Genomic selection of purebred animals for crossbred performance in the presence of dominant gene action

Background

Genomic selection is an appealing method to select purebreds for crossbred performance. In the case of crossbred records, single nucleotide polymorphism (SNP) effects can be estimated using an additive model or a breed-specific allele model. In most studies, additive gene action is assumed. However, dominance is the likely genetic basis of heterosis. Advantages of incorporating dominance in genomic selection were investigated in a two-way crossbreeding program for a trait with different magnitudes of dominance. Training was carried out only once in the simulation.

Results

When the dominance variance and heterosis were large and overdominance was present, a dominance model including both additive and dominance SNP effects gave substantially greater cumulative response to selection than the additive model. Extra response was the result of an increase in heterosis but at a cost of reduced purebred performance. When the dominance variance and heterosis were realistic but with overdominance, the advantage of the dominance model decreased but was still significant. When overdominance was absent, the dominance model was slightly favored over the additive model, but the difference in response between the models increased as the number of quantitative trait loci increased. This reveals the importance of exploiting dominance even in the absence of overdominance. When there was no dominance, response to selection for the dominance model was as high as for the additive model, indicating robustness of the dominance model. The breed-specific allele model was inferior to the dominance model in all cases and to the additive model except when the dominance variance and heterosis were large and with overdominance. However, the advantage of the dominance model over the breed-specific allele model may decrease as differences in linkage disequilibrium between the breeds increase. Retraining is expected to reduce the advantage of the dominance model over the alternatives, because in general, the advantage becomes important only after five or six generations post-training.

Conclusion

Under dominance and without retraining, genomic selection based on the dominance model is superior to the additive model and the breed-specific allele model to maximize crossbred performance through purebred selection.     

Comparing genomic prediction accuracy from purebred,crossbred and combined purebred and crossbred reference populations in sheep

Background

The accuracy of genomic prediction depends largely on the number of animals with phenotypes and genotypes. In some industries, such as sheep and beef cattle, data are often available from a mixture of breeds, multiple strains within a breed or from crossbred animals. The objective of this study was to compare the accuracy of genomic prediction for several economically important traits in sheep when using data from purebreds, crossbreds or a combination of those in a reference population.

Methods

The reference populations were purebred Merinos, crossbreds of Border Leicester (BL), Poll Dorset (PD) or White Suffolk (WS) with Merinos and combinations of purebred and crossbred animals. Genomic breeding values (GBV) were calculated based on genomic best linear unbiased prediction (GBLUP), using a genomic relationship matrix calculated based on 48 599 Ovine SNP (single nucleotide polymorphisms) genotypes. The accuracy of GBV was assessed in a group of purebred industry sires based on the correlation coefficient between GBV and accurate estimated breeding values based on progeny records.

Results

The accuracy of GBV for Merino sires increased with a larger purebred Merino reference population, but decreased when a large purebred Merino reference population was augmented with records from crossbred animals. The GBV accuracy for BL, PD and WS breeds based on crossbred data was the same or tended to decrease when more purebred Merinos were added to the crossbred reference population. The prediction accuracy for a particular breed was close to zero when the reference population did not contain any haplotypes of the target breed, except for some low accuracies that were obtained when predicting PD from WS and vice versa.

Conclusions

This study demonstrates that crossbred animals can be used for genomic prediction of purebred animals using 50 k SNP marker density and GBLUP, but crossbred data provided lower accuracy than purebred data. Including data from distant breeds in a reference population had a neutral to slightly negative effect on the accuracy of genomic prediction. Accounting for differences in marker allele frequencies between breeds had only a small effect on the accuracy of genomic prediction from crossbred or combined crossbred and purebred reference populations.     

Maximizing crossbred performance through purebred genomic selection

Background

In livestock production, many animals are crossbred, with two distinct advantages: heterosis and breed complementarity. Genomic selection (GS) can be used to select purebred parental lines for crossbred performance (CP). Dominance being the likely genetic basis of heterosis, explicitly including dominance in the GS model may be an advantage to select purebreds for CP. Estimated breeding values for CP can be calculated from additive and dominance effects of alleles that are estimated using pure line data. The objective of this simulation study was to investigate the benefits of applying GS to select purebred animals for CP, based on purebred phenotypic and genotypic information. A second objective was to compare the use of two separate pure line reference populations to that of a single reference population that combines both pure lines. These objectives were investigated under two conditions, i.e. either a low or a high correlation of linkage disequilibrium (LD) phase between the pure lines.

Results

The results demonstrate that the gain in CP was higher when parental lines were selected for CP, rather than purebred performance, both with a low and a high correlation of LD phase. For a low correlation of LD phase between the pure lines, the use of two separate reference populations yielded a higher gain in CP than use of a single reference population that combines both pure lines. However, for a high correlation of LD phase, marker effects that were estimated using a single combined reference population increased the gain in CP.

Conclusions

Under the hypothesis that performance of crossbred animals differs from that of purebred animals due to dominance, a dominance model can be used for GS of purebred individuals for CP, without using crossbred data. Furthermore, if the correlation of LD phase between pure lines is high, accuracy of selection can be increased by combining the two pure lines into a single reference population to estimate marker effects.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-015-0099-3) contains supplementary material, which is available to authorized users.     

Genomic selection of purebreds for crossbred performance

Background

One of the main limitations of many livestock breeding programs is that selection is in pure breeds housed in high-health environments but the aim is to improve crossbred performance under field conditions. Genomic selection (GS) using high-density genotyping could be used to address this. However in crossbred populations, 1) effects of SNPs may be breed specific, and 2) linkage disequilibrium may not be restricted to markers that are tightly linked to the QTL. In this study we apply GS to select for commercial crossbred performance and compare a model with breed-specific effects of SNP alleles (BSAM) to a model where SNP effects are assumed the same across breeds (ASGM). The impact of breed relatedness (generations since separation), size of the population used for training, and marker density were evaluated. Trait phenotype was controlled by 30 QTL and had a heritability of 0.30 for crossbred individuals. A Bayesian method (Bayes-B) was used to estimate the SNP effects in the crossbred training population and the accuracy of resulting GS breeding values for commercial crossbred performance was validated in the purebred population.

Results

Results demonstrate that crossbred data can be used to evaluate purebreds for commercial crossbred performance. Accuracies based on crossbred data were generally not much lower than accuracies based on pure breed data and almost identical when the breeds crossed were closely related breeds. The accuracy of both models (ASGM and BSAM) increased with marker density and size of the training data. Accuracies of both models also tended to decrease with increasing distance between breeds. However the effect of marker density, training data size and distance between breeds differed between the two models. BSAM only performed better than AGSM when the number of markers was small (500), the number of records used for training was large (4000), and when breeds were distantly related or unrelated.

Conclusion

In conclusion, GS can be conducted in crossbred population and models that fit breed-specific effects of SNP alleles may not be necessary, especially with high marker density. This opens great opportunities for genetic improvement of purebreds for performance of their crossbred descendents in the field, without the need to track pedigrees through the system.     

Imputation of genotypes in Danish purebred and two-way crossbred pigs using low-density panels

Background

Genotype imputation is commonly used as an initial step in genomic selection since the accuracy of genomic selection does not decline if accurately imputed genotypes are used instead of actual genotypes but for a lower cost. Performance of imputation has rarely been investigated in crossbred animals and, in particular, in pigs. The extent and pattern of linkage disequilibrium differ in crossbred versus purebred animals, which may impact the performance of imputation. In this study, first we compared different scenarios of imputation from 5 K to 8 K single nucleotide polymorphisms (SNPs) in genotyped Danish Landrace and Yorkshire and crossbred Landrace-Yorkshire datasets and, second, we compared imputation from 8 K to 60 K SNPs in genotyped purebred and simulated crossbred datasets. All imputations were done using software Beagle version 3.3.2. Then, we investigated the reasons that could explain the differences observed.

Results

Genotype imputation performs as well in crossbred animals as in purebred animals when both parental breeds are included in the reference population. When the size of the reference population is very large, it is not necessary to use a reference population that combines the two breeds to impute the genotypes of purebred animals because a within-breed reference population can provide a very high level of imputation accuracy (correct rate ≥ 0.99, correlation ≥ 0.95). However, to ensure that similar imputation accuracies are obtained for crossbred animals, a reference population that combines both parental purebred animals is required. Imputation accuracies are higher when a larger proportion of haplotypes are shared between the reference population and the validation (imputed) populations.

Conclusions

The results from both real data and pedigree-based simulated data demonstrate that genotype imputation from low-density panels to medium-density panels is highly accurate in both purebred and crossbred pigs. In crossbred pigs, combining the parental purebred animals in the reference population is necessary to obtain high imputation accuracy.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-015-0134-4) contains supplementary material, which is available to authorized users.     

Prenatal mortality and its effects on subsequent fertility in purebred and crossbred Gir cows

A study was conducted on 851 cows (Gir cows and their crossbreeds) between 1971 and 1984 at the Livestock Farm of the All India Coordinated Research Project on Cattle in Jabalpur, India. The cows were impregnated with frozen or fresh semen from 90 bulls and a total of 2699 pregnancies resulted. An overall incidence of prenatal mortality was recorded at 4.89%. A significantly higher (P / 0.01) incidence of prenatal mortality was encountered in three breed crosses in comparison with Gir cows and their first generation crossbreeds. Parity of dam had no effect on prenatal mortality. However, effects of season and individual sires on prenatal mortality were documented. It was observed that reproductive performance was improved significantly after prenatal mortality.     

Optimizing purebred selection for crossbred performance using QTL with different degrees of dominance

A method was developed to optimize simultaneous selection for a quantitative trait with a known QTL within a male and a female line to maximize crossbred performance from a two-way cross. Strategies to maximize cumulative discounted response in crossbred performance over ten generations were derived by optimizing weights in an index of a QTL and phenotype. Strategies were compared to selection on purebred phenotype. Extra responses were limited for QTL with additive and partial dominance effects, but substantial for QTL with over-dominance, for which optimal QTL selection resulted in differential selection in male and female lines to increase the frequency of heterozygotes and polygenic responses. For over-dominant QTL, maximization of crossbred performance one generation at a time resulted in similar responses as optimization across all generations and simultaneous optimal selection in a male and female line resulted in greater response than optimal selection within a single line without crossbreeding. Results show that strategic use of information on over-dominant QTL can enhance crossbred performance without crossbred testing.     

Progesterone determinations and clinical examinations of reproductive organs in purebred and crossbred female Zebu cattle

Five experiments were undertaken to investigate variation in progesterone concentrations as related to the collection and handling of samples of milk and blood, to determine reference values for progesterone and to evaluate clinical findings in relation to progesterone data from pure- and crossbred Zebu cattle reared in the Peruvian tropics. Whole-milk progesterone concentrations obtained from 12 Holstein x Nellore pregnant cows at hourly intervals over a 24-h period were highest immediately after milking; this peak was followed by a sharp drop over the next 3 h. Milk-fat content from 28 Brown Swiss x Nellore pregnant cows increased from 2.4% before milking to 6.7% afterwards (P < 0.05), whereas progesterone concentrations in whole milk increased from 18.4 to 59.5 nmol/1 (P < 0.05). Progesterone concentrations in fat-free milk were stable, with the exception of the fore-milk sample, which was lower than subsequent samples collected during milking (P < 0.05). Blood samples collected from 23 purebred, pregnant Nellore cows, were divided into four aliquots, and plasma and serum were harvested periodically over the next 120 h of storage at +4 degrees C, or in the shade at ambient air temperatures. The results indicate that blood samples can be stored unseparated at both temperatures studied for up to 3 h without severe loss of progesterone. Milk samples collected at different intervals during the luteal phase of the estrous cycle and during early and mid-pregnancy from crossbred cows showed no significant differences in progesterone concentrations between Days 9 to 13 of the cycle and Days 9 to 13 of gestation. Progesterone levels during advanced gestation were higher (P < 0.05). Out of 2,607 clinical examinations, the level of agreement between palpatory findings and progesterone determinations was 95.6 and 81.9% in diagnosing non-luteal and luteal structures, respectively. Significant differences in the level of agreement between palpators were found (P < 0.01). It is concluded that milk samples, preferably composite milk or strippings, must be consistently collected at the same stage of milking, and that centrifugation of blood samples should be done as soon as possible and not later than 3 h after collection.     

Dissecting total genetic variance into additive and dominance components of purebred and crossbred pig traits

The partition of the total genetic variance into its additive and non-additive components can differ from trait to trait, and between purebred and crossbred populations. A quantification of these genetic variance components will determine the extent to which it would be of interest to account for dominance in genomic evaluations or to establish mate allocation strategies along different populations and traits. This study aims at assessing the contribution of the additive and dominance genomic variances to the phenotype expression of several purebred Piétrain and crossbred (Piétrain × Large White) pig performances. A total of 636 purebred and 720 crossbred male piglets were phenotyped for 22 traits that can be classified into six groups of traits: growth rate and feed efficiency, carcass composition, meat quality, behaviour, boar taint and puberty. Additive and dominance variances estimated in univariate genotypic models, including additive and dominance genotypic effects, and a genomic inbreeding covariate allowed to retrieve the additive and dominance single nucleotide polymorphism variances for purebred and crossbred performances. These estimated variances were used, together with the allelic frequencies of the parental populations, to obtain additive and dominance variances in terms of genetic breeding values and dominance deviations. Estimates of the Piétrain and Large White allelic contributions to the crossbred variance were of about the same magnitude in all the traits. Estimates of additive genetic variances were similar regardless of the inclusion of dominance. Some traits showed relevant amount of dominance genetic variance with respect to phenotypic variance in both populations (i.e. growth rate 8%, feed conversion ratio 9% to 12%, backfat thickness 14% to 12%, purebreds-crossbreds). Other traits showed higher amount in crossbreds (i.e. ham cut 8% to 13%, loin 7% to 16%, pH semimembranosus 13% to 18%, pH longissimus dorsi 9% to 14%, androstenone 5% to 13% and estradiol 6% to 11%, purebreds-crossbreds). It was not encountered a clear common pattern of dominance expression between groups of analysed traits and between populations. These estimates give initial hints regarding which traits could benefit from accounting for dominance for example to improve genomic estimated breeding value accuracy in genetic evaluations or to boost the total genetic value of progeny by means of assortative mating.     

Managing the transition from purebred to rotational crossbred dairy cattle herds: three technical pathways from a retrospective case-study analysis

The growing interest in rotational crossbreeding in Western countries is due to its potential to improve reproductive and health performances of cows. Although a large amount of research focuses on assessing crossbred cows’ performances, how to manage the transition from purebred to rotational crossbred herds is under-explored. Based on a retrospective analysis of French dairy herd case studies, we aimed to identify and characterise technical pathways to make such a transition. In 2018, we performed semi-directive interviews on 26 commercial dairy farms. Data were collected to describe changes in breeding, replacement and culling management practices from the first crossbred mating with purebred cows to the management of a mainly crossbred herd in 2018. Based on a multivariate analysis, we identified two main guidelines structuring technical pathways to move towards rotational crossbred herds: (i) the depth and scale of change (i.e. farm v. herd) associated with the introduction of rotational crossbreeding in the whole-farm dynamics and (ii) the changes in herd replacement and breeding practices to adapt to the evolution of herd demographics induced by the evolution of the dairy crossbred mating rate over time (high from the beginning v. distributed over time). Hierarchical clustering discriminated three groups of farmers differing in their technical pathway to move towards a rotational crossbred herd. In pathway 1, farmers customised one or several rotational crossbreeding schemes to support whole-farm transition towards an organic or grass-based system. Once the scheme stabilised, they quickly implemented it and had to readjust replacement and culling practices to regulate imbalance in herd demographics induced by the improvement in cow fertility. In pathway 2, farmers also customised one or several rotational crossbreeding schemes to support whole-farm redesign but they implemented it more gradually in the herd, which induced no major imbalance in herd demographics. In pathway 3, farmers predefined a relatively well-known rotational crossbreeding scheme to correct fertility issues of purebred cows without any changes at the farm level. They implemented it quickly from the beginning and had to adapt herd replacement and culling to regulate imbalance in herd demographics induced by the improvement in cow fertility. These first empirical evidences on how dairy farmers manage the transition from a purebred to rotational crossbred herd provide original scientific and operational contributions.     

Milk protein composition in purebred Holsteins and in first/second-generation crossbred cows from Swedish Red,Montbeliarde and Brown Swiss bulls

The aim of this study was to analyze milk protein composition in purebred and crossbred dairy cattle and estimate the effects of individual sources of variation on the investigated traits. Milk samples were collected from 505 cows from three commercial farms located in Northern Italy, some of which had originated from crossbreeding programs, although most were purebred Holsteins (HO). The basic crossbreeding scheme was a three-breed rotational system using Swedish Red (SR) semen on HO cows (SR×HO), Montbeliarde (MO) semen on SR×HO cows (MO×(SR×HO)) and HO semen again on MO×(SR×HO) cows. A smaller number of purebred HO from each of the herds were mated inverting the breed order (MO×HO and SR×(MO×HO)) or using Brown Swiss (BS) bulls (BS×HO) then MO bulls (MO×(BS×HO)). Milk samples were analyzed by reverse-phase HPLC to obtain protein fraction amounts (g/l) and proportions (% of total true protein). Traits were analyzed using a linear model, which included the fixed effects of herd-test-day (HTD), parity, days in milk and breed combination. Results showed that milk protein fractions were influenced by HTD, stage of lactation, parity and breed combination. The increase in protein concentration during lactation was due in particular to β-casein (β-CN), α_S1-CN and β-lactoglobulin (β-LG). The higher protein content of primiparous milk was mainly due to higher concentrations of all casein fractions. The milk from crossbred cows had higher contents and proportions of κ-CN and α-lactalbumin (α-LA), lower proportions of β-LG and greater proportion of caseins/smaller in whey proteins on milk true protein than purebred HO. The three-way crossbreds differed from two-way crossbreds only in having greater proportions of α-LA in their milk. Of the three-way crossbreds, the SR sired cows yielded milk with a smaller content and proportion of β-LG than the MO sired cows, and, consequently, a higher proportion of caseins than whey proteins. Results from this study support the feasibility of using crossbreeding programs to alter milk protein profiles with the aim of improving milk quality and cheese-making properties.     

Heterogeneous variances and genetics by environment interactions in genetic evaluation of crossbred lambs

Accounting for environmental heteroscedasticity and genetics by environment interaction (G×E) in genetic evaluation is important because animals may not perform predictably across environments. The objectives of this study were to evaluate the presence and consequences of heteroscedasticity and G×E on genetic evaluation. The population considered was crossbred lambs sired by terminal sires and reared under commercial conditions in the UK. Data on 6325 lambs sired by Charollais, Suffolk and Texel rams were obtained. The experiment was conducted between 1999 and 2002 on three farms located in England, Scotland and Wales. There were 2322, 2137 and 1866 lambs in England, Scotland and Wales, respectively. A total of 89 sires were mated to 1984 ewes of two types (Welsh and Scottish Mules). Most rams were used for two breeding seasons with some rotated among farms to create genetic links. Lambs were reared on pasture and had their parentage, birth, 5 week, 10 week, and slaughter weights recorded. Lambs were slaughtered at a constant fatness, at which they were ultrasonically scanned for fat and muscle depth. Heteroscedasticity was evaluated in two ways. First, data were separated into three subsets by farm. Within-farm variance component estimates were then compared with those derived from the complete data (Model 1). Second, the combined data were fitted, but with a heterogeneous (by farm) environmental variance structure (Model 2). To investigate G×E, a model with a random farm by sire (F×S) interaction was used (Model 3). The ratio of the F×S variance to total variance was a measure of the level of G×E in the population. Heterogeneity in environmental variability across farm was identified for all traits (P<0.01). Rank correlations of sire estimated breeding value between farms differed for Model 1 for all traits. However, sires ranked similarly (rank correlation of 0.99) for weight traits with Model 2, but less so for ultrasonic measures. Including the F×S interaction (Model 3) improved model fit for all traits. However, the F×S term explained a small proportion of variation in weights (<2%) although more in ultrasonic traits (at least 10%). In conclusion, heteroscedasticity and G×E were not large for these data, and can be ignored in genetic evaluation of weight but, perhaps, not ultrasonic traits. Still, before incorporating heteroscedasticity and G×E into routine evaluations of even ultrasonic traits, their consequences on selection response in the breeding goal should be evaluated.     

Comparison of joint versus purebred genomic evaluation in the French multi-breed dairy goat population

Background

All progeny-tested bucks from the two main French dairy goat breeds (Alpine and Saanen) were genotyped with the Illumina goat SNP50 BeadChip. The reference population consisted of 677 bucks and 148 selection candidates. With the two-step approach based on genomic best linear unbiased prediction (GBLUP), prediction accuracy of candidates did not outperform that of the parental average. We investigated a GBLUP method based on a single-step approach, with or without blending of the two breeds in the reference population.

Methods

Three models were used: (1) a multi-breed model, in which Alpine and Saanen breeds were considered as a single breed; (2) a within-breed model, with separate genomic evaluation per breed; and (3) a multiple-trait model, in which a trait in the Alpine was assumed to be correlated to the same trait in the Saanen breed, using three levels of between-breed genetic correlations (ρ): ρ = 0, ρ = 0.99, or estimated ρ. Quality of genomic predictions was assessed on progeny-tested bucks, by cross-validation of the Pearson correlation coefficients for validation accuracy and the regression coefficients of daughter yield deviations (DYD) on genomic breeding values (GEBV). Model-based estimates of average accuracy were calculated on the 148 candidates.

Results

The genetic correlations between Alpine and Saanen breeds were highest for udder type traits, ranging from 0.45 to 0.76. Pearson correlations with the single-step approach were higher than previously reported with a two-step approach. Correlations between GEBV and DYD were similar for the three models (within-breed, multi-breed and multiple traits). Regression coefficients of DYD on GEBV were greater with the within-breed model and multiple-trait model with ρ = 0.99 than with the other models. The single-step approach improved prediction accuracy of candidates from 22 to 37% for both breeds compared to the two-step method.

Conclusions

Using a single-step approach with GBLUP, prediction accuracy of candidates was greater than that based on parent average of official evaluations and accuracies obtained with a two-step approach. Except for regression coefficients of DYD on GEBV, there were no significant differences between the three models.     

Genomic and proteomic evaluation of antibiotic resistance in Salmonella strains

Using Salmonella strains identical to those present in the gastrointestinal tract of different animals we aim to determine and compare the proteome of two serotypes, Salmonella Typhimurium and Enteritidis recovered from faecal samples of wild boars and wild rabbits, respectively. The presence of genes responsible for antibiotic resistance was detected by PCR. Proteomes of the two distinct serotypes were determined using 2-DE in order to identify proteins associated with antibiotic resistance or virulence. Through 2-DE we obtained a total of 229 spots from both strains. All were suitable for MALDI-TOF/TOF and, in correlation with bioinformatic databases, allowed accurate identification and characterization of proteins. S. Enteritidis recovered from wild rabbits was sensitive to all the antibiotics tested in contrast to S. Typhimurium isolated from wild boars which presented a resistance phenotype to ampicillin, streptomycin and chloramphenicol. Nevertheless, despite the different ratio of proteins observed in each proteome according to their biological function, no significant difference was observed in the involvement of these proteins in pathogenicity. Bearing in mind that serotypes are related to infectious processes in humans and animals, it is important to explore the proteome of new strains which might serve as protein biomarkers for biological activity.     

Genomic and functional evaluation of TNFSF14 in multiple sclerosis susceptibility

Among multiple sclerosis (MS) susceptibility genes, the strongest non-human leukocyte antigen (HLA) signal in the Italian population maps to the TNFSF14 gene encoding LIGHT, a glycoprotein involved in dendritic cell (DC) maturation. Through fine-mapping in a large Italian dataset (4,198 patients with MS and 3,903 controls), we show that the TNFSF14 intronic SNP rs1077667 is the primarily MS-associated variant in the region. Expression quantitative trait locus (eQTL) analysis indicates that the MS risk allele is significantly associated with reduced TNFSF14 messenger RNA levels in blood cells, which is consistent with the allelic imbalance in RNA-Seq reads (P < 0.0001). The MS risk allele is associated with reduced levels of TNFSF14 gene expression (P < 0.01) in blood cells from 84 Italian patients with MS and 80 healthy controls (HCs). Interestingly, patients with MS are lower expressors of TNFSF14 compared to HC (P < 0.007). Individuals homozygous for the MS risk allele display an increased percentage of LIGHT-positive peripheral blood myeloid DCs (CD11c⁺, P = 0.035) in 37 HCs, as well as in in vitro monocyte-derived DCs from 22 HCs (P = 0.04). Our findings suggest that the intronic variant rs1077667 alters the expression of TNFSF14 in immune cells, which may play a role in MS pathogenesis.     

Model comparison for genetic evaluation of milk yield in crossbred Holsteins in the tropics

The objective of this study was to compare models for appropriate genetic parameter estimation for milk yield (305-day) in crossbred Holsteins in the tropics, where only records from crossbred cows were available. Eleven models with different effects of contemporary group (CG) at calving (herd-year-season or herd-year-month as fixed, and herd-year-month as random), age at calving (as linear or quadratic covariates, age-class, and age-class x lactation), and dominance were considered. On-farm records from small herds (n < 50) were included or excluded to validate the parameter estimates. Average Information Restricted Maximum Likelihood (AIREML) and Best Linear Unbiased Prediction (BLUP) were used to estimate variance components and breeding values. R-square (R2) and standard error of heritability (h2) were used to determine the appropriate model. The estimates of heritability from most models ranged from 0.18 to 0.22. CG formation of herd-year-month as a random effect slightly lowered the additive genetic variance but considerably decreased the permanent environmental variance. The model with age-class x lactation gave better R2 than other age adjustments. The models including records from smallholders gave similar estimates of heritability and a lower standard error than the models excluding them. The estimate of dominance variance as a proportion of total variance was close to zero. The low ratio of dominance to additive genetic variance suggested that the inclusion of dominance effects in the model was unjustified. In conclusion, the model including the effects of herd-year-month, age-class x lactation, as well as additive genetic, permanent environmental and residual effects, was the most appropriate for genetic evaluation in crossbred Holsteins, where records from smallholders could be included.