeding {Azevedo, 1982 #28;Bohlin, 1975 #65;Faversham, 1999 #79;Klemetsdal, 1996 #113;Poto nik, 2010 #29;Radomska, 1984 #152;Rozhdestvenskaya, 1972 #150;VanRaden, 1992 #229}.
The consequences of inbreeding are manifested in terms of inbreeding depression, causing a depression of performance {Pern, 1976 #139}; an increase in undesirable recessive disorders, and a loss in genetic variation {Kearney, 2004 #210}.  {Uimari, 1990 #168;de Boer, 1992 #74} studied inbreeding depression by phenotypic selection, utilizing a stochastic simulation. As pedicted by the earlier studies, they found a regression in performance in relation to the inbreeding coefficient.
{Lande, 1995 #116} suggests that inbreeding depression is mainly due to a limited number of recessive mutations. By implication, selection may then rid the population of some of its mutational load {Falconer, 1989 #224}. In studies specifically done with Norwegian cold-blooded trotters, utilizing a Finite Locus model, it was found that phenotypic selection will affect the gene frequencies and the genetic bias will depend on an unknown genetic model {Groen, 1995 #96}. {Frankham, 1993 #89} also argues that selection is able to counteract a significant part of inbreeding depression over time.  Relatively, {Falconer, 1989 #224} and {Fikse, 1997 #81} both stress that the problem with the estimation of inbreeding depression is the effect of selection. This is specifically relevant to Warmblood analysis as Warmblood breeding heavily utilizes phenotypic selection.
Models utilizing a linear regression of performance on inbreeding (with an additive genetic effect) including and excluding the dominance genetic effect were compared on field data {Miglior, 1995 #126;Misztal, 1997 #129}. An inclination towards increased inbreeding depression was obtained from the simpler of the models in each of the studies. Based on this, {Miglior, 1995 #126} and {Klemetsdal, 1998 #15} recommended that one should therefore utilize a simple model.
Studies on other animal species should be observed as in theory there will be a certain amount of relevant crossover. Since the early 1990’s there has been a dramatic increase in the level of inbreeding in the Holstein Dairy Cow population of the UK {Roughsedge, 1999 #149}, in both production traits {Smith, 1998 #156;Thompson, 2000 #211} and nonproduction traits {Cassell, 2003 #72;Wall, 2003 #171;Kearney, 2004 #210}. This has resulted in a loss in genetic variation {Kearney, 2004 #210}. Numerous other studies in cows have shown an unfavorable association between performance for production traits {Smith, 1998 #156;Thompson, 2000 #211} and non production traits {Cassell, 2003 #72;Wall, 2003 #171} with increasing inbreeding {Kearney, 2004 #210}. {Weigel, 2002 #173} concluded that optimized selection could be used to control the rate of inbreeding in the US dairy cattle populations. Inbreeding depression in sheep has been found related to several traits {Analla, 1998 #35;Ercanbrack, 1991 #33;Lamberson, 1984 #32;Wiener, 1992 #175}.
 Inbreeding depression is not only measured in terms of “competition or production performance”. Many other negative effects from inbreeding have been reported in the literature. {Van Eldik, 2006 #21} reported possible negative effects of inbreeding on semen quality in Shetland pony stallions. {Struik, 2009 #287} found that the Thai Pony has a high inbreeding coefficient and attributes this as the main cause to a result a decrease in ejaculate concentration. {Klemetsdal, 1989 #112} studied the negative effect of inbreeding and its effect on fertility in Norwegian trotters.  Conversely, {Cothran, 1984 #1} found that inbreeding does not appear to be a significant factor influencing reproductive performance of Standardbred horses. The probability of Arthritis in the carpal joints of the Norwegian Trotter was found to be greater in the case of horses with a higher degree on inbreeding {Cothran, 1984 #1;Dolvik, 1994 #20}.
It is recommended to maintain F of at most 0.5 to 1.0% per generation {Bijma, 2000 #43;Norberg, 2007 #39}.

Inbreeding in relevant horse populations

The Thoroughbred and Standardbred Racing Horses

Inbreeding is especially pertinent to the Thoroughbred horse {Cunningham, 1991 #18;Field, 1976 #88;Gillespie, 1971 #95;James, 1990 #219;Langlois, 1980 #120;Pern, 1973 #138;Pern, 1972 #137;Porter, 1971 #144;Thiruvenkadan, 2009 #216}.
Phenotypic racing performance results form the basis for selection of trotters and thoroughbreds without any use of genetic parameters {Àrnason, 1994 #4}. As such it is one of the most inbred breeds in the world concerning base genetic material {Klemetsdal, 1992 #110}. Ten founder females account for 72% of maternal lineages, while one founder stallion is responsible for 95%of paternal lineages {Cunningham, 2001 #14}. Four stallions account for a third of the genes and a total of 21 horses account for 80% of the makeup of the modern population {Cunningham, 1991 #18}. This is therefore an extremely inbred breed and with a closed studbook, this situation can only get worse.
Federico Tesio (1869-1954) is considered in Thoroughbred circles as the father of inbreeding lore. Through his philosophy of inbreeding, he won twenty-four Italian derby races {Tesio, 1947 #282}. His focus was on obtaining the appropriate sex balance in his inbreeding models. Following from Tesio’s work, Leon Rasmussen {Faversham, 1999 #79} devised a system of categorizing inbreeding and coined the phrase the “Rasmussen Factor” that indicated a certain level of inbreeding. He claimed that the Rasmussen Factor (RF) occurred 50% more often among the top class runners than it did in the general population. Later however, Roger Lyons {Lyons, 2010 #223}disputed Rasmussen’s findings, and concluded that in general that higher-class runners descend from a narrower range of breeding stock than the population at large. 
Although Standardbred breeders tend to avoid very close inbreeding, nearly all horses with complete pedigrees are inbred within the first five ancestral generations {MacCluer, 1983 #107}. The frequency of inbred animals in a population of  Norwegian Trotters increased with inclusion of a steadily larger number of ancestral generations {Klemetsdal, 1998 #15;Klemetsdal, 1996 #113} In the former USSR, extensive research has been undertaken to compare the phenotypic performance of inbred animals compared to that of outbred animals in both the Orlov trotter and the Thoroughbred. Reduced performance with inbreeding was seen in most cases. {Afanasjev, 1965 #2;Fomin, 1983 #85;Odnoletkova, 1975 #132;Pern, 1973 #138;Pern, 1972 #137;Rozhdestvenskaya, 1978 #151}. Of recent USSR studies, only one did not report an effect of inbreeding depression on Racing Performance {Rozhdestvenskaya, 1972 #150}.
Interestingly, the Thoroughbred shows a foaling rate of only 50% and their racing performance as, measured by the winning times of horses since 1910, has been static {Hill, 1988 #218}. This means that despite intensive selection for performance, race times have not improved {Cunningham, 1975 #217;Cunningham, 1991 #18;Hill, 1988 #218;Langlois, 1980 #120;Ricard, 2000 #146}, indicating that the limited effective population size and intense selection have led to a plateau in performance, and the apparent heritability of performance is illusory. In comparison the human athlete, where inbreeding is specifically avoided, has maintained a steady increase in speed and constantly we are seeing track records broken.  In the 1900 Olympics in Paris, Frank Jarvis of the USA won the 100 meter sprint in 11 seconds {database, 2010 #296}. Usain Bolt currently holds the world record for 100m at 9.72 seconds {database, 2010 #296}. In the marathon, in the 1900 Olympics, Michel Theato from Luxemburg won the event in 2 hours 59 minutes and 45 seconds {database, 2010 #296}. The current world record for the Marathon is held by Haile Gebrselassie of Ethiopia at 2 hours, 3 minutes and 50 seconds {database, 2010 #296}.

Arab Horses

{Radomska, 1984 #152} reported inbreeding depression as a cause for the demise in performance of Polish Arabian Horses.

The European Warmblood

Inbreeding in Warmbloods is generally utilized to improve one or more specific traits {Wallins, 2010 #262}. However correlations in genetic improvement of the one trait has shown to yield on average only a limited response in the other trait {Koenen, 2004 #36}. 
Since the 70’s, the European Warmblood has attracted a number of research projects that have dealt with various genetic issues concerned with improving the breed types. The mid 70’s  saw a number of studies undertaken involving the genetic parameters of the relationship between inbreeding and Warmblood riding horses {Bohlin, 1975 #65;Philipsson, 1975 #140;Ronningen, 1975 #148;Strom, 1982 #159}. By 1980 it was reported that the Hanoverian population had an inbreeding co-efficient of 1,2% (whole Hanoverian population N=152.717) and in 2006 it was 1,4% (only Hanoverian horses with Hanoverian parents) {Niemann, 2009 #268}.  In 2008, {Hamann, 2008 #22}, analyzed a set of 310,109 Hanoverian Warmbloods and established the mean coefficient of inbreeding was 1.33%, 1.19%, and 1.29% for the reference population (all Hanoverian horses born between 1980 and 1995), stallions, and breeding mares, respectively. In 1998, “Interstallion” was an organization formed by the European Association for Animal Production, the World Breeding Federation of Sport Horses and the International Committee for Animal Recording to help the advancement of the European Sporthorse {Koenen, 2002 #41}. The aims of the Interstallion organisation were to describe breeding values, testing procedures and genetic evaluation methods of Warmblood breeding organizations so that genetic evaluations can be compared across countries {Bruns, 2004 #68}. Regretfully, the Interstallion has since been disbanded.
BLUP (Best Linear Unbiased Predictor) animal model indices for genetic evaluation of Warmblood, based on the quality field test, have been published for conformation, dressage and jumping ability since 1987 {Àrnason, 1994 #4; {Hugason, 1994 #103;A ́rnason, 1988 #3;A ́rnason, 1996 #191;Meinardus, 1987 #50;Tavernier, 1988 #31;Tavernier, 1990 #187;Àrnason, 1991 #5}.

From casual observation, the level of performance of the Sporthorse is increasing and thus it seems that the level of inbreeding within these horses is not yet at a restrictive level. In a recent study, the degree of inbreeding of the Hanoverian Horse was found to be in the range of 1.19% to 1.33% {Hamann, 2008 #22}. This is an acceptable level and would not constitute a threat to impact negatively on performance {Azevedo, 1982 #28}. The Hanoverian and most other Warmblood breed types operate an open studbook system {Christmann, 1995 #73} and by definition would migrate towards an outcrossed animal. As a Phd dissertation, {Niemann, 2009 #268} studied how changes in breeding structure in the last 20 years have affected the breeding structure of the Hanoverian. She found no influence of inbreeding on the breeding values.
{Langlois, 1983 #47}, advocates the use of crossbreeding to allow the exploitation of the effects of complementarity on various breeding stocks. The Hanoverian (as do most other Warmblood types) utilized a variety of breeds to “outcross, and has Anglo Arab, Trakehner and Thoroughbred blood within it’s breeding program {Niemann, 2009 #268}

Wild Horse Populations

It is interesting to note that in wild populations, inbreeding coefficients are generally lower than expected on the hypothesis that mating is random with respect to relatedness {Berger, 1987 #13;Duncan, 1984 #12} showing that nature itself has an inherent restrictive mechanism protecting against the effect of inbreeding depression.

Calculating Inbreeding

Various Algorithms have been developed for methods to select and mate animals to maximise the conditioned response on effective population size or rate of inbreeding {Brisbane, 1995 #213;Caballero, 1996 #71;Grundy, 1998 #98;Woolliams, 1998 #215}.
Geneticists have used several different measures to rank genetic closeness. The most commonly used one is the coefficient of inbreeding (F) defined by Sewall Wright in the early 1920s {Wright, 1920 #290}.
When calculating the inbreeding coefficient of an individual or the average inbreeding coefficient of a population, the completeness of a pedigree have a large effect on the outcome of the calculation {Cothran, 1984 #1} and as such estimates of inbreeding depression are least accurate when based on performance data of animals with incomplete pedigrees {Fikse, 1997 #81}.

Influence of inbreeding on Breeding Values

 

Influence of inbreeding in the Hannoverian horse on breeding values (dressage)
{Niemann, 2009 #268}

Description: ::inbreeding 1a.jpg
KEY:
Description: ::bar.jpg             Inbreeding coefficient (%)
Description: ::line.jpg        Number of animals
Middle in-breeding coefficients and number of animals of the Hanoverian total population within different groups of the integrated Pedigree Breeding VALUEs (Dressage) (starting from birth year 1981)

 Of the animals tested, with respect to dressage horses, the average inbreeding coefficient remained around 1,2% and as such breeding values remained independent of the inbreeding coefficient.
Influence of inbreeding in the Hannoverian horse on the breeding values (jumping)
{Niemann, 2009 #268}:

 

Description: ::inbreeding 2a.jpg
KEY:
Description: ::bar.jpg             Inbreeding coefficient (%)
Description: ::line.jpg        Number of animals

“Middle in-breeding coefficients and number of animals of the Hanoverian total population within different groups of the integrated Pedigree Breeding values (branch) (starting from birth year 1981)”

When analysing jumpers, although the number of animals above a breeding value of 141 was minimal, these animals demonstrated a high level of inbreeding (near 1,6 %). There was another slight peak in the inbreeding of jumpers with horses around the breeding value of 100, though  this could be a simple reflection of an average population.

 

Regression of inbreeding on the integrated breeding values in the Hannoverian horse
(Dressage and Jumping)
{Niemann, 2009 #268}:

Description: ::combined 2c.jpg

 

Involution of in-breeding coefficients (%) on the breeding values Dressage and Jumping (breeding appreciation integrated)

{Niemann, 2009 #268} found no meaningful relationship between breeding values and inbreeding in the Hanoverian Population.

A relevant threshold of inbreeding depression

 

Research indicates a number of horse populations to exist with a high inbreeding coefficient. {Valera, 2005 #301} found F = 8.58 in a population of Andalusians. {Sevinga, 2004 #154} found F=15.7 in Fresian horses. {Castle, 2007 #300} in their study of the Australian Thoroughbred found an F of 14.33. The European Warmblood however, generally operating on an open stud book system, has been found to have a low inbreeding coefficient.  In their study of Hanoverians between 1980 and 2000, {Hamann, 2008 #22} found an inbreeding coefficient of 1,33.

There has been little inbreeding depression recorded in Warmblood Horses.  {Meyer, 2003 #302} in a study of OCD in 100 Southern German Coldblood foals, found that foals had a far greater prevalence of OCD above a 3% inbreeding coefficient threshold. However, it seems that the relevance of such a threshold does not translate to other breeds as in Niemann’s study {Niemann, 2009 #268} of the Hanoverian breed type she found that there was no significant change in OCD levels with relevance to a change in F.

Research indicates that there is no quantitative inbreeding threshold that translates across breeds and this seems especially relevant to a quality such as performance.

For this study the relevant threshold should be revealed through the results.

 

METHODOLOGY and DATA ANALYSIS

The hypothesis will be tested by calculating individual inbreeding coefficients and correlating degree of inbreeding with corresponding levels of performance in the disciplines of Dressage, Jumping and Eventing as observed for the competition records for the 2009/2010 season supplied by the WBFSH.
We will utilize 5 generations in our pedigree analysis as beyond 5 generations the ancestors influence is less than 1,57% of the total genetic influence.

INBREEDING

INBREEDING COEFFICIENT

Various models for calculating the inbreeding coefficient have been presented {Elliott, 2009 #231;Croquet, 2006 #232;Blackwell, 1995 #234;Waples, 1989 #235;Reed, 1991 #236;Kearney, 2004 #210;Meuwissen, 1992 #125;Uimari, 1990 #168;de Boer, 1992 #74;VanRaden, 1992 #229}.
{Falconer, 1989 #224} and {Sevinga, 2004 #154} calculated Individual inbreeding coefficients using an online Pedigree Viewer;  (http://www-personal.une.edu.au/~bkinghor/pedigree.htm).  {Berg, 2003 #42} also created a software package called Inbred to calculate individual PCI and inbreeding coefficients. For simplicity, we will utilize the same or similar inbreeding co-efficient program.
The inbreeding co-efficient does not entirely reflect the inbred nature of many of today’s warmbloods due to a selective breeding approach. The parents can be highly inbreed, yet the competing offspring have an inbreeding co-efficient of zero. To address this I will calculate the blood quota of double or common ancesters in each of the pedigree’s within the first 5 generations. This will be used as a secondary indication of inbreeding.  

BlOOD QUOTA OF DOUBLE ANCESTORS

It is assumed that each parent gives 50% of the genetic material to the offspring. Thus, an ancestor in one generation will give 50% material, in 2 generations will give 25%, three generations 12.5%, 4 generations 6,25%  and so on. For each horse that occurs more than once in a pedigree, by utilizing this methodology, we will calculate the percantage of genetic material reflected in the competing offspring.

PARENTAL INBREEDING COEFFICIENT

A further measure of inbreeding that must be taken into account is the inbreeding co-efficient of the parents. As an example, Moorlands Totilas, the current number one dressage horse in the world has an inbreeding coefficient of 0, and yet both his parents are inbreed ( 3,125% and 1,562% respectively).

PERFORMANCE

A list of current competition standings will be used as it equates to performance according to the published 2010 list from the WBFSH.  However, the WBFSH list is a linear indication of ranking. It does not reflect a true indication of what is suspected to be a skewed relationship of ranking to earnings, where earnings is a function of performance.  As such various measures of performance will need to be utilized to understand the full relationship of Inbreeding to performance.

PERFORMANCE vs INBREEDING

Our aim is to understand the relationship between inbreeding and performance. To this end, the relationship of independent variables to dependent variables will be established and a multiple regression analysis performed to account for the confounding variables.
Independent variables relate to Inbreeding and include:

  • The inbreeding co-efficient
  • The sire and dam inbreeding coefficient
  • Blood quota of double ancestors

Dependant variables relate to performance and include:

  • The wbfsh ranking
  • Year earnings
  • Lifetime earnings

The confounding variables include:

  • age
  • gender,
  • rider status (professional of amateur)
  • country of training
  • breed

A multiple regression analysis will be performed to analyze the influence confounding variables have on the relationship between inbreeding and performance.

RESTRICTION OF RANGE

The limitation of this study is that we are only utilizing the top 100 sport horses of each discipline. This study reflects the relationship only in the top percentage of Sport Horses.