Investigation of country bias in international genetic evaluations using full-brother information

International Bull Evaluation Service evaluations from May 2005 were examined for country bias by comparing Holstein full-brother families. Countries with ≥25 bulls in multicountry full-brother families were included. The model fit evaluations of US estimated breeding values (EBV) by absorbing full-brother family and producing solutions for country of brothers. For yield and somatic cell score, 24,611 and 22,802 bulls, respectively, were included in the analysis. The study was repeated fitting evaluations on the scales of 9 countries other than the United States. On all countries’ scales, bulls from Australia, Germany, Great Britain, and Japan had greater EBV for milk yield than did their full brothers from the United States; Italian bulls had lower EBV. Bulls from Australia, Great Britain, and South Africa had an advantage in EBV for fat yield. For EBV for protein yield, bulls from Germany, Great Britain, Japan, and South Africa had an advantage, whereas bulls from the Netherlands were disadvantaged. For somatic cell score, US bulls were advantaged compared with bulls from South Africa. Significance and rankings of apparent biases were similar across country scales of the international evaluations. Causes of those differences are unknown; differences in incorporation of parental data in national and International Bull Evaluation Service evaluations are a possibility.     

Impact of paternity errors in cow identification on genetic evaluations and international comparisons.

The impact of paternity identification errors on US genetic evaluations and international comparisons of Holstein dairy bulls for milk, fat, and protein yields was investigated. Sire identification was replaced for 11% of Holstein cows that were sired by AI bulls and had records in the US database for national genetic evaluations; US evaluations were computed based on those modified pedigrees and compared with official national evaluations. Estimated breeding values from the data with introduced paternity errors were biased, especially for later generations. Estimated genetic trends decreased by 11 to 15%. Estimates of standard deviations of sire transmitting ability also decreased by 8 to 9%. International multitrait across-country comparisons of bulls were computed based on national evaluations from the United States, Canada, New Zealand, and The Netherlands. Estimates of genetic correlations between the United States and other countries decreased by 0.04 to 0.06 when US evaluations were based on modified pedigree. The resulting bias toward selection of domestic bulls and the inability to identify truly superior animals that are available internationally could decrease potential selection differentials by 0.07 to 0.09 standard deviation units on the US scale, which corresponds to sire breeding values of approximately 50 kg for milk, 3 kg for fat, and 1.7 kg for protein. Losses for the other countries were lower and ranged from 0.02 to 0.05 standard deviation units, because a correlation of less than unity with the United States decreased the impact of US cow paternity errors on the scales of other countries. Although paternity verification is desirable and technically feasible, commercial implementation would require low testing costs.     

Accuracy of foreign dairy bull evaluations in predicting United States evaluations for yield

Combining foreign daughter data with domestic information in dairy bull genetic evaluations has been shown to improve prediction of future domestic evaluations for US bulls. This study focused on the accuracy of Interbull evaluations of bulls with only foreign daughters, in predicting the latest domestic evaluations (based on US daughters). August 2003 USDA evaluations based only on US daughters were matched with the most recent February or August Interbull evaluations without US daughters. A minimum reliability of at least 80% for yield and 70% for somatic cell score (SCS) was required in both evaluations. This provided pairs of evaluations based on different daughters (foreign or US) for 286 bulls (60 bulls for SCS). Mean Interbull reliabilities on the US scale were 88% for yield and 84% for SCS, and the mean US reliability for the current evaluations was 91% for yield and 80% for SCS. Correlations between the Interbull and domestic evaluations were 0.90, 0.87, 0.90, and 0.87 for milk, fat, protein, and SCS respectively. Expected correlations were 0.89 for yield and 0.82 for SCS. Mean differences between the Interbull and current domestic evaluations were near zero. These foreign bulls had graduated from progeny test programs (selected for positive Mendelian sampling) before being marketed in the United States. Thus, parent average was a substantial underestimate of merit. The small average differences between evaluations from foreign and US daughters and high correlations indicate that Interbull evaluations based solely on foreign daughters are useful predictors of the US evaluations for yield and SCS, providing accuracy in agreement with reliabilities and much better estimates than the alternative, parent averages.     

Impact of estimated genetic correlations on international evaluations to predict milk traits

The Interbull procedure for combining dairy bull evaluations uses estimated genetic correlations between countries. It is important to know whether the resulting difficulties from differences in ranking in each country are justified by improved accuracy relative to a system assuming unity correlations. Data submitted for the May 2001 yield and somatic cell score (SCS) Interbull evaluations were processed once with the usual estimated genetic correlations (E01) and again assuming these correlations to be essentially unity (0.995; U01). The 2 sets of resulting evaluations were compared with August 2004 national evaluations (N04) for bulls not having local evaluations used in the 2001 evaluations. Thus, the examination was of Interbull evaluations from foreign data in predicting national evaluations. Countries in the study for yield were Australia, Canada, France, Germany, Great Britain, Ireland, Italy, The Netherlands, New Zealand, and the United States. Countries included for SCS were Canada, France, Germany, Great Britain, The Netherlands, and the United States. For most countries’ evaluations, standard deviations of differences between E01 or U01 and N04 were smaller for E01 by about 5 to 7% and correlations between E01 and N04 were higher by 0.01 or the same as for U01 and N04. Although use of estimated correlations tended to improve prediction, the advantage was small. A previous study had concluded no difference in accuracy for yield but did not include Australia and New Zealand, countries with the lowest correlations with other countries. Excluding bulls from those countries produced results for the other 8 countries more like the previous study, but still favoring E01 slightly. Those 2 countries were not in the SCS data. Estimated genetic correlations improved the prediction of future national evaluations slightly in most countries but more substantially for the evaluations and bulls of Australia and New Zealand.     

Methods to Combine Estimated Breeding Values Obtained from Separate Sources

Separate estimates of breeding value can be combined using meta-analysis if a combined analysis of all data is not possible or efficient. Computation is fast but not exact if the reliabilities of the separate estimates are approximate, if the extent of overlap of the datasets is unknown, or if selection has occurred across the datasets. Selection index methods were used to combine single-trait evaluations into approximate multitrait evaluations for productive life and to combine single-country rankings into multicountry rankings for yield traits. The same methods are used for males and females. To avoid iteration, parent evaluations were included in the data and combined before progeny evaluations. A little information is lost because foreign progeny contribute to domestic parents but not to domestic grandparents. Exchange of sire and dam evaluations provides a closer connection between national and international evaluations and may be more accurate than the current sire-maternal grandsire model used internationally. Correlations of the two evaluation methods were about 0.99 for 35,414 bulls from eight countries. The estimated breeding value of each bull was adjusted separately for information from foreign parents and foreign progeny. Reliabilities of the animal, its sire, and its dam were used to determine how much information came from the parents of the animal versus from its progeny and records. Multitrait reliabilities for productive life were higher than single-trait reliabilities by a mean of 7% for recent bulls and 3% for recent cows. Selection index methods may allow current multitrait across-country evaluations for bulls to be improved and to be extended to cows.     

Application of a multiple-trait, multiple-country genetic evaluation model for female fertility traits

The need to implement a method that can handle multiple traits per country in international genetic evaluations is evident. Today, many countries have implemented multiple-trait national genetic evaluations and they may expect to have their traits simultaneously analyzed in international genetic evaluations. Traits from the same country are residually correlated and the method currently in use, single-trait multiple across-country evaluation (ST-MACE), cannot handle nonzero residual correlations. Therefore, multiple-trait, multiple across-country evaluation (MT-MACE) was proposed to handle several traits from the same country simultaneously. To test the robustness of MT-MACE on real data, female fertility was chosen as a complex trait with low heritability. Data from 7 Holstein populations, 3 with 2 traits and 4 with 1 trait, were used. The differences in the estimated genetic correlations by MT-MACE and the single ST-MACE analysis (average absolute deviation of 0.064) were due to the bias of considering several traits from the same country in the ST-MACE analysis. However, the differences between the estimated genetic correlations by MT-MACE and multiple ST-MACE analyses avoiding more than one trait per country in each analysis (average absolute deviation of 0.066) were due to the lack of analysis of the correlated traits from the same country together and using the reported within-country genetic correlations. Applying MT-MACE resulted in reliability gain in international genetic evaluations, which was different from trait to trait and from bull to bull. The average reliability gain by MT-MACE over ST-MACE was 3.0 points for domestic bulls and 6.3 points for foreign bulls. Even countries with 1 trait benefited from the joint analysis of traits from the 2-trait countries. Another superiority of MT-MACE over ST-MACE is that the bulls that do not have national genetic evaluation for some traits from multiple trait countries will receive international genetic evaluations for those traits. Rank correlations were high between ST-MACE and MT-MACE when considering all bulls. However, the situation was different for the top 100 bulls. Simultaneous analysis of traits from the same country affected bull ranks, especially for top 100 bulls. Multi-trait MACE is a recommendable and robust method for international genetic evaluations and is appropriate for handling multiple traits per country, which can increase the reliability of international genetic evaluations.     

Progeny testing and selection intensity for Holstein bulls in different countries

International Bull Evaluation Service (Interbull) Holstein evaluations from February 1995 through February 2003 were used to determine characteristics of progeny testing for Holstein bulls in Australia, Canada, Denmark, France, Germany, Italy, New Zealand, Sweden, The Netherlands, and the United States. The decision to graduate a bull from progeny test (PT) was assumed to have been made based on the second Interbull evaluation, and graduation was defined as the addition of 200 daughters in the period 2.5 to 4.5 yr later. Mean bull age at PT decision varied across countries by 12 mo. Mean numbers of herds and daughters ranged from 39 to 111 and 54 to 144, respectively. Countries with higher requirements for official evaluations generally had more herds and daughters but older bulls at PT decision. Mean estimated breeding values for yield traits of sires of tested bulls were most similar across countries for fat, differing by only 6.4 kg. The four countries highest for sire protein differed only by 1 kg; however, the range was 12 kg. Percentages of bulls graduated ranged from 4.4 to 14.7 across countries. Selection intensities (standardized selection differentials) tended to be about 1.0 for yield traits. Selection intensities for somatic cell score were generally unfavorable, reflecting selection for negatively correlated yield traits. Reflecting variation in national breeding goals, selection intensities for stature were positive for most countries and highly negative for New Zealand. Selection intensity for fore udder was generally the lowest among the traits examined. All but one country showed positive selection for udder support. These statistics permit comparison of the components of PT programs across country, illustrating possible opportunities for improvement.     

Extent and benefits of multi-country progeny testing of young dairy sires

One of the current trends within the artificial insemination industry is to progeny test young dairy bulls in multiple countries. The objectives of this study were to assess the extent of multi-country progeny testing and to measure the corresponding gains in reliability of international breeding value estimates. Data of Holstein bulls that were born between July 1, 1992, and December 31, 1994, and progeny tested in countries that participate in the International Bull Evaluation Service were used in the present study, because these were the youngest bulls that had completed multi-country progeny testing before the study. Based on August 1999 international sire evaluation data, a total of 562 bulls from 10 countries were multi-country sampled for production traits during this 2.5-yr period, and 233 bulls from seven countries were multi-country sampled for type traits. The United States, Canada, The Netherlands, France, and Germany were most active in multicountry progeny testing, and Germany, New Zealand, Australia, France, and The Netherlands were the most common countries of foreign sampling. Mean reliabilities of international breeding values were calculated within each country. Means for milk yield were 0.89 for single-country sampled bulls with local progeny (i.e., progeny in the home country), 0.71 for single-country sampled bulls with no local progeny, 0.90 for multicountry sampled bulls with local progeny, and 0.78 for multi-country sampled bulls with no local progeny. Mean reliabilities for teat placement for these groups of bulls were 0.80, 0.71, 0.88, and 0.83, respectively, and means for rear udder width were 0.79, 0.60, 0.85, and 0.68, respectively. Gains in reliability in the country of foreign sampling were greatest when foreign progeny were located in countries that had low genetic correlations with the home country.     

Genetic analysis of an artificial insemination progeny test program

The success of the progeny test (PT) program from one Spanish artificial insemination (AI) organization was evaluated. The annual genetic trend for the organization was compared with PT programs from other countries. The relationships among parents' estimated breeding values (EBV) and PT results for sons were also studied. Estimated breeding values for type and production traits were obtained from international genetic evaluations from February 2004. The annual genetic gain of the Spanish PT program was similar to that of other international programs. The Spanish AI organization graduated 13% of its sampled bulls, and 52% of primiparous cows were daughters of Spanish bulls (32% from proven bulls and 20% from sampling bulls).Correlations between EBV for PT bulls and their pedigree indices (0.52 to 0.70) were slightly lower than correlations between EBV for PT bulls and their parent averages (0.63 to 0.73). Both young and mature cows contributed to genetic progress. Success of PT bulls (defined by number of second-crop daughters) depended mainly on their EBV for final score, protein yield, and the type-production index. Significant correlations of sire EBV were found for final score and type-production index with the number of second-crop daughters (0.22 and 0.17). Likewise, significant correlations of dam EBV for final score and type-production index with the number of second crop daughters were found (0.25 and 0.18). Final score and protein yield were the main factors in success of a PT bull. The type-production index for PT bulls was not important for success unless it was 2.5 standard deviations above average. The PT bulls with low EBV for type-production index were used as proven bulls when they had higher EBV either for protein or final score.     

Genomic evaluation of Brown Swiss dairy cattle with limited national genotype data and integrated external information

This study demonstrated the feasibility of a genomic evaluation for the dairy cattle population for which the small national training population can be complemented with foreign information from international evaluations. National test-day milk yield data records for the Slovenian Brown Swiss cattle population were analyzed. Genomic evaluation was carried out using the single-step genomic best linear unbiased prediction method (ssGBLUP), resulting in genomic estimated breeding values (GEBV). The predominantly female group of genotyped animals, representing the national training population in the single-step genomic evaluation, was further augmented with 7,024 genotypes of foreign progeny-tested sires from an international Brown Swiss InterGenomics genomic evaluation (https://interbull.org/ib/whole_cop). Additionally, the estimated breeding values for the altogether 7,246 genotyped domestic and foreign sires from the 2019 sire multiple across-country evaluation (MACE), were added to the ssGBLUP as external pseudophenotypic information. The ssGBLUP method, with integration of MACE information by avoiding double counting, was then performed, resulting in MACE-enhanced GEBV (GEBVM). The methods were empirically validated with forward prediction. The validation group consisted of 315 domestic males and 1,041 domestic females born after 2012. Increase, inflation, and bias of the GEBV(M) reliability (REL) were assessed for the validation group with a focus on females. All individuals in the validation benefited from genomic evaluations using both methods, but the GEBV(M) REL increased most for the youngest selection candidates. Up to 35 points of GEBV REL could be assigned to national genomic information, and up to 17 points of GEBVM REL could additionally be attributed to the integration of foreign sire genomic and MACE information. Results indicated that the combined foreign progeny-tested sire genomic and external MACE information can be used in the single-step genomic evaluation as an equivalent replacement for domestic phenotypic information. Thus, an equal or slightly higher genomic breeding value REL was obtained sooner than the pedigree-based breeding value REL for the female selection candidates. When the abundant foreign progeny-tested sire genomic and MACE information was used to complement available national genomic and phenotypic information in single-step genomic evaluation, the genomic breeding value REL for young-female selection candidates increased approximately 10 points. Use of international information provides the possibility to upgrade small national training populations and obtain satisfying reliability of genomic breeding values even for the youngest female selection candidates, which will help to increase selection efficiency in the future.     

Retrospective analysis of the accuracy of conversion equations and multiple-trait, across-country evaluations of Holstein bulls used internationally

In 1995, the multiple-trait across country genetic evaluation procedure replaced regression-based conversion equations as the preferred method for international genetic comparisons of dairy bulls. In the present study, February 1999 estimated breeding values of 632 foreign Holstein bulls that were used in Canada, Germany, Italy, The Netherlands, Sweden, and the US were compared with January 1995 predictions from home country data only. January 1995 predicted breeding values for each importing country were calculated using three methods: the multiple-trait, across-country evaluation procedure; conversion equations based on the multiple-trait, across-country evaluations; and conversion equations based on the Wilmink method. Mean correlations between 1999 estimated breeding values in the importing countries and 1995 predictions from international data were from 0.76 to 0.81 for all methods. The multiple-trait, across-country evaluation procedure is expected to lead to selection of different bulls, because bulls were allowed to be ranked differently in each country, but no significant increase in accuracy of selection was observed. The lack of improvement in accuracy of prediction was most likely due to limitations in data structure. International genetic comparisons are largely driven by data from a relatively small number of evaluated bulls with exported semen. Data from siblings and more distant relatives provide only weak, indirect genetic links between countries, and inclusion of such data seems to provide a minimal improvement in accuracy. Limitations in data structure might be alleviated by methods that define environments by climate or management factors rather than country borders.     

International genetic evaluations of Holstein sires for milk somatic cell and clinical mastitis

International genetic evaluations for milk somatic cell and clinical mastitis have been implemented on a routine basis by Interbull. This paper examines possible genetic consequences of such evaluations. Holstein data from 12 countries were used for this purpose. Trait definitions and national genetic evaluation procedures were first summarized and showed that differences between countries existed. Estimated genetic correlations among milk somatic cell in these countries ranged from 0.47 to 0.97, with a median of 0.88. Estimated genetic correlations among clinical mastitis in three Nordic countries ranged from 0.59 to 0.83, and estimated genetic correlations between clinical mastitis in the three Nordic countries and milk somatic cell in the non-Nordic countries ranged from 0.37 to 0.78 with a median of 0.55. Bulls without daughter information in the Nordic countries had low reliabilities on the Nordic clinical mastitis scales. International genetic evaluations for milk somatic cell and clinical mastitis enable a broader selection among foreign bulls, and higher selection differentials were found when using international evaluations compared with national evaluations.     

Weighting factors of sire daughter information in international genetic evaluations

International genetic evaluations of dairy bulls are currently based on national genetic evaluation results. Total number of daughters in a country is used to weight national information, but may not optimally reflect the precision of a sire's daughter contribution to international genetic evaluations. This study investigates the impact of alternative weighting factors on international evaluation results. A conventional progeny test scheme was simulated for two dairy cattle populations, with semen exchange at a fixed rate after each generation. True breeding values for both populations were generated as bivariate normal deviates. Each cow had three lactation records in one country only. After 10 generations of selection, all records were used in national breeding value prediction. National breeding values of bulls were used as input to international evaluations. Seven different weighting factors were evaluated: 1) total number of daughters; 2) total number of lactations; 3) as (one) also adjusted for finite contemporary group size; 4) as (three) also adjusted for distribution of daughters over contemporary groups; 5) effective daughter contribution considering finite contemporary group size and correlation between repeated records; 6) as (five) also considering the reliability of the daughter dam evaluation; and 7) as (five) also considering the reliability of the daughter female ancestors' evaluations. Using the last two weighting factors yielded empirically unbiased estimates of sire variance. Using total number of daughters overestimated genetic variance by up to 7%. In general, international breeding values were marginally affected by choice of weighting factor. The effect was larger when different national evaluation models had been applied in the two countries. International reliabilities for the last two weighting factors were close to expectation, whereas using total number of daughters resulted in 1 to 4% negative bias. In practice, different countries apply a wide range of national evaluation models, and genetic ties may be weak between some populations, thereby increasing the potential effect of weighting factors on international comparisons. The weighting factor developed in this study, which considers contemporary group structure, correlation between repeated records, and reliability of dams of daughters, should replace total number of daughters in international genetic evaluations of dairy sires.     

Comparison of international dairy sire evaluations from meta-analysis of national estimated breeding values and direct analysis of individual animal performance records

Our objective was to assess the predictive ability of different methodologies for international genetic evaluation of milk yield and to determine the magnitude of differences in the resulting sire estimated breeding values (EBV). Data included first lactation records of 16,057,335 Holstein-sired cows from 237,049 herds in 14 countries. Meta-analysis of national sire EBV using the multiple-trait across country evaluation (MACE) procedure, single-trait analysis of individual animal performance records, multiple-trait analysis of individual animal performance records, and borderless herd cluster model were compared by assessing predictive ability. Comparisons were based on root mean square error of sire EBV from a subset of records from cows calving between 1990 and 1995 and corresponding pedigree indices for sires that received their first genetic evaluations in 1996 or 1997. The number of bulls first evaluated in 1996 or 1997 that were in common between the top 25, 100, and 250 for pedigree index and the top 25, 100, and 250 for EBV were also determined for each method. Average root mean square error of prediction was 10.3 kg2 for the borderless single-trait model, 6.6 kg2 for the borderless herd cluster model, and 6.7 kg2 for both the borderless multiple-trait model and meta-analysis of national sire EBV using MACE. The mean numbers of common bulls among the top 25, 100, and 250, respectively, when selected on pedigree index and subsequent EBV were 11, 48, and 154 for the borderless single-trait model; 16, 66, and 176 for the borderless multiple-trait model; 16, 66, and 178 for the borderless herd cluster model; and 15, 66, and 178 for meta-analysis of national sire EBV using MACE. Rank correlations between sire EBV from different models ranged from 0.77 for the single-trait borderless model and the meta-analysis using MACE to 0.92 for the borderless multiple-trait and the borderless herd cluster models.     

Consistency of maturity rate for milk yield across countries and generations

Differences among bulls in maturity rate of their daughters for milk yield were investigated. Milk records for US Holsteins with first-parity calving dates between 1960 and 1998 were used to calculate 3 evaluations for bulls based on daughter records from parity 1, parities 1 and 2, and parities 1, 2, and 3. The 3 evaluations were used to estimate parity-specific evaluations for parities 2 and 3. Maturity rate of Holstein bull daughters in Canada and the Netherlands was compared with that for daughters of the same bulls in the United States by using official November 2004 Canadian and August 2005 Dutch parity-specific evaluations. For bulls with ≥500 first-parity daughters, correlations among parity-specific evaluations within country and birth year of bull were 0.88 between parities 1 and 2, 0.84 between parities 1 and 3, and 0.96 between parities 2 and 3 for the United States; 0.90, 0.86, and 0.97, respectively, for Canada; and 0.92, 0.89, and 0.98, respectively, for the Netherlands. Correlations between Canada and the United States for within-country differences between evaluations for parities 1 and 2 were 0.72 for bulls with ≥50 first-parity daughters and 0.89 for bulls with ≥500 first-parity daughters; corresponding correlations between the Netherlands and the United States were 0.66 and 0.82. Correlations between countries for differences between evaluations for parities 1 and 3 were slightly less, and corresponding correlations between evaluations for parities 2 and 3 were still lower. To establish whether differences between parity-specific evaluations were genetic, comparisons were made across a generation. Coefficients for regression of son on sire within country and birth years of sire and son for parity-specific evaluations and differences between parity-specific evaluations ranged from 0.35 to 0.53, with standard errors of ≤0.04. Differences in maturity rate of bull daughters were quite consistent across country, and those differences were transmitted to the sons’ daughters. Modeling to account for maturity differences should increase the accuracy of US evaluations and reduce fluctuation between evaluations, especially for bulls with daughters that deviate substantially from the population mean for maturity rate for milk yield.     

Accuracy and stability of national and international somatic cell score evaluations

Somatic cell score (SCS) evaluations have been published in the United States since 1994 and international evaluations have been available through Interbull since May 2001. The accumulated data provides an opportunity to investigate the accuracy and stability of SCS evaluations. United States domestic evaluations from January 1995 through August 2004, for 21,500 Holstein bulls were considered, over time and sequentially within bull, for changes to the November 2004 evaluation. On average, predicted transmitting ability (PTA) SCS increased (worsened) by 0.002 from earlier evaluations to November 2004. Although bias was small, PTA changes were more than expected based on change in reliability. When looked at sequentially, bulls’ earlier evaluations were generally lower (i.e., merit was overestimated) relative to November 2004. Differences were small, and PTA SCS increased steadily with the addition of second-crop daughters. All 524,081 evaluations were considered pairwise providing over 8,000,000 pairs of bulls’ evaluations for analysis of PTA differences relative to change in reliability. Agreement of observed and expected SD improved for larger changes in reliability. The November 2004 US and Interbull PTA were matched with US and Interbull PTA from May 2001 (US04, IB04, US01, and IB01, respectively) for 14,652 Holstein bulls. For bulls having only US daughters in IB01, correlations were similar for US01 and IB01 with US04, and IB01 with IB04. Corresponding regressions were all nearly 1.00. For bulls also having nonUS daughters in IB01, correlations with yield deviations calculated for later daughters (used as source of independent data) were higher (0.747 vs. 0.714) for IB01 than for US01. For bulls with added US daughters, correlation with US04 was also higher for IB01 than US01, showing that inclusion of foreign data improved predictive value of SCS evaluations.     

Invited review: Genomic selection in dairy cattle: Progress and challenges

A new technology called genomic selection is revolutionizing dairy cattle breeding. Genomic selection refers to selection decisions based on genomic breeding values (GEBV). The GEBV are calculated as the sum of the effects of dense genetic markers, or haplotypes of these markers, across the entire genome, thereby potentially capturing all the quantitative trait loci (QTL) that contribute to variation in a trait. The QTL effects, inferred from either haplotypes or individual single nucleotide polymorphism markers, are first estimated in a large reference population with phenotypic information. In subsequent generations, only marker information is required to calculate GEBV. The reliability of GEBV predicted in this way has already been evaluated in experiments in the United States, New Zealand, Australia, and the Netherlands. These experiments used reference populations of between 650 and 4,500 progeny-tested Holstein-Friesian bulls, genotyped for approximately 50,000 genome-wide markers. Reliabilities of GEBV for young bulls without progeny test results in the reference population were between 20 and 67%. The reliability achieved depended on the heritability of the trait evaluated, the number of bulls in the reference population, the statistical method used to estimate the single nucleotide polymorphism effects in the reference population, and the method used to calculate the reliability. A common finding in 3 countries (United States, New Zealand, and Australia) was that a straightforward BLUP method for estimating the marker effects gave reliabilities of GEBV almost as high as more complex methods. The BLUP method is attractive because the only prior information required is the additive genetic variance of the trait. All countries included a polygenic effect (parent average breeding value) in their GEBV calculation. This inclusion is recommended to capture any genetic variance not associated with the markers, and to put some selection pressure on low-frequency QTL that may not be captured by the markers. The reliabilities of GEBV achieved were significantly greater than the reliability of parental average breeding values, the current criteria for selection of bull calves to enter progeny test teams. The increase in reliability is sufficiently high that at least 2 dairy breeding companies are already marketing bull teams for commercial use based on their GEBV only, at 2 yr of age. This strategy should at least double the rate of genetic gain in the dairy industry. Many challenges with genomic selection and its implementation remain, including increasing the accuracy of GEBV, integrating genomic information into national and international genetic evaluations, and managing long-term genetic gain.     

Prediction of genetic correlations and international breeding values for missing traits

Prediction of genetic merit for missing traits is possible by combining available indicator traits. Indicator traits were combined using genetic correlations obtained from multiple regression equations of estimated genetic correlations among available indicator traits on variables explaining production circumstances and trait definitions. This prediction of missing traits was closer to actual breeding values than breeding values for any of the indicator traits. This was verified by evaluating clinical mastitis in each of the Nordic countries as a missing trait. The derived methodology was used to predict breeding values for clinical mastitis in the United States for local and international bulls with an average reliability of 43%.     

International genetic evaluation of Holstein bulls for overall type traits and body condition score

The study documents the procedures used to estimate genetic correlations among countries for overall conformation (OCS), overall udder (OUS), overall feet and legs (OFL), and body condition score (BCS) of Holstein sires. Major differences in traits definition are discussed, in addition to the use of international breeding values (IBV) among countries involved in international genetic evaluations, and similarities among countries through hierarchical clustering. Data were available for populations from 20 countries for OCS and OUS, 18 populations for OFL, and 11 populations for BCS. The IBV for overall traits and BCS were calculated using a multi-trait across-country evaluation model. Distance measures, obtained from genetic correlations, were used as input values in the cluster analysis. Results from surveys sent to countries participating in international genetic evaluation for conformation traits showed that different ways of defining traits are used: the overall traits were either computed from linear or composite traits or defined as general characteristics. For BCS, populations were divided into 2 groups: one scored and evaluated BCS, and one used a best predictor. In general, populations were well connected except for Estonia and French Red Holstein. The average number of common bulls for the overall traits ranged from 19 (OCS and OUS of French Red Holstein) to 514 (OFL of United States), and for BCS from 17 (French Red Holstein) to 413 (the Netherlands). The average genetic correlation (range) across countries was 0.75 (0.35 to 0.95), 0.80 (0.41 to 0.95), and 0.68 (0.12 to 0.89) for OCS, OUS, and OFL, respectively. Genetic correlations among countries that used angularity as best predictor for BCS and countries that scored BCS were negative. The cluster analysis provided a clear picture of the countries distances; differences were due to trait definition, trait composition, and weights in overall traits, genetic ties, and genotype by environment interactions. Harmonization of trait definition and increasing genetic ties could improve genetic correlations across countries and reduce the distances. In each national selection index, all countries, except Estonia and New Zealand, included at least one overall trait, whereas none included BCS. Out of 18 countries, 9 have started genomic evaluation of conformation traits. The first were Canada, France, New Zealand, and United States in 2009, followed by Switzerland, Germany, and the Netherlands in 2010, and Australia and Denmark-Finland-Sweden (joint evaluation) in 2011. Six countries are planning to start soon.     

Major advances in globalization and consolidation of the artificial insemination industry

The artificial insemination (AI) industry in the United States has gone through many consolidations, mergers, and acquisitions over the past 25 yr. There are 5 major AI companies in the United States today: 3 large cooperatives, 1 private company, and 1 public company. The latter 2 have majority ownership outside of the United States. The AI industry in the United States progeny-tests more than 1,000 Holstein young sires per year. Because healthy, mature dairy bulls are capable of producing well over 100,000 straws of frozen semen per year, only a relatively small number of bulls are needed to breed the world's population of dairy cows. Most AI companies in the United States do not own many, if any, females and tend to utilize the same maternal families in their breeding programs. Little differences exist among the selection programs of the AI companies in the United States. The similarity of breeding programs and the extreme semen-production capabilities of bulls have contributed to difficulties the AI companies have had in developing genetically different product lines. Exports of North American Holstein genetics increased steadily from the 1970s into the 1990s because of the perceived superiority of North American Holsteins for dairy traits compared with European strains, especially for production. The breeding industry moved towards international genetic evaluations of bulls in the 1990s, with the International Bull Evaluation Service (Interbull) in Sweden coordinating the evaluations. The extensive exchange of elite genetics has led to a global dairy genetics industry with bulls that are closely related, and the average inbreeding level for the major dairy breeds continues to increase. Genetic markers have been used extensively and successfully by the industry for qualitative traits, especially for recessive genetic disorders, but markers have had limited impact for quantitative traits. Selection emphasis continues to migrate away from production traits and towards nonproduction traits, especially towards health and fitness traits. Specifically, fertility has arguably become the major breeding and management issue facing dairy farmers today. Some producers have implemented crossbreeding programs in an effort to capitalize on heterosis, and crossbreeding will almost certainly need to be a bigger part of the AI companies business in the years ahead.