GENETIC PARAMETERS OF SOME TRAITS OF STALLIONS EVALUATED FOR PERFORMANCE BY TRAINING CENTRES DURING 1977-2000

Henryk Geringer de Oedenberg, Maciej Dobrowolski, Magdalena Zatoń-Dobrowolska, Heliodor Wierzbicki

ABSTRACT

Studies on genetic parameters that describe performance traits of stallions managed in Training Centres are of great importance, since they allow breeders to rationally design their breeding programmes. The outcomes of such studies also include breeding methods that are adequate for an improvement of traits which are of different heritability or which have different coefficients of correlation between them. As an effect of evaluated stallions will be significant over the decades to follow, the studies we have performed are important in terms of both theory and practice. The study covered performance test records of 2825 stallions tested in Training Centres in 1977-2000. Heritability estimates for the selected traits evaluated during performance tests were estimated with the REML method. The recorded values ranged from 0.09, for the pulse rate, to an exceptionally high record for dressage, as judged by foreign riders, reaching 0.95. Phenotypic, genetic, and environmental correlations were es

Key words: stallions, genetic parameters, performance tests.

INTRODUCTION

The problem of heritability of performance traits in horses has been targeted by many researchers for years. Velsen-Zerweck and Bruns [10] estimated the heritability basing on 2826 stallions, which had undergone 100-day tests, and 156 mares, which had passed field trial for appearance, walk, canter, trot, racing potential, and jumping. Gerber et al. [1] estimated the heritability of traits basing on performance achievements by stallions tested in Sweden during 1979-1993. Kaproń et al. [3] used performance test records of 1865 stallions obtained from Training Centres. These authors analysed two systems: (I) applied between 1973-1982 and (II) used between 1983-1992. The previous system comprised 9 traits, while the latter included 14 traits. The authors have distinguished groups of stallions by breed. Kaproń [7], in a subsequent study, included three performance testing systems into his analysis (I – 1973-1982, II – 1983-1992, and III – 1993-1996).

Kaproń et al. [2, 5] estimated heritability of the traits included in the performance tests in Training Centres, which had been passed in the systems I and II for various breeds of stallions with various methods (REML, SAS, Henderson method 3). The results obtained displayed large differences depending on the computation method used. For the same trait, the levels of heritability differed considerably. These results suggest that the most recent methods (REML) should be applied, which will allow estimation of the coefficient of heritability also when other methods yield no results. The same authors [6], who studied conformation traits of Małopolski horses, demonstrated a strong variability in the coefficients of heritability in relation to applied statistical model (18-65% in stallions and 18-57% in mares). Heritability of these traits were higher for mares than for stallions. Kownacki et al. [8] estimated genetic parameters for 743 stallions tested for performance during 1983-1988, which re presented the basis for an analysis of Training Centres selection success.

Coefficients of correlation represent the strength with which one trait depends on another. Gerber-Olsson et al. [1] estimated correlation for the traits evaluated in performance tests in Sweden. Kaproń [7] analysed a population of stallions tested for performance in Training Centres during 1973-1996.

The aim of this study was to estimate genetic parameters of selected traits of stallions tested for performance in Training Centres over the last twenty years.

MATERIALS AND METHODS

The material comprised records on 2825 stallions from performance tests carried out during 1977-2000 in Training Centres. Analyses included 15 traits: conformation score, evaluation by the Training Leader, dressage, jumping, pulse rate, walk, moving, trot, pulling test, endurance tests (depending on the performance testing system, these were: steeplechase, endurance driving test, and cross-country – overall score), foreign riders' test, i.e. dressage and jumping, and the general score.

The following linear model was used in the analyses:

Y_ijklmn = ľ + year_i + r_j + TS_k + b_l + s_m + a_n + e_ijklmn

where:
Y_ijklmn – phenotypic value of a given individual,
ľ – overall mean,
year_i – fixed effect of i-th year of testing,
r_j – fixed effect of j-th breed,
TS_k – fixed effect of k-th training centre,
b_l – fixed effect of l-th breeder,
s_m – fixed effect of m-th system of performance testing,
a_n – additive genetic effect of n-th animal,
e_ijklmn – random error.

The stallions comprised of paternal half-sib as well as full-sib groups. Heritabilities were estimated using the DFREML software package [9]. Correlations were estimated based on covariances using the paternal component. The NESTED procedure of the SAS package was used for estimations; the package was also used for calculating means, standard deviations, and the maxima of the traits.

RESULTS AND DISCUSSION

The results of the analyses obtained from the population of 2825 stallions undergoing performance tests in Training Centres are presented in Tables 1-3.

The studied population showed highly varied levels of the coefficient of heritability, which ranged from 0.09, for the pulse rate, to an exceptionally high 0.95, for the dressage test assessed by foreign riders. The value obtained for the dressage by foreign riders, however, showed a high standard error. No results were obtained for the driving test, most probably due to lack of pedigree data (sire performance) of the horses that had been tested for this trait.

The results of heritability we have obtained are similar to those reported by Kaproń et al. [3, 7]. Under the system I, the values of the coefficient of heritability ranged from 0.02, for trotting speed, to 0.568, for the steeplechase. Under the system II, the values ranged within 0.132, for the canter during cross-country I, and 0.564, for walking speed. The ranges of the heritability of the traits subjected to performance test changed from 0.02 (walking speed) to 0.568 (steeplechase) under the system I, from 0.084 (time of cross-country II) to 0.564 (walking speed) under the system II, and from 0.055 (trotting speed) to 0.858 (free-jumping test). The results obtained were much lower than those presented by Kownacki et al. [8], who estimated heritability for 10 traits. The levels of the coefficient of heritability estimated by these authors ranged from 0.16, for Training Leader's assessment, to 0.85, for cross-country. The heritability of gaits was 0.69 (trot) and 0.76 (walk). The coeffici ent of heritability for jumping reached 0.68, while for dressage – as low as 0.39. The general, final score was inheritable at the level of 0.5. This inconsistency may result from different time spans of the studies (24 years in our analyses and 5 years in the studies by Kownacki et al. [8]). On the other hand, the values were considerably lower than those obtained for stallions undergoing performance tests in Sweden or in Germany [1, 10]. This may result from a different performance testing and evaluation system. All the h² values there were higher for stallions than for mares. The coefficients of heritability for stallions oscillated around 0.5 (from 0.43, for the walk, to 0.52, for the racing ability). In terms of gaits, they obtained the following heritability: 0.46 for the walk, 0.37 for the trot, and 0.39 for the canter. The authors also estimated heritability of free-jumping (0.47) and jumping under saddle (0.32).

The most interesting traits of those included in the performance test are jumping, dressage, endurance test (steeplechase or cross-country, depending on the system), as well as the general score. For all these traits, heritability remained at a level of about 0.3, except for the cross-country (overall score), for which the coefficient of heritability was lower and reached 0.147. According to Kaproń et al. [3], the coefficients changed depending on the system of performance testing. In the system I, these were: 0.236 for dressage, 0.164 for jumping, 0.568 for endurance test (steeplechase), and 0.436 for the general score. In the system II, the values remained at a level of 0.208 for dressage; 0.224 for jumping; from 0.084 and 0.148 for time, 0.336 and 0.464 for style, and 0.132 and 0.192 for canter in the endurance test (cross-country I and II, the score split into time, style, and canter); and 0.312 for the general score. In the study by Kownacki et al. [8], we find the coefficients of her itability for these traits being much higher, namely: 0.68 for jumping, 0.39 for dressage, 0.85 for endurance test (cross-country), and 0.5 for the general score. For the population of both German and Swedish horses [1, 10], the estimated level of heritability of jumping was also higher, about 0.47.

Heritability of the gaits was 0.307 for the walk and 0.244 for the trot. According to Kaproń et al. [3], the results for these traits under the system I remained at a level 0.172 and 0.02, respectively, while being much higher under the system II, 0.564 and 0.264, respectively. In the study by Kownacki et al. [8], heritability in this case was also higher, 0.79 for the trot and 0.69 for the walk. Despite the differences in the absolute values of this index, the trend remains that it is higher for the walk. In the above-mentioned foreign studies [1, 10], heritability of gaits was estimated as follows:

- for the population of German horses: 0.43 – walk, 0.50 – trot, 0.47 – canter;
- for the population of Swedish horses: 0.46 – walk, 0.37 – trot, 0.39 – canter.

The Swedish studies reveal the same pattern of a higher heritability of the walk; according to the German studies, this trend is reversed.

All the types of correlation were found in the range between –1 and +1. In the case of phenotypic and genetic correlations, their maximum values were found between the scores for dressage and jumping (+1), both assessed by foreign riders, while negative correlation was found between foreign riders' assessment and the general score (at a level of –0.97). For environmental correlations, on the other hand, the pattern was reversed. 21 of all the estimated genetic correlations, and 25 of the phenotypic correlations, were close to zero. These included nearly all correlations between the conformation score and the remaining traits, correlations between the gaits, jumping/dressage and endurance tests, as well as between the general score and endurance tests results. Among the environmental correlations, 15 were close to zero (mainly the conformation score vs. other traits).

Among phenotypic correlations, 13 were significant (including eight negative), of which the most important seem to be negative correlations between dressage by foreign riders and free jumping, between jumping assessment by foreign riders and dressage, positive correlation between free jumping and jumping assessed by foreign riders and the general score. What is interesting, a significant negative correlation was found between the general score and gaits (walk and trot). Ten phenotypic correlations were highly significant (4 negative). A highly significant relationship was found between the walk and dressage, and the trot and dressage. Also, some correlations between the Training Leader's evaluation and other traits (free jumping, walk, trot, cross-country) were highly significant. A highly significant negative correlation was also found between the general score and dressage. Among the phenotypic correlations, as many as 19 (including 9 negative) were very highly significant. An interesting is a very highly significant relationship between jumping and dressage as well as between the trot and the walk (at a level of 0.5). So significant correlations were also found between the pulling test and other traits (some were negative). Highly significant were also relationships between foreign riders' assessment (of both dressage and jumping) and cross-country score, and between dressage and jumping assessment by foreign riders. The general score was highly significantly correlated with Training Leader's assessment (positively) and with pulse rate, pulling test, foreign riders' assessment (negatively). Kaproń et al. [4, 7] also reported numerous highly significant correlations, both within the analysed system I and the system II. In the system I, highly significant correlations were found between Training Leader's assessment and nearly all the remaining traits, between the trot and the walk, starting, jumping and general score, between the walk and starting, as well as between steeplechase and dressage, and between jumping and general score. Also, a highly significant relationship was found between jumping and dressage and between these two tests and the general score. In the system II, a number of correlations were also found, both highly significant and significant, which were distributed similarly as in the system I. Kaproń [7] obtained the highest coefficients of correlation between jumping score and dressage (system I), 0.855, in the system II between the style of cross-country 1st run and the style of cross-country 2nd run, 0.838, while in the system III between jumping under saddle assessment and free jumping, 0.884. Kaproń et al. [4], like in the studies on heritability of traits, analysed genetic and phenotypic relationships between the traits of stallion performance test, both in relation to system, I and II, and the breeds. For the total population in the system I, the authors found significant correlations between the walk and Training Leader's assessment, while highly significant between Trainin g Leader's assessment and the following traits: trot, starting, dressage, jumping, general score; between the trot and the following traits: walk, starting, jumping, general score; between the walk and starting; between endurance test (steeplechase) and: pulse rate, dressage, jumping, general score; between pulse rate and dressage and general score; between dressage and jumping and general score; between jumping and general score. For the system II, significant and highly significant correlations were very numerous. The highest phenotypic correlation occurred between the trot and canter of cross-country I (0.612). Other correlations remained below 0.5; the correlations between the general score and other traits were close to this value. Genetic correlations were much higher. Phenotypic correlations between the gaits ranged from 0.36 (walk – trot) to 0.67 (trot – canter). Correlations between the gaits and free jumping and between the gaits and jumping under saddle were low and reached from 0.11 (jumping under saddle – walk) to 0.14 (free jumping – walk) to 0.37 (jumping under saddle – canter) and 0.31 (free jumping – canter). The relationship between free jumping and jumping under saddle was strong and reached 0.57. Genetic correlations between the gaits were similar. On the other hand, correlations between the gaits and jumping were higher and were formed differently, since the lowest values were found between jumping and the trot (0.24 for free jumping) and 0.14 (for jumping under saddle), while the highest occurred between jumping and the canter, respectively 0.40 and 0.54.

Interesting are phenotypic and genetic correlations between the general score and Training Leader's evaluation, which are high and positive, whereas the correlations with foreign riders' assessment are high and negative. For the environmental correlations, this pattern is reversed. This demonstrates a rather strong effect of the score awarded to a stallion by the Training Leader on its final achievement. A strong relationship is also important between the evaluation of the gaits and a significant relationship between the gaits and dressage, which was also observed by Kaproń et al. [4] and Gerber et al. [1]. Conversely, no association was found between endurance tests and the gaits as well as other tests (jumping, dressage). A strong positive correlation occurs between jumping and dressage (also demonstrated by Kaproń et al. [4, 7]) and the evaluation of these abilities by foreign riders, which reveals versatile abilities of the horses. On the other hand, a negative correlation occurs betwee n dressage assessment by foreign riders and jumping and between jumping assessment by foreign riders and dressage, which indicates that the assessment by foreign riders would confirm the horse's predisposition to a certain type of management. This would be confirmed by a significant positive relationship between jumping assessment by foreign riders and free jumping, but also by other studies, both Polish [7] and Swedish [1]. Unfortunately, a negative (and non-significant) correlation occurs between dressage assessment by foreign riders and dressage. These significant correlations between the gaits and jumping and dressage abilities may be applied in the breeding and selection process, since evaluating and improvement of one trait is enough for the other trait to improve. Therefore, it seems important to confirm the abilities of a horse by foreign riders, whose assessment seems adequate and may create an important element of stallion evaluation in the Training Centres. Some concern, however, may arise from the negative, highly significant correlation between the assessment by foreign riders and the general score.

Three values of environmental correlation exceeded the theoretical range for these coefficients (they reached the level > –1). This could have been due to a small amount of data that the authors had for the computations.

CONCLUSIONS

1. The calculated high coefficient of heritability for important performance traits of the horses demonstrate the possibility to achieve rapid selection response.

2. Genetic and phenotypic correlations indicate importance of foreign riders' assessment. This factor should be maintained in the evaluation of half-bred stallions.

3. The positive relationship between the assessment of the walk and the trot, as well as between the assessment of jumping and dressage, seems important from the breeding standpoint; selection for one trait should contribute to a faster improvement of the other.

1. Gerber Olsson E., Árnanson Th., Näsholm A., Philipsson J., 2000. Genetic parameters for traits at performance test of stallions and correlations with traits at progeny tests in Swedish warmblood horses. Livest. Prod. Sci. 65, 81-89.

2. Kaproń M., Pięta M., Zięba G., Łukaszewicz M., Kaproń H., Stachurska A., 1996. Wpływ zastosowanych metod statystycznych na wyniki szacowania odziedziczalności użytkowych cech ogierów trenowanych w zakładach treningowych w latach 1973-1992 [Effect of applied statistical methods on the results of estimations of performance traits of stallions trained in training centres during 1973-1992]. Zesz. Nauk. Prz. Hod. 25, 43-50 [in Polish].

3. Kaproń M., Zięba G., Łukaszewicz M., Kaproń H., Janczarek I., 1996. Odziedziczalność użytkowych cech ogierów trenowanych w zakładach treningowych w latach 1973-1992 [Heritability of performance traits in stallions trained in training centres during 1973-1992]. Pr. Mater. Zootech. 49,67-74 [in Polish].

4. Kaproń M., Zięba G., Łukaszewicz M., Kaproń H., Janczarek I., 1996. Genetyczne i fenotypowe zależności między cechami użytkowymi ogierów trenowanych w zakładach treningowych w latach 1973-1992 [Genetic and phenotypic relationships between performance traits of stallions trained in training centres during 1973-1992]. Pr. Mater. Zootech. 49, 77-89 [in Polish].

5. Kaproń M., Zięba G., Strzelec K., Stachurska A., Kaproń H., Janczarek I., 1997. Ocena zmienności poziomu odziedziczalności cech użytkowych ogierów półkrwi trenowanych w zakładach treningowych (1973-1992) z uwzględnieniem ich przynależności rasowej i wykorzystywanych metod statystycznych [Evaluation of variability of the level of heritability of performance traits of half-bred stallions trained in training centres (1973-1992) with consideration to their breed and plied statistical methods]. Zesz. Nauk. AR Szczec., Zootechnika 177 (35), 177-186 [in Polish].

6. Kaproń M., Pięta M., Patkowski K., Kaproń H., 1998. Zmienność Odziedziczalności cech pokroju koni małopolskich zależnie od statystycznego modelu szacowania [Variability of heritability of conformation traits in Małopolski horses depending on the statistical model used for estimations]. Pr. Mater. Zootech. 53, 49-56 [in Polish].

7. Kaproń M., 2001. Wykorzystanie parametrów genetycznych w modyfikacji systemów prób dzielności ogierów półkrwi (w: Wykorzystanie wyników prób dzielności koni półkrwi) [Using genetic parameters to modify systems of half-bred stallion performance testing (in: Application of results of half-bred horses performance tests)]. IGHZ PAN, Jastrzębiec [in Polish].

8. Kownacki M., Lipińska Z., Kozaczyński K., 1993. Selekcja ogierów w Zakładach Treningowych na podstawie wyników oceny użytkowości [Stallion selection in Training Centres basing on their performance evaluation results]. Rocz. Nauk. Zootech. 20 (2), 31-38 [in Polish].

9. Meyer K., 1998. DFREML User Notes version 3.0 b. Animal Genetics and Breeding Unit, University of New England, Armidale, Australia.

10. Velsen-Zerweck A., Bruns E., 1997. Estimation of genetic parametres using data from performance tests of young german riding horses. The 48th Annual Meeting of The European Association for Animal Production, Wiedeń 25-28 August 1997. Wageningen Pers.

Magdalena Zatoń-Dobrowolska, Heliodor Wierzbicki
Department of Genetics and Animal Breeding
Agricultural University of Wrocław
Kożuchowska 7, 51-631 Wrocław, Poland
e-mail: magda@gen.ar.wroc.pl

Responses to this article, comments are invited and should be submitted within three months of the publication of the article. If accepted for publication, they will be published in the chapter headed ‘Discussions’ in each series and hyperlinked to the article.