Improvement in the national genetic evaluation of warmblood riding horses by including information from related studbooks

Abstract

Young horse performance test data from two warmblood riding horse populations, Norwegian warmblood (NWB) and Swedish warmblood (SWB), were analysed to examine whether including information from a related studbook would increase the accuracy of the genetic evaluations within a population. Ten conformation and performance traits from 31,588 horses, 774 NWB and 30,814 SWB were analysed separately and jointly using single trait animal models. Heritabilities were moderate to high, and varied from 0.15 (conformation, joint data) to 0.74 (jumping technique, NWB data). The genetic similarity (GS) between populations was 31%, with the SWB, as expected given the size of the populations, contributing most to the GS (98%). Genetic correlations between the same traits in the two populations were 0.43–0.90 but with large standard errors (0.2–0.3). Including information from the other population increased the average accuracy of estimated breeding values for common stallions, on average 4% for SWB and 110% for NWB.     

Modelling repeated competition records in genetic evaluations of Danish sport horses

Genetic evaluations of sport performance typically consider competition records of ranking points in each competition, accumulated lifetime points or annual earnings. Repeated observations have the advantage of allowing for adjustment of effects associated with each competition such as rider experience, judge and competing horses, but also demands more computer capacity than single‐trait records, which could prohibit multiple‐trait evaluations. The aim of the study was to compare CPU times, estimated breeding values (EBVs), reliabilities and model prediction abilities when modelling repeated competition ranking points (run A), mean ranking points (runs B and C), mean ranking points precorrected for effects associated with each competition (run D) and accumulated lifetime points (run E) for Danish Warmblood horses. CPU times for run A were 632–776 times (show jumping) and 59–96 times (dressage) as high as for runs B–E. EBVs of run D were perfectly correlated (1.00) with those of run A. Reliabilities were highest in runs E and A. Best model prediction ability and least bias were found in run C (dressage) and run E (show jumping), but the best choice in each discipline was not preferable for the other. Run D was the second best in both disciplines (D), and is expected to increase in performance over time as omission of a relatively large amount of historic data becomes less important.     

Analysis of competition performance in dressage and show jumping of Dutch Warmblood horses

Most Warmblood horse studbooks aim to improve the performance in dressage and show jumping. The Dutch Royal Warmblood Studbook (KWPN) includes the highest score achieved in competition by a horse to evaluate its genetic ability of performance. However, the records collected during competition are associated with some aspects that might affect the quality of the genetic evaluation based on these records. These aspects include the influence of rider, censoring and preselection of the data. The aim of this study was to quantify the impact of rider effect, censoring and preselection on the genetic analysis of competition data of dressage and show jumping of KWPN. Different models including rider effect were evaluated. To assess the impact of censoring, genetic parameters were estimated in data sets that differed in the degree of censoring. The effect of preselection on variance components was analysed by defining a binary trait (sport‐status) depending on whether the horse has a competition record or not. This trait was included in a bivariate model with the competition trait and used all horses registered by KWPN since 1984. Results showed that performance in competition for dressage and show jumping is a heritable trait (h² ~ 0.11–0.13) and that it is important to account for the effect of rider in the genetic analysis. Censoring had a small effect on the genetic parameter for highest performance achieved by the horse. A moderate heritability obtained for sport‐status indicates that preselection has a genetic basis, but the effect on genetic parameters was relatively small.     

A reaction norm model approach to estimate the genetic effect of temperature on sportive performance of trotter horses

The performance results of Spanish Trotter Horses were merged with the information from the nearest official weather stations, to estimate the effects of temperature (T) on speed during the day of the race. The resulting data set included: 104 125 individual results from 7653 races with 3772 participant horses and driven by 1402 jockeys. After a classical repeatability animal model in which the effects of T were assumed null, a total of five different random regression models (RRM) with the same fixed part were applied in a parsimonious nested process in which the (co)variance components for each random effect were estimated considering homo/heterogeneous responses across the trajectory of T during the race. The results demonstrate that performance is better when T increases from 4°C to 12°C, followed by a comfort zone (12–19°C) and a light but more stressful zone (>20°C). A significant change in the T during 20 years (+4°C) was observed whilst the speed of the race improved by +2.1 m/s during the same period. The information criterion (LogL, AIC and BIC) confirmed the existence of heterogeneity of (co)variance components along the trajectory of T scale. The model in which T effect was assumed non‐constant across the T scale for animal genetic effects; the individual environmental permanent effects and the jockey effects fit the data better. Heritability had higher values (0.19–0.20) when T was <12°C and slightly decreases within the comfort and stressful zones. Genetic correlations were high and, as was expected, it decreases when difference between T increases. The best 150 animals for speed during the comfort zone were selected according to the Expected Breeding Value (EBV) estimated with the best RRM, and important variations were detected in the evolution of EBV across the trajectory of T. The presence of a genetic base related to T effect during the race in this type of performance is demonstrated.     

Multi-trait animal model estimation of genetic parameters for linear type and gait traits in the Belgian warmblood horse

Genetic parameters for the height at withers, 27 linear type and six linear gait traits were estimated for the Belgian warmblood horse. Observations on 987 mares, mostly 3 years old, were analysed using a multi-trait animal model. The statistical model included appraiser, age and location (date × place of appraisal) as fixed effects. Genetic parameters were estimated using a canonical transformation and an expectation-maximization restricted maximum likelihood algorithm with an additional deceleration step. Estimates of heritability for the 33 linear traits were between 0.15 and 0.55. Heritability of the height at withers was 0.34 ± 0.06. Estimated genetic correlations ranged from −0.60 to 0.98 with an average SE of 0.10. The highest positive correlations were found among traits of walk and among traits of trot. Volume and the quality of legs were the most negatively correlated. Estimated genetic parameters indicated that the linear scoring system is a valuable tool to assess conformation. The full (co)variance matrix is now available for breeding value estimation to support selection for conformation and gaits.     

Plasticity effect of rider–horse interaction on genetic evaluations for Show Jumping discipline in sport horses

To obtain a sport horse that excels in the highest levels of competition, breeders must take into account certain genetic and environmental factors that could influence the sport horse's performance, such as the rider–horse interaction (RHI). The main aim of this study was to describe this interaction in a genetic model by modelling it in relation to the horse's age. A total of 31,129 sport results from Spanish Sport Horses were used from a total of 1,101 animals evaluated, and these were grouped in three age levels and had been ridden by 606 different riders. Only riders who had ridden more than one horse (and vice‐versa) were considered for the analyses. Five linear models with different random effects were analysed according to the covariates, the homogeneity/heterogeneity of the RHI and the relevant residual random effects. The model of best fit was then selected for the genetic evaluation of the animal. In general, models including the RHI effect (M2, M4 and M5) fitted better than the other models, and the best fit was obtained for M4 (with heterogeneous residual variance). The genetic variance increased constantly with age, whereas heritability showed a response on three intervals. This study revealed the varied evolution of the RHI with age, showing the different “plastic abilities” of this relationship.     

Genetic connections between dressage and show-jumping horses in Dutch Warmblood horses

Abstract

During the last decades, the breeding practice within the Dutch Warmblood studbook (KWPN) has resulted in an increasing specialisation of horses into show-jumping (JH) and dressage (DH). The objective of this study was to describe the effect of the specialisation on the connectedness between the subpopulations of JH and DH horses registered by KWPN. The subpopulations comprised 23,800 JH horses and 18,125 DH horses, born between 1995 and 2009. Genetic similarity (GS), genetic pool in common (GC_x) based on the marginal genetic contribution of common ancestors and coefficient of relationship (r) between and within subpopulations were analysed in three periods of time to describe changes in genetic connectedness between subpopulations. A decline in GS (0.97–0.45), GC_0.5 (0.69–0.13) and r (0.018–0.014) in the recent years was observed. Both subpopulations have a common genetic pool; however, if the specialisation process continues, it will result in two unrelated (genetic) groups.     

Evaluation of the length of competitive life in Hungarian sport horses

The aim of this study was to analyse the duration of the competitive life of jumping horses in Hungary to assess the potential for genetic evaluation and to get information about the risk factors affecting jumping performance of sport horses. The data set included lifetime performance in jumping competitions for 8475 horses born from 1992 onwards with records between 1996 and 2011 (32.7% were right‐censored). Longevity was measured as the number of years spent in the competition. Discrete time survival model included fixed effects of gender, age at first competition, and the time‐dependent effect of competition level. All fixed effects were highly significant (p < 0.001). The highest risk of ending the race career was estimated for horses competing in the lowest level (category 1) events. The smallest risk was found in horses competing in the highest category. The culling risk for stallions was 15% higher, for geldings 20% lower when compared to mares. As the age at the first competition result increased, the risk ratio also increased. Risk ratios of horses starting as 8 years old or older were 56% higher than those of 5‐year‐old horses. The longest time in sport tracks belonged to horses starting their career as 4 years old. Their culling risk was 20% lower than those of 5‐year‐old horses. The heritability of the length of competitive life was 0.17. Breeding values were predicted for each animals, and 47 showed an accuracy greater than 0.60. The clear benefit of the analysis was shown by comparison of offspring to extreme stallion groups.     

Genetic correlations between competition traits and traits scored at breeding field-tests in Icelandic horses

Relationships between breeding field-test traits and competition traits were studied to investigate whether the latter could be usefully included in the genetic evaluation of Icelandic horses. The current method of genetic evaluation is based on records from breeding field-tests only. The breeding field-test data included 16 401 individual records of Icelandic horses evaluated in 11 countries during 1990–2005. Competition results included 18 982 records of 3790 horses competing in sport and gæðinga competitions in Iceland and Sweden during 1998–2004. In the breeding field-tests, eight conformation traits and eight riding ability traits were scored; height of withers was also recorded. These traits were analysed together with the competition traits tölt(comp), 4-gait, 5-gait and pace test, in bivariate analyses. Animal models were used; the fixed effects for breeding field-test traits included sex by age interaction and country by year interaction. For the competition traits the model included fixed effects of sex, age and event, and a random permanent environmental effect. Estimated heritabilities and genetic correlations for breeding field-test traits were consistent with earlier results; heritabilities ranged from 0.20 to 0.67, and moderate to high genetic correlations were estimated between many of the riding ability traits, and between riding ability traits and some conformation traits. The estimated heritabilities for competition traits were about 0.20, and genetic correlations between competition traits varied from − 0.12 to 1.00. In general, high genetic correlations were estimated between breeding field-test riding ability traits and competition traits. Moderately positive genetic correlations were found between some breeding field-test conformation traits and competition traits. Competition traits add information relating to the breeding goal of the Icelandic horse; they should therefore be added to genetic evaluation in future.     

Effects of integrated genetic evaluations for Icelandic horses on predictive ability, accuracy and selection bias

The genetic evaluation of Icelandic horses is currently based on results from breeding field tests of riding ability and conformation. The effect of integrating competition traits and/or test status into the genetic evaluation was studied concerning estimation bias, predictive ability, accuracy, correlations between breeding values and ranking of sires. Breeding field test data included 19 954 records from horses assessed in 11 countries during 1994-2008. Competition data included 44 160 records from 7687 horses competing in Iceland and Sweden in 1998-2008. Test status was defined as attendance of horses born in Iceland at breeding field tests and/or in competition. Overall, there were trivial differences between different genetic evaluation models regarding estimation bias and predictive ability. Very strong correlations were estimated between breeding values for combined indexes of conformation, riding ability and total score from different models. Higher accuracy was achieved for most of the traits when competition traits and/or test status were added to the model. Sires ranked differently when the new traits were added to the genetic evaluation model. It was concluded that competition traits should be integrated into the genetic evaluation. Further analyses on genetic parameters for test status and its relationship with the other traits are needed for future inclusion of test status in the genetic evaluation.     

Genetic analysis of the longevity of French sport horses in jumping competition

The longevity of sport horses is an economic and ethical issue. The aim of this study was to analyze the duration of the competitive life of jumping horses in France to assess the potential for genetic evaluation and to propose rules of management for sport horses. Data included lifetimes spent in jumping competitions for the 209,296 horses born from 1968 onward with performances between 1972 and 2008; the data set contained 22% right-censored records. Longevity was measured in years. Discrete survival analysis included fixed effects of region of birth, month of birth, year of recording, age at first competition, interaction between sex and level of jumping performance as measured by the logarithm of earnings adjusted for sex, age, year, and random sire and maternal grand-sire effects. There were 16,668 sires and maternal grand-sires. All fixed effects were highly significant (P < 0.001). Management of the sports career had an important effect on longevity: against common belief, the younger the horse started competing, the longer it stayed in competition. For horses that started competing at an age of 6 yr, the risk of culling was 1.33-fold that of horses having started at 4 yr of age. The less success in competition, the greater was the chance for leaving competition, especially for horses without earnings. For a gelding without earnings, the risk of culling was 1.40-fold that of an average-performance gelding and 2.57-fold that of a top-rated gelding (performance at least 2 SD above the mean). Mares always had greater relative risk than geldings or stallions because they may be culled from competition to be used for breeding. The risk of culling for females was 1.45-fold that of a gelding with the same performance. The heritability of the length of competitive life was 0.10. Breeding values were predicted for sires, and 3,303 sires showed an accuracy greater than 0.60. Among these sires, 262 were used for breeding in 2008.     

Modelling genetic evaluation for dressage in Pura Raza Español horses with focus on the rider effect

The most popular use of the Pura Raza Español horse in sport is for dressage competitions. Tests on young sport horses were first established in 2004 in Spain to collect data for the genetic evaluation of this breed's suitability for dressage. The aim of this study was to compare eight different models to find out the most appropriate way to include the rider in the genetic evaluation of dressage. A progressive removal of systematic effects from model was also analysed. A total of 8867 performance records collected between 2004 and 2011 from 1234 horses aged between 4 and 6 years old were used. The final score in the dressage test was used as the performance trait. The pedigree matrix contained 8487 individuals. A BLUP animal model was applied using a Bayesian approach with TM software. The horse's age, gender, travelling time, training level, stud of birth and event were included as systematic effects in all the models. Apart from the animal and residual effects that were present in all models, different models were compared combining random effects such as the rider, match (i.e. rider–horse interaction) and permanent environmental effects. A cross‐validation approach was used to evaluate the models' prediction ability. The best model included the permanent environmental, rider and match random effects. As far as systematic effects are concerned, the event or the stud of birth was essential effects needed to fit the data.     

Genetic analyses of new movement traits using detailed evaluations of warmblood foals and mares

Detailed movement evaluations of warmblood foals and mares were performed in connection with regular breeding events of the Oldenburg horse breeding societies in 2009 and 2010. Unfavourable movement characteristics considered indicative for impaired balance were noted by a special judge (SJ) and the regular judges of the breeding events (RJ) and served as the basis for definition of new movement traits. Detailed movement information on 3374 foals and 2844 mares showed that more severe findings like irregular motion pattern in hind legs or irregularity in general motion pattern occurred only sporadically (prevalences of 1-2%). Irregular tail tone or posture was documented for 4% of the foals and 5% of the mares, resulting in prevalences of the comprehensive trait indications of imbalance (IMB) of 6.2% (foals) and 5.5% (mares). Binary coding was used for all traits, and genetic parameters were estimated bivariately in linear animal models with residual maximum likelihood. Comparative analyses between judges revealed that differences between trait definitions of SJ and RJ were larger in the mares than in the foals, but justified combined use of SJ and RJ information in both age groups. Heritability estimates for the movement traits ranged on the original scale from 0.02 to 0.26 in the foals and from 0.03 to 0.12 in the mares, with heritabilities for IMB on the underlying liability scale of 0.46 (foals) and 0.22 (mares). Comparative analyses between age groups indicated that common genetic factors may be responsible for findings of impaired balance in foals and mares. The results implied that horse breeding may benefit from using the early available information on the movement of foals obtained by detailed movement evaluations, although favourable combination of foal and mare data in future genetic evaluations may require refined recording of unfavourable movement characteristics in the adult horses.     

Genetic connectedness between seven European countries for performance in jumping competitions of warmblood riding horses

The aim of this study was to evaluate the genetic connectedness between seven European populations of show jumping horses. Data on stallions bred for jumping and their pedigree up to three generations were provided by seven countries, including 6317 stallions and 22 324 different horses. Four different methods were used to estimate the genetic connectedness. Common stallions between countries varied from 0 to 308 and common horses from 130 to 1166. “Genetic similarities” were not the optimal measure of connectedness as number of progeny by country by stallion was not balanced. Correlation between estimates of country effects was computed in a mixed model including the genetic value of horses (h² = 0.20, all relationships included). For Germany, Belgium, France and The Netherlands, these correlations were 0.32 to 0.51 corresponding to a balanced scheme with 11 to 19 progeny per stallion per country. Other correlations ranged from 0.08 to 0.27. Another method to evaluate the connectedness was (i) to simulate a systematic difference between the true genetic level of countries, (ii) performing a BLUP and (iii) measuring the percentage of the initial difference that could be found between the estimated genetic levels of the corresponding countries. In 50% of the country pairs, 50% or more of the initial genetic difference was estimated. It was concluded that the genetic connectedness will allow the estimation of genetic correlations.     

Genetic analysis of competition data on Icelandic horses

In a study of the possibility of using competition data in the genetic evaluation of Icelandic horses, data from competitions held in Iceland between 1999 and 2004 and in Sweden between 1998 and 2004 were analyzed to estimate the genetic parameters of competition performance traits. The data-sets from both countries included 18 982 records of 3790 horses in 379 different events. Two types of competition were included: sport competitions and gæðinga competitions performed on oval tracks and on a straight track. Each type of competition involves several disciplines in which the horses are ridden in the various gaits. The traits analyzed were two different measures of four-gait, five-gait and tölt, and one pace trait. In both the four-gait and five-gait tests the gaits walk, trot, tölt and gallop are exhibited. In the five-gait test, pace is also exhibited. The traits tölt and pace are performances of these single gaits. Highly correlated and similar traits were combined, and three new traits relating to tölt, four-gait and five-gait were formed. No large differences in means or standard deviations of traits were found between countries. All traits were approximately normally distributed. Genetic parameters were estimated using linear animal models including the fixed effects of sex, age and event for all traits, and the level of discipline was included for some traits. Random permanent environmental effects were also included. Estimated heritabilities were moderate to high, ranging from 0.18 to 0.21 for sport-competition traits, from 0.33 to 0.35 for gæðinga-competition traits and from 0.19 to 0.22 for combined traits. Estimated genetic correlations between different sport-competition traits varied from 0.63 to 0.96, and between the two gæðinga-competition traits it was estimated at 0.43. Genetic correlations between sport- and gæðinga-competition traits ranged from − 0.42 to 1.00. It was concluded that competition traits are suitable to include in genetic evaluations.     

Genetic diversity and population structure of Polish Arabian horses assessed through breeding and microsatellite data

Polish Arabian horses are one of the most important populations of this breed in the world. Their post‐war history can be divided into two periods, with the dominant role of state studs until 1989, and the increasing significance of private breeding in the next years. The goal of the study was to evaluate genetic diversity and structure of the population under a new breeding policy. The analyses of breeding and microsatellite data from 1996 to 2012 provide a coherent picture of the population with constant flow of horses only in one direction from state to private studs. An increase in the number of broodmares was observed, from 396 mares in 1996 to 1021 mares in 2012. The proportion of foreign sires used in Polish studs also increased, from 7.1% to 37.0%. An increasing number of alleles and progressive differentiation in mares were observed. STRUCTURE analysis indicated that the Polish horses were clearly separated from foreign horses used in Polish breeding, although only one of the 75 alleles found can be considered as typically Polish. The high heterozygosity is an important feature of the Polish population; however, the decrease of heterozygosity in state broodmares was noted. This issue needs to be studied further.     

Effect of specialization on genetic parameters of studbook–entry inspection in Dutch Warmblood horses

Recent studies on data from the Dutch Warmblood Studbook (KWPN) have shown that the ongoing specialization of horses for either dressage (DH) or show jumping (JH) has led to a decreasing genetic relationship between the two subpopulations. The aim of this study was to analyse the effect of the specialization process on the genetic parameters of traits measured in the studbook–entry inspection of KWPN during the last fifteen years. Data from 18 125 DH and 23 800 JH recorded from 1998 until 2013 were used to analyse 13 traits scored in both DH and JH. Analyses were performed in a Bayesian framework. Firstly, variance components were estimated based on the whole data set. Secondly, genetic correlations between traits measured in DH or JH were estimated using bivariate analyses. Thirdly, three time periods were defined and genetic correlations between subpopulations were estimated within each period. Heritability was moderate (0.17–0.39) for both DH and JH. Genetic correlations between traits measured in DH or JH were not different from one considering the posterior standard deviation of the estimation; however, in most of the traits, a clear trend in reduction of the genetic correlation for traits expressed in DH and JH and an increase in their posterior standard deviation for recent years was observed. These results suggest that specialization could lead to differences in traits measured in DH and JH in the recent years.     

Evaluation of poor performance in competition horses: A musculoskeletal perspective. Part 1: Clinical assessment

Lack of willingness to go forward freely, lack of power, shortened steps, stiffness of the cervical or thoracolumbosacral regions are common nonspecific signs of musculoskeletal causes of poor performance in sports horses. Understanding musculoskeletal causes of poor performance requires knowledge of how normal horses move, the requirements of specific work disciplines, the nomenclature used by riders to describe how a horse is performing and the interactions between horses and riders. Determining the underlying causes needs an in‐depth history and clinical assessment, including in hand, on the lunge and ridden. Ridden exercise should include all aspects with which the rider is experiencing problems. Change of the rider can sometimes help to differentiate between horse and rider problems, but most normal horses are compliant and just because a horse goes better for a more skilled rider does not preclude an underlying pain‐related condition. Lungeing and ridden exercise should include not only trot but also transitions and canter which may highlight gait abnormalities not seen at trot. An accurate history combined with thorough clinical examination of the whole horse should permit the establishment of a list of problems requiring further investigation.