Dairy Sires Research Articles

Effect of Heterogeneity of Variance on Dairy Sire Evaluation

The effects of heterogeneity of variance among herds on dairy sire evaluation were examined. Data consisted of 111,196 first lactation Holstein production records of heifers calving in Ontario between 1978 and 1983. To account for selection, older sires without first crop daughters in the data were treated as fixed effects. Herds were classified into one of nine groups based on their estimated sire and residual components. Sires were evaluated by three models. Heterogeneity of variance was ignored in the first model. When accounting for heterogeneity of variance, a genetic correlation of unity between the genetic values of the same sire in the nine variance groups was assumed in one model, and correlations of <1 were assumed for a second model. Data were randomly divided into two equal subsets and sires were evaluated by each method within the subsets. Correlations for the various models were above .92 indicating small differences among the models. Accounting for heterogeneity of variance did not improve the accuracy of sire evaluations.

Potential Impact of Bovine Somatotropin on Dairy Sire Evaluation

Use of bovine somatotropin to increase production of milk and milk solids poses the problem of allocation within herd and implications for genetic evaluatation records with and without bovine somatotropin were simulated for different correlations between performance with and without treatment and response to it. Resulting within-herd variances and accuracy and bias in sire evaluations were compared for different treatment strategies. Uniform allocation of bovine somatotropin to all cows in the herd appeared to have the least potential to cause erroneous sire rankings. Random allocation of bovine somatotropin to a part of the herd is unlikely to cause serious problems in sire evaluation, although it is expected to affect cow evaluations. Systematic use of bovine somatotropin on low or high yielding cows or progeny groups may cause substantial errors in genetic evaluation.

Heterogeneity of Variances Among Herds and Its Effects on Dairy Sire Evaluation

Heterogeneity of Variances Among Herds and Its Effects on Dairy Sire Evaluation

Evaluation of dairy sires in Swiss herds using a rolling months model

Evaluation of dairy sires in Swiss herds using a rolling months model

Developments in methods for breeding value and parameter estimation in livestock

In an effective animal breeding programme the objectives are appropriate, the population structure is good, suitable records are collected, and these records are used in an optimal way so that animals likely to have the best offspring are chosen.Estimates of genetic parameters such as heritabilities and correlations influence the design of the programme: for example, the optimal number of daughters to progeny test per dairy sire or, more fundamentally, whether to adopt performance or progeny testing, whether to record food conversion efficiency or simply growth rate, and whether or not to collect records on some indicator trait such as blood urea nitrogen level. The estimates of parameters also influence the weightings given to different traits and to individuals and relatives' performance in breeding value estimation and selection. Methods for estimating parameters have become much more sophisticated as a result of advances in theory, particularly, restricted maximum likelihood (REML), in computer power, and in development of special computer programs.

Relationships among Growth Hormone and Prolactin Secretory Parameter Estimates in Holstein Bulls and Their Predicted Differences for Lactational Traits

Selection of dairy sires is based on the production records of their female ancestors, half-sibs and daughters. No trait expressed by the sire is used. Concentrations of growth hormone (GH) and prolactin (PRL), hormones produced in both males and females that are fundamental in lactation, may be correlated with production. A study was conducted to determine whether measures of these hormones in the sire would be useful predictors of lactational ability of daughters. Blood samples were collected at 15-min intervals for 8 h from 26 Holstein bulls (5.5 yr of age) that had one progeny summary available. Plasma concentrations of GH and PRL were quantified and the mean and baseline concentrations and the frequency and mean amplitude of the secretory peaks were determined for each bull. Concentrations among these values and bulls' predicted differences (PD) were determined. Significant negative correlations were detected for frequency of GH peaks and PD for yield of milk, fat and protein; correlations were positive for PRL baseline concentrations and PD for fat and protein (P less than .10), and correlations were negative for frequency of PRL peaks and PD for milk, fat and protein (P less than .10). Addition of estimates of bull hormone secretory parameters to breeding values based on performance of relatives considerably improved the accuracy (R2) for predicting progeny performance from sire information. Certain characteristics of the patterns of GH and PRL secretion may be heritable and aid in identification of superior dairy animals.

SIRE-SON CORRELATIONS OF GENETIC EVALUATIONS FOR FAT PERCENTAGE IN HOLSTEIN-FRIESIANS

The correlation between sire and son genetic evaluations for fat percentage has been shown to be lower than corresponding correlations for milk or fat yield for the population of bulls in artificial insemination. The purpose of this study was to determine the cause of this low correlation. Based on 302 935 first-lactation daughters of 11 396 sires, the heritabilities of milk, fat and fat percentage were 0.28, 0.27 and 0.52, respectively. Indirect versus direct calculation of sire evluations for fat percentage were correlated at 0.997, and resulting sire-son correlations were 0.45 for directly calculated proofs and 0.43 for indirectly calculated proofs. The low sire-son correlations were due to the intense selection of sires of sons, especially for fat percentage, which resulted in a change in genetic parameters among the bulls used as sires of sons compared to the parameters in the overall population. Key words: Fat percentage, pedigree indexing, dairy sires

Dairy sire evaluation fitting some of the herd-year-season effects as random

Three models of sire evaluation using different environmental groupings were compared. Effects fitted were herd, period (either 6 or 12 months) within herd, season (either 1 or 2 or 4 months) within period within herd, sire and linear and quadratic regressions on age at calving. Models differed in fitting (1) the effect of herd-period-season fixed, or (2) herd-period fixed and herd-period-season random, or (3) herd fixed, herd-period and herd-period-season random. The overall effects of period and season of calving were regarded as fixed, and were removed by precorrection. Records of first lactation fat yield on 49 242 progeny of 69 widely used proven Friesian-Holstein sires in 1628 herds in England and Wales were used. Compared to Model 1, Model 2 required about four-fifths and Model 3 required two-thirds of the effective number of daughters to give the equivalent variance of the estimates of sire effects. Using random effects models the relative advantage, in terms of a smaller variance of sire effects, increased as the size of herd-period-season subclass decreased. In herd-period-season fixed effects models subclasses with a single of few records, or subclasses with all or almost all records of the same sire, contribute nothing or little to the progeny group comparisons. The random effects models could avoid these losses, and were considered to be useful especially where herds are small, provided sires can be assumed as randomly distributed over environmental subclasses.

Assessing the comparative economics of a biotechnology: Artificial insemination dairy sires

Much opportunity exists to increase economic returns from the artificial insemination, dairy sire selection decision. This article measures potential returns to producers with alternative objective functions, management and financial characteristics. The costs of wrong decisions from incorrect information are measured by differences between alternative sets of top-20 Holstein sires selected from more than 400 available. Correct specification of the objective function is most important. Determination of planning horizon and conception rate also merits allocation of some scarce management resources. Correct information about length of calving interval, female mortality rate, and opportunity cost of capital, however, is economically inconsequential.

Sire evaluation for type traits with linear and non-linear procedures

Breeding values of dairy sires for some categorically scored type traits were predicted by BLUP and by a non-linear procedure called GFCAT, which is based on the threshold concept and takes the categorical nature of the data into account. Comparisons of the two procedures were made by product-moment correlations between breeding values predicted by the two procedures. The correlations were very high and in the range 0.998–0.999. Reasons for the close relationship between the breeding values predicted by the two procedures are discussed. It was concluded that it was not worth the extra computational costs to use the non-linear procedure for the actual data. This conclusion may well be different for other traits, other designs or other modes of classification.

Considering relationships and genetic grouping in mixed model sire evaluation: A simulation study

In sire evaluation using the best linear unbiased prediction (BLUP) method a relationship matrix including only the relationships among sires can be used. The aim of the present study was to show properties of this method when models with or without grouping of sires are applied. The examination consisted of Monte-Carlo simulation of special situations occurring in dairy sire evaluation. Up to 20 repetitions were generated for each data set. Sire effects were equal for all repetitions of a particular set, cow and residual effects being sampled afresh. Thus, in addition to the correlation between true and estimated breeding value, the true standard deviation of estimated breeding value could be calculated from the results of repeated sire evaluation. With respect to these parameters, the results show that models which do not contain sire grouping but include relationships were superior to grouping models even when relationships are considered in addition. This superiority increases with decreasing number of daughters per sire. The hypothesis that a model without sire grouping but including a relationship matrix leads to a bias in the estimation of breeding values when dams of bulls are selected, could not be confirmed. On condition that the frequency of sires which are related to other sires is high, a model including a relationship matrix but without grouping can be recommended.

Computing algorithm for dairy sire evaluation on several lactations considered as the same trait

Computing algorithm for dairy sire evaluation on several lactations considered as the same trait

Computing algorithm for dairy sire evaluation on several lactations considered as the same trait

Computing algorithm for dairy sire evaluation on several lactations considered as the same trait

Model for international evaluation of dairy sires

A linear statistical model is put forward for use in comparing the genetic level of dairy sires based on their progeny test evaluations from one or more countries. Additive genetic relationships among bulls are included to provide more connections or comparisons among countries. Several definitions of genetic differences among countries are presented. A small example illustrates the use of this model.

Evaluating dairy sires for conformation of their daughters: use of first classification records

Type classification records for 18 132 Australian Holstein-Friesian heifers were analysed. These consisted of 27 traits scored in three or six categories, from three rounds of classification between 1981 and 1983. Only first lactation, first classification records were considered. The model of analysis included herd-round-classifier subclasses as fixed and sires as random effects, fitting age at classification as a linear and quadratic covariable within subclasses. Herd-round-classifier effects explained between 18% and 37% of total sums of squares. Age accounted for 5.8-2.4% for traits related to body size and for 1.6% or less for the other traits. Heritability estimates obtained using a Restricted Maximum Likelihood procedure ranged from 0.44 for total score, 0.42 for stature and 0.40 for dairy character, to 0.10 or less for feet and legs, rear heel, rear legs set, bone quality and rear teat placement. On average, values were higher than corresponding 'all-lactation' estimates. Covariance components between all traits were obtained. The resulting genetic variance/covariance matrix was then forced to be positive semi-definite before calculating genetic and phenotypic correlations. Breeding value estimates for all sires and traits were determined using a univariate Best Linear Unbiased Prediction procedure for the above model. In addition, the relationship matrix between males was incorporated. There were 2 597 sires with an average of 23.2 effective daughters, and 474 sires without daughter records included as male ancestors. The association between breeding value estimates for different traits was examined by multiple regression. Sire-son regressions were determined and compared with their expectations. Australian Breeding Values for production were obtained for a subset of sires and contrasted to the type proofs. There seemed to be little correlation between genetic merit for type and milk production.

Ranking Dairy Sires by a Linear Programming Dairy Farm Model

Dairy sires were ranked for overall merit by an average daughter's contribution to farm net profit. Biological characteristics of sires and economic factors of a dairy farm were linked by linear programming. Availability and constraints of resources were in the model. Average daughter's returns over variable costs attributable to sire proofs for several traits was the measure of sire's net merit. The index of total economic merit for sires was the amount of change of net profit by milking progeny of different sires. Ranking considered sire's contribution to milk yield and feed intake of daughters caused by variations of proofs for milk, fat percent, and size, sire's nonreturn rate, veal calf sales of the offspring, and labor costs of slow versus fast milking daughters. Size of quota for daily milk shipments, cow housing capacity, labor for milking, and milk tank capacity were critical in determining ranking and forced greater emphasis on traits other than milk yield. Correlations between sire ranking on returns over variable costs and sire proofs were highest for milk and significant for fat percent, milking speed, size, and nonreturn rate. These traits had high standard partial correlations with and explained most of the variation of the index of total economic merit. This method aims at maximizing farm profits and may be applied to rank dairy sires on a national basis or to select sires for specific dairy farm operations. Production dollar index ratings on the same sires were closely correlated (−.87 to −.96) with profit index ranking but are potentially misleading if constraints exist that limit maximum milk output per farm.

Increasing Pedigree Contribution to Dairy Sire Evaluation

Methods to improve contributions of ancestor information to sire evaluations computed by the Modified Contemporary Comparison were studied. Pedigree index (one-half evaluation of sire plus one-fourth evaluation of maternal grandsire) was used to indicate ancestor merit with adjustment for type of sampling program and birth date. For Holsteins, effect of sampling through artificial insemination was 35kg for milk yield; effect of birth date (trend of daughter merit not accounted for by pedigree index or type of sampling) was 7.3kg/yr for milk yield and .36kg/yr for fat yield. Trends were larger for milk and fat yields of Ayrshires, Brown Swiss, Guernseys, and Jerseys except for fat yield of Ayrshires, which had no significant trend. Correlations of daughter merit with pedigree index and with present group average (average daughter merit of bulls with similar pedigree indexes for milk yield) were computed. For milk yield, correlations of daughter merit with pedigree index and with group average were nearly the same; however, for fat yield, pedigree index had a higher correlation with daughter merit than did group average. Accounting for residual trend resulted in further increase of correlation. Replacing one-fourth evaluation of maternal grandsire with one-half Cow Index of dam also improved correlations slightly. Evaluations can be modified so that pedigree index adjusted by birth date replaces group average in representing ancestor merit. Such evaluations better predict performance of future daughters than do evaluations with grouping by pedigree index for milk yield, particularly for evaluations for fat yield. Dam's Cow Index provides little additional information.

Scope for evaluating dairy sires using first and second lactation records

Model calculations on the value of including second lactation records in a size evaluation procedure suggest an increase in accuracy by about 6% for sires with 25 or more first lactation daughters and 50–70% surviving to second lactation. A two-trait Best Linear Unbiased Prediction procedure utilises information from both lactations simultaneously, so that accuracies are improved through additional genetic information and extra connections in the data. Measures are described to assess the relative importance of both sources. Analyses of records for milk yield of daughters of British Friesian sires tested between 1968/69 and 1975/76 gave an average increase in accuracy of 5–6% for a multivariate first and second versus a univariate first lactation evaluation. Overall, extra genetic information and the improvement of the data structure proved to be equally important, while for the last batch of sires tested, with very few second records available, an increase in accuracy by 3% was almost entirely due to the latter. The rank correlation between breeding value estimates from the two procedures was 0.96 and only small effects of culling of cows were found. Heifer performance, therefore, is an efficient selection criterion to improve lifetime performance. Nevertheless, there is scope, albeit small, for the use of later lactation records in evaluating dairy sires.

Genetic Groups in Dairy Sire Evaluation under a Selection Model

The frequently arbitrary application and definition of genetic groups makes sire grouping the “weak link” of the sire evaluation process. A selection model for records not in the model for observations is similar in form to the genetic groups model. Implications of this correspondence are discussed and illustrated through a small example. Guidelines for possible grouping strategies also are discussed.

Associations Among Progeny Tests of Single or Pooled Lactations1,2

Genetic correlations between Predicted Difference for first and second, first and third, and second and third lactations milk were .82, .62, and .94 and for Predicted Difference percent fat .94 for the three associations.Regression coefficients for following on previous lactation Predicted Difference were less than unity; thus, Predicted Difference from first lactation might be overweighted when age adjusted and pooled with second and third lactations.Regressions of Predicted Difference milk of sons on sires for first, second, third, and pooled lactation records were .32±18, .42±.21, .56±.31, .35±.15, and intraclass correlations between half brothers were .09±.08, .28±.11, .25±.11, and .12±10. Predicted Differences for second and third lactations were, thus, near the theoretical expectation of .5 and .25 for regressions of son on sire and between half-brother correlations but lower for Predicted Differences of first and pooled lactations.Variance of Predicted Difference of first lactation between sires of sons was small compared with the variance of sons within sires. It was postulated that effective selection on the sire-to-sire path on Predicted Difference of first lactation had reduced predictability on the sire-to-sire improvement path. An interaction of sire by number of lactations may be inferred from the incomplete genetic correlations between Predicted Differences of first and following lactations and from the reduction in the variance between sire in first but not in later lactations. Progeny tests of single lactation seem warranted for the production profile of dairy sires.