Economic values of production and functional traits,including residual feed intake,in Finnish milk production

Improving the feed efficiency of dairy cattle has a substantial effect on the economic efficiency and on the reduction of harmful environmental effects of dairy production through lower feeding costs and emissions from dairy farming. To assess the economic importance of feed efficiency in the breeding goal for dairy cattle, the economic values for the current breeding goal traits and the additional feed efficiency traits for Finnish Ayrshire cattle under production circumstances in 2011 were determined. The derivation of economic values was based on a bioeconomic model in which the profit of the production system was calculated, using the generated steady state herd structure. Considering beef production from dairy farms, 2 marketing strategies for surplus calves were investigated: (A) surplus calves were sold at a young age and (B) surplus calves were fattened on dairy farms. Both marketing strategies were unprofitable when subsidies were not included in the revenues. When subsidies were taken into account, a positive profitability was observed in both marketing strategies. The marginal economic values for residual feed intake (RFI) of breeding heifers and cows were −25.5 and −55.8 €/kg of dry matter per day per cow and year, respectively. The marginal economic value for RFI of animals in fattening was −29.5 €/kg of dry matter per day per cow and year. To compare the economic importance among traits, the standardized economic weight of each trait was calculated as the product of the marginal economic value and the genetic standard deviation; the standardized economic weight expressed as a percentage of the sum of all standardized economic weights was called relative economic weight. When not accounting for subsidies, the highest relative economic weight was found for 305-d milk yield (34% in strategy A and 29% in strategy B), which was followed by protein percentage (13% in strategy A and 11% in strategy B). The third most important traits were calving interval (9%) and mature weight of cows (11%) in strategy A and B, respectively. The sums of the relative economic weights over categories for RFI were 6 and 7% in strategy A and B, respectively. Under production conditions in 2011, the relative economic weights for the studied feed efficiency traits were low. However, it is possible that the relative importance of feed efficiency traits in the breeding goal will increase in the future due to increasing requirements to mitigate the environmental impact of milk production.     

Neglect of lactation stage leads to naive assessment of residual feed intake in dairy cattle

Residual feed intake (RFI) is a candidate trait for feed efficiency in dairy cattle. We investigated the influence of lactation stage on the effect of energy sinks in defining RFI and the genetic parameters for RFI across lactation stages for primiparous dairy cattle. Our analysis included 747 primiparous Holstein cows, each with recordings on dry matter intake (DMI), milk yield, milk composition, and body weight (BW) over 44 lactation weeks. For each individual cow, energy-corrected milk (ECM), metabolic BW (MBW), and change in BW (ΔBW) were calculated in each week of lactation and were taken as energy sinks when defining RFI. Two RFI models were considered in the analyses; RFI model [1] was a 1-step RFI model with constant partial regression coefficients of DMI on energy sinks (ECM, MBW, and ΔBW) over lactation. In RFI model [2], data from 44 lactation weeks were divided into 11 consecutive lactation periods of 4 wk in length. The RFI model [2] was identical to model [1] except that period-specific partial regressions of DMI on ECM, MBW, and ΔBW in each lactation period were allowed across lactation. We estimated genetic parameters for RFI across lactation by both models using a random regression method. Using RFI model [2], we estimated the period-specific effects of ECM, MBW, and ΔBW on DMI in all lactation periods. Based on results from RFI model [2], the partial regression coefficients of DMI on ECM, MBW, and ΔBW differed across lactation in RFI. Constant partial regression coefficients of DMI on energy sinks over lactation was not always sufficient to account for the effects across lactation and tended to give roughly average information from all period-specific effects. Heritability for RFI over 44 lactation weeks ranged from 0.10 to 0.29 in model [1] and from 0.10 to 0.23 in model [2]. Genetic variance and heritability estimates for RFI from model [2] tended to be slightly lower and more stable across lactation than those from model [1]. In both models, RFI was genetically different over lactation, especially between early and later lactation stages. Genetic correlation estimates for RFI between early and later lactation tended to be higher when using model [2] compared with model [1]. In conclusion, partial regression coefficients of DMI on energy sinks differed across lactation when modeling RFI. Neglect of lactation stage when defining RFI could affect the assessment of RFI and the estimation of genetic parameters for RFI across lactation.     

Reliability of breeding values for feed intake and feed efficiency traits in dairy cattle: When dry matter intake recordings are sparse under different scenarios

Currently, routine recordings of dry matter intake (DMI) in commercial herds are practically nonexistent. Recording DMI from commercial herds is a prerequisite for the inclusion of feed efficiency (FE) traits in dairy cattle breeding goals. To develop future on-farm phenotyping strategies, recording strategies that are low cost and less demanding logistically and that give relatively accurate estimates of the animal's genetic merit are therefore needed. The objectives of this study were (1) to estimate genetic parameters for daily DMI and FE traits and use the estimated parameters to simulate daily DMI phenotypes under different DMI recording scenarios (SCN) and (2) to use the simulated data to estimate for different scenarios the associated reliability of estimated breeding value and accuracies of genomic prediction for varying sizes of reference populations. Five on-farm daily DMI recording scenarios were simulated: once weekly (SCN1), once monthly (SCN2), every 2 mo (SCN3), every 3 mo (SCN4), and every 4 mo (SCN5). To estimate reliability of estimated breeding values, DMI and FE observations and true breeding values were simulated based on variance components estimated for daily observations of Nordic Red cows. To emulate realistic on-farm recording, 5 data set replicates, each with 36,037 DMI and FE records, were simulated for real pedigree and data structure of 789 Holstein cows. Observations for the 5 DMI recording scenarios were generated by discarding data in a step-wise manner from the full simulated data per the scenario's definitions. For each of these scenarios, reliabilities were calculated as correlation between the true and estimated breeding values. Variance components and genetic parameters were estimated for daily DMI, residual feed intake (RFI), and energy conversion efficiency (ECE) fitting the random regression model. Data for variance components were from 227 primiparous Nordic Red dairy cows covering 8 to 280 d in milk. Lactation-wise heritability for DMI, RFI, and ECE was 0.33, 0.12, and 0.32, respectively, and daily heritability estimates during lactation ranged from 0.18 to 0.45, 0.08 to 0.32, and 0.08 to 0.45 for DMI, RFI, and ECE, respectively. Genetic correlations for DMI between different stages of lactation ranged from ?0.50 to 0.99. The comparison of different on-farm DMI recording scenarios indicated that adopting a less-frequent recording scenario (SCN3) gave a similar level of accuracy as SCN1 when 17 more daughters are recorded per sire over the 46 needed for SCN1. Such a strategy is less demanding logistically and is low cost because fewer observations need to be collected per animal. The accuracy of genomic predictions associated with the 5 recording scenarios indicated that setting up a relatively larger reference population and adopting a less-frequent DMI sampling scenario (e.g., SCN3) is promising. When the same reference population size was considered, the genomic prediction accuracy of SCN3 was only 5.0 to 7.0 percentage points lower than that for the most expensive DMI recording strategy (SCN1). We concluded that DMI recording strategies that are sparse in terms of records per cow but with slightly more cows recorded per sire are advantageous both in genomic selection and in traditional progeny testing schemes when accuracy, logistics, and cost implications are considered.     

Depression in feed intake by a highly fermentable diet is related to plasma insulin concentration and insulin response to glucose infusion

The effects of dietary starch fermentability on feed intake and nutrient digestibility were evaluated in a crossover study, which was also designed to find factors that predict individual variation in feed intake response to starch fermentability. Thirty-two multiparous Holstein cows (121 ± 48 d in milk, 44 ± 7 kg/d of milk yield; mean ± SD) were randomly assigned to treatment sequence and were fed a diet intermediate to the treatments during a preliminary 28-d period. Treatments were dry ground corn grain and high-moisture corn harvested from the same field. Treatment periods were 14 d, with the final 4 d used for data and sample collection. Diets included corn silage and alfalfa haylage at a 2:1 ratio and were 26% neutral detergent fiber, 17% crude protein, 32% starch, and 3.5% fatty acids. High-moisture corn decreased dry matter intake (DMI) by 8%, but did not significantly alter digestible DMI. Individual DMI responses were highly variable, and variables from preliminary plasma analyses, propionate challenge tests, glucose tolerance tests, and hepatic mRNA analysis were assessed as potential predictors of DMI response to increased dietary starch fermentability. Of the covariates tested, only preliminary plasma insulin concentration and insulin response to glucose infusion were significant predictors of DMI response. High preliminary plasma insulin concentration was correlated with greater depression in DMI with increased fermentability; conversely, greater insulin secretion in response to glucose infusion was associated with minimal depression in DMI. These insulin variables were not significantly correlated. Consistent with past results, increased dietary starch fermentability decreased DMI. Significant correlations between insulin variables and individual DMI responses may warrant further investigation.     

Genetic parameters of coagulation properties, milk yield, quality, and acidity estimated using coagulating and noncoagulating milk information in Brown Swiss and Holstein-Friesian cows

The aim of this study was to estimate heritabilities of rennet coagulation time (RCT) and curd firmness (a₃₀) and their genetic correlations with test-day milk yield, composition (fat, protein, and casein content), somatic cell score, and acidity (pH and titratable acidity) using coagulating and noncoagulating (NC) milk information. Data were from 1,025 Holstein-Friesian (HF) and 1,234 Brown Swiss (BS) cows, which were progeny of 54 HF and 58 BS artificial insemination sires, respectively. Milk coagulation properties (MCP) of each cow were measured once using a computerized renneting meter and samples not exhibiting coagulation within 31 min after rennet addition were classified as NC milk. For NC samples, RCT was unobserved. Multivariate analyses, using Bayesian methodology, were performed to estimate the genetic relationships of RCT or a₃₀ with the other traits and statistical inference was based on the marginal posterior distributions of parameters of concern. For analyses involving RCT, a right-censored Gaussian linear model was used and records of NC milk samples, being censored records, were included as unknown parameters in the model implementing a data augmentation procedure. Rennet coagulation time was more heritable [heritability (h²) = 0.240 and h² = 0.210 for HF and BS, respectively] than a₃₀ (h² = 0.148 and h² = 0.168 for HF and BS, respectively). Milk coagulation properties were more heritable than a single test-day milk yield (h² = 0.103 and h² = 0.097 for HF and BS, respectively) and less heritable than milk composition traits whose heritability ranged from 0.275 to 0.275, with the only exception of fat content of BS milk (h² = 0.108). A negative genetic correlation, lower than −0.85, was estimated between RCT and a₃₀ for both breeds. Genetic relationships of MCP with yield and composition were low or moderate and favorable. The genetic correlation of somatic cell score with RCT in BS cows was large and positive and even more positive were those of RCT with pH and titratable acidity in both breeds, ranging from 0.80 to 0.94. Including NC milk information in the data affected the estimated correlations and decreased the uncertainty associated with the estimation process. On the basis of the estimated heritabilities and genetic correlations, enhancement of MCP through selective breeding with no detrimental effects on yield and composition seems feasible in both breeds. Milk acidity may play a role as an indicator trait for indirect enhancement of MCP.     

The Effect of Silage and Concentrate Type on Intake Behavior, Rumen Function, and Milk Production in Dairy Cows in Early and Late Lactation

The objective of this experiment was to evaluate the effect of feeding total mixed rations (TMR) that differ in structural and nonstructural carbohydrates to dairy cows in early and late lactation on short-term feed intake, dry matter intake (DMI), rumen fermentation variables, and milk yield. A 5 × 5 Latin square experiment with 15 dairy cows was repeated during early and late lactation. The 5 treatments were a TMR with (all on dry matter basis) 55% roughage (a 50:50 mixture of corn silage and grass silage) and 45% concentrate (a 50:50 mixture of concentrate rich in structural carbohydrates and concentrate rich in nonstructural carbohydrates; treatment CON), a TMR with the concentrate mixture and 55% grass silage (RGS) or 55% corn silage (RCS), and a TMR with the roughage mixture and 45% of the concentrate rich in structural carbohydrates (CSC) or the concentrate rich in nonstructural carbohydrates (CNS). Meal criteria, determined using the Gaussian-Gaussian-Weibull method per animal per treatment, showed an interaction between lactation stage and treatment. Feed intake behavior variables were therefore calculated with meal criteria per treatment-lactation stage combination. Differences in feed intake behavior were more pronounced between treatments differing in roughage composition than between treatments differing in concentrate composition, probably related to larger differences in chemical composition and particle size between corn silage and grass silage than between the 2 concentrates. The number of meals was similar between treatments, but eating time was greater in RGS (227 min/d) and lesser in RCS (177 min/d) than the other treatments. Intake rate increased when the amount of grass silage decreased, whereas meal duration decreased simultaneously. These effects were in line with a decreased DMI of the RGS diet vs. the other treatments, probably related to the high neutral detergent fiber (NDF) content. However, this effect was not found in CSC, although NDF content of the TMR, fractional clearance rate of NDF, and fractional degradation rate of NDF was similar between CSC and RGS. Rumen fluid pH was lesser, and molar proportions of acetic acid and of propionic acid were lesser and greater, respectively, in RCS compared with all other diets. Milk production did not differ between treatments. There was no effect of type of concentrate on milk composition, but diet RCS resulted in a lesser milk fat content and greater milk protein content than diet RGS. Lactation stage did affect short-term feed intake behavior and DMI, although different grass silages were fed during early and late lactation. The results indicate that short-term feed intake behavior is related to DMI and therefore may be a helpful tool in optimizing DMI and milk production in high-production dairy cows.     

Genetic parameters for milk urea nitrogen in relation to milk production traits

The aim of this study was to estimate genetic parameters for test-day milk urea nitrogen (MUN) and its relationships with milk production traits. Three test-day morning milk samples were collected from 1,953 Holstein-Friesian heifers located on 398 commercial herds in the Netherlands. Each sample was analyzed for somatic cell count, net energy concentration, MUN, and the percentage of fat, protein, and lactose. Genetic parameters were estimated using an animal model with covariates for days in milk and age at first calving, fixed effects for season of calving and effect of test or proven bull, and random effects for herd-test day, animal, permanent environment, and error. Coefficient of variation for MUN was 33%. Estimated heritability for MUN was 0.14. Phenotypic correlation of MUN with each of the milk production traits was low. The genetic correlation was close to zero for MUN and lactose percentage (−0.09); was moderately positive for MUN and net energy concentration of milk (0.19), fat yield (0.41), protein yield (0.38), lactose yield (0.22), and milk yield (0.24), and percentage of fat (0.18), and percentage of protein (0.27); and was high for MUN and somatic cell score (0.85). Herd-test day explained 58% of the variation in MUN, which suggests that management adjustments at herd-level can reduce MUN. This study shows that it is possible to influence MUN by herd practice and by genetic selection.     

Random Forests approach for identifying additive and epistatic single nucleotide polymorphisms associated with residual feed intake in dairy cattle

Feed efficiency is an economically important trait in the beef and dairy cattle industries. Residual feed intake (RFI) is a measure of partial efficiency that is independent of production level per unit of body weight. The objective of this study was to identify significant associations between single nucleotide polymorphism (SNP) markers and RFI in dairy cattle using the Random Forests (RF) algorithm. Genomic data included 42,275 SNP genotypes for 395 Holstein cows, whereas phenotypic measurements were daily RFI from 50 to 150 d postpartum. Residual feed intake was defined as the difference between an animal’s feed intake and the average intake of its cohort, after adjustment for year and season of calving, year and season of measurement, age at calving nested within parity, days in milk, milk yield, body weight, and body weight change. Random Forests is a widely used machine-learning algorithm that has been applied to classification and regression problems. By analyzing the tree structures produced within RF, the 25 most frequent pairwise SNP interactions were reported as possible epistatic interactions. The importance scores that are generated by RF take into account both main effects of variables and interactions between variables, and the most negative value of all importance scores can be used as the cutoff level for declaring SNP effects as significant. Ranking by importance scores, 188 SNP surpassed the threshold, among which 38 SNP were mapped to RFI quantitative trait loci (QTL) regions reported in a previous study in beef cattle, and 2 SNP were also detected by a genome-wide association study in beef cattle. The ratio of number of SNP located in RFI QTL to the total number of SNP in the top 188 SNP chosen by RF was significantly higher than in all 42,275 whole-genome markers. Pathway analysis indicated that many of the top 188 SNP are in genomic regions that contain annotated genes with biological functions that may influence RFI. Frequently occurring ancestor-descendant SNP pairs can be explored as possible epistatic effects for further study. The importance scores generated by RF can be used effectively to identify large additive or epistatic SNP and informative QTL. The consistency in results of our study and previous studies in beef cattle indicates that the genetic architecture of RFI in dairy cattle might be similar to that of beef cattle.     

Evaluating the performance of machine learning methods and variable selection methods for predicting difficult-to-measure traits in Holstein dairy cattle using milk infrared spectral data

Fourier-transform infrared (FTIR) spectroscopy is a powerful high-throughput phenotyping tool for predicting traits that are expensive and difficult to measure in dairy cattle. Calibration equations are often developed using standard methods, such as partial least squares (PLS) regression. Methods that employ penalization, rank-reduction, and variable selection, as well as being able to model the nonlinear relations between phenotype and FTIR, might offer improvements in predictive ability and model robustness. This study aimed to compare the predictive ability of 2 machine learning methods, namely random forest (RF) and gradient boosting machine (GBM), and penalized regression against PLS regression for predicting 3 phenotypes differing in terms of biological meaning and relationships with milk composition (i.e., phenotypes measurable directly and not directly in milk, reflecting different biological processes which can be captured using milk spectra) in Holstein-Friesian cattle under 2 cross-validation scenarios. The data set comprised phenotypic information from 471 Holstein-Friesian cows, and 3 target phenotypes were evaluated: (1) body condition score (BCS), (2) blood β-hydroxybutyrate (BHB, mmol/L), and (3) κ-casein expressed as a percentage of nitrogen (κ-CN, % N). The data set was split considering 2 cross-validation scenarios: samples-out random in which the population was randomly split into 10-folds (8-folds for training and 1-fold for validation and testing); and herd/date-out in which the population was randomly assigned to training (70% herd), validation (10%), and testing (20% herd) based on the herd and date in which the samples were collected. The random grid search was performed using the training subset for the hyperparameter optimization and the validation set was used for the generalization of prediction error. The trained model was then used to assess the final prediction in the testing subset. The grid search for penalized regression evidenced that the elastic net (EN) was the best regularization with increase in predictive ability of 5%. The performance of PLS (standard model) was compared against 2 machine learning techniques and penalized regression using 2 cross-validation scenarios. Machine learning methods showed a greater predictive ability for BCS (0.63 for GBM and 0.61 for RF), BHB (0.80 for GBM and 0.79 for RF), and κ-CN (0.81 for GBM and 0.80 for RF) in samples-out cross-validation. Considering a herd/date-out cross-validation these values were 0.58 (GBM and RF) for BCS, 0.73 (GBM and RF) for BHB, and 0.77 (GBM and RF) for κ-CN. The GBM model tended to outperform other methods in predictive ability around 4%, 1%, and 7% for EN, RF, and PLS, respectively. The prediction accuracies of the GBM and RF models were similar, and differed statistically from the PLS model in samples-out random cross-validation. Although, machine learning techniques outperformed PLS in herd/date-out cross-validation, no significant differences were observed in terms of predictive ability due to the large standard deviation observed for predictions. Overall, GBM achieved the highest accuracy of FTIR-based prediction of the different phenotypic traits across the cross-validation scenarios. These results indicate that GBM is a promising method for obtaining more accurate FTIR-based predictions for different phenotypes in dairy cattle.     

Genetic parameters for major milk fatty acids and milk production traits of Dutch Holstein-Friesians

The objective of this study was to estimate genetic parameters for major milk fatty acids and milk production traits. One morning milk sample was collected from 1,918 Holstein-Friesian heifers located in 398 commercial herds in the Netherlands. Each sample was analyzed for total percentages of fat and protein, and for detailed fatty acid percentages (computed as fatty acid weight as a proportion of total fat weight). Intraherd heritabilities were high for C4:0 to C16:0, ranging from 0.42 for C4:0 to 0.71 for C10:0. Saturated and unsaturated C18 fatty acids had intraherd heritability estimates of approximately 0.25, except for C18:2 cis-9, trans-11, which was 0.42. Standard errors of the heritabilities were between 0.07 and 0.12. Genetic correlations were high and positive among C4:0 to C14:0, as well as among unsaturated C18, but correlations of C4:0 to C14:0 with unsaturated C18 were generally weak. The genetic correlation of C16:0 with fat percentage was positive (0.65), implying that selection for fat percentage should result in a correlated increase of C16:0, whereas unsaturated C18 fatty acids decreased with increasing fat percentage (−0.74). Milk fat composition can be changed by means of selective breeding, which offers opportunities to meet consumer demands regarding health and technological aspects.     

Short communication: Estimates of heritabilities and genetic correlations among milk fatty acid unsaturation indices in Canadian Holsteins

The objectives of the present study were to estimate genetic parameters of milk fatty acid unsaturation indices in Canadian Holsteins. Data were available on milk fatty acid composition of 2,573 Canadian Holstein cows from 46 commercial herds enrolled in the Québec Dairy Production Centre of Expertise, Valacta (Sainte-Anne-de-Bellevue, Quebec, Canada). Individual fatty acid percentages (g/100 g of total fatty acids) were determined for each milk sample by gas chromatography. The unsaturation indices were calculated as the ratio of an unsaturated fatty acid to the sum of that unsaturated fatty acid and its corresponding substrate fatty acid, multiplied by 100. A mixed linear model was fitted under REML for the statistical analysis of milk fatty acid unsaturation indices. The statistical model included the fixed effects of parity, age at calving, and stage of lactation, each nested within parity, and the random effects of herd-year-season of calving, animal, and residual. Estimates of heritabilities for the C14, C16, C18, conjugated linoleic acid, and total unsaturation indices were 0.48, 0.25, 0.29, 0.14, and 0.19, respectively. Phenotypic and genetic correlation estimates among unsaturation indices were all positive and ranged from 0.20 to 0.65 and 0.23 to 0.81, respectively. The estimates of heritabilities and genetic correlations for milk fatty acid unsaturation indices suggest that genetic variation exists among cows in milk fatty acid unsaturation, and the proportions of desirable unsaturated fatty acids from a human health point of view may be increased in bovine milk through genetic selection.     

Mitochondrial protein gene expression and the oxidative phosphorylation pathway associated with feed efficiency and energy balance in dairy cattle

Feed efficiency and energy balance are important traits underpinning profitability and environmental sustainability in animal production. They are complex traits, and our understanding of their underlying biology is currently limited. One measure of feed efficiency is residual feed intake (RFI), which is the difference between actual and predicted intake. Variation in RFI among individuals is attributable to the metabolic efficiency of energy utilization. High RFI (H_RFI) animals require more energy per unit of weight gain or milk produced compared with low RFI (L_RFI) animals. Energy balance (EB) is a closely related trait calculated very similarly to RFI. Cellular energy metabolism in mitochondria involves mitochondrial protein (MiP) encoded by both nuclear (NuMiP) and mitochondrial (MtMiP) genomes. We hypothesized that MiP genes are differentially expressed (DE) between H_RFI and L_RFI animal groups and similarly between negative and positive EB groups. Our study aimed to characterize MiP gene expression in white blood cells of H_RFI and L_RFI cows using RNA sequencing to identify genes and biological pathways associated with feed efficiency in dairy cattle. We used the top and bottom 14 cows ranked for RFI and EB out of 109 animals as H_RFI and L_RFI, and positive and negative EB groups, respectively. The gene expression counts across all nuclear and mitochondrial genes for animals in each group were used for differential gene expression analyses, weighted gene correlation network analysis, functional enrichment, and identification of hub genes. Out of 244 DE genes between RFI groups, 38 were MiP genes. The DE genes were enriched for the oxidative phosphorylation (OXPHOS) and ribosome pathways. The DE MiP genes were underexpressed in L_RFI (and negative EB) compared with the H_RFI (and positive EB) groups, suggestive of reduced mitochondrial activity in the L_RFI group. None of the MtMiP genes were among the DE MiP genes between the groups, which suggests a non-rate limiting role of MtMiP genes in feed efficiency and warrants further investigation. The role of MiP, particularly the NuMiP and OXPHOS pathways in RFI, was also supported by our gene correlation network analysis and the hub gene identification. We validated the findings in an independent data set. Overall, our study suggested that differences in feed efficiency in dairy cows may be linked to differences in cellular energy demand. This study broadens our knowledge of the biology of feed efficiency in dairy cattle.     

Genetic heterogeneity of feed intake,energy-corrected milk,and body weight across lactation in primiparous Holstein,Nordic Red,and Jersey cows

In this study, we aimed to estimate and compare the genetic parameters of dry matter intake (DMI), energy-corrected milk (ECM), and body weight (BW) as 3 feed efficiency–related traits across lactation in 3 dairy cattle breeds (Holstein, Nordic Red, and Jersey). The analyses were based on weekly records of DMI, ECM, and BW per cow across lactation for 842 primiparous Holstein cows, 746 primiparous Nordic Red cows, and 378 primiparous Jersey cows. A random regression model was applied to estimate variance components and genetic parameters for DMI, ECM, and BW in each lactation week within each breed. Phenotypic means of DMI, ECM, and BW observations across lactation showed to be in very similar patterns between breeds, whereas breed differences lay in the average level of DMI, ECM, and BW. Generally, for all studied breeds, the heritability for DMI ranged from 0.2 to 0.4 across lactation and was in a range similar to the heritability for ECM. The heritability for BW ranged from 0.4 to 0.6 across lactation, higher than the heritability for DMI or ECM. Among the studied breeds, the heritability estimates for DMI shared a very similar range between breeds, whereas the heritability estimates for ECM tended to be different between breeds. For BW, the heritability estimates also tended to follow a similar range between breeds. Among the studied traits, the genetic variance and heritability for DMI varied across lactation, and the genetic correlations between DMI at different lactation stages were less than unity, indicating a genetic heterogeneity of feed intake across lactation in dairy cattle. In contrast, BW was the most genetically consistent trait across lactation, where BW among all lactation weeks was highly correlated. Genetic correlations between DMI, ECM, and BW changed across lactation, especially in early lactation. Energy-corrected milk had a low genetic correlation with both DMI and BW at the beginning of lactation, whereas ECM was highly correlated with DMI in mid and late lactation. Based on our results, genetic heterogeneity of DMI, ECM, and BW across lactation generally was observed in all studied dairy breeds, especially for DMI, which should be carefully considered for the recording strategy of these traits. The genetic correlations between DMI, ECM, and BW changed across lactation and followed similar patterns between breeds.     

Genetic parameters for milk coagulation properties in Estonian Holstein cows

The objective of this study was to estimate heritabilities and repeatabilities for milk coagulation traits [milk coagulation time (RCT) and curd firmness (E₃₀)] and genetic and phenotypic correlations between milk yield and composition traits (milk fat percentage and protein percentage, urea, somatic cell count, pH) in first-lactation Estonian Holstein dairy cattle. A total of 17,577 test-day records from 4,191 Estonian Holstein cows in 73 herds across the country were collected during routine milk recordings. Measurements of RCT and E₃₀ determined with the Optigraph (Ysebaert, Frepillon, France) are based on an optical signal in the near-infrared region. The cows had at least 3 measurements taken during the period from April 2005 to January 2009. Data were analyzed using a repeatability animal model. There was substantial variation in milk coagulation traits with a coefficient of variation of 27% for E₃₀ and 9% for the log-transformed RCT. The percentage of variation explained by herd was 3% for E₃₀ and 4% for RCT, suggesting that milk coagulation traits are not strongly affected by herd conditions (e.g., feeding). Heritability was 0.28 for RCT and 0.41 for E₃₀, and repeatability estimates were 0.45 and 0.50, respectively. Genetic correlation between both milk coagulation traits was negligible, suggesting that RCT and E₃₀ have genetically different foundations. Milk coagulation time had a moderately high positive genetic (0.69) and phenotypic (0.61) correlation with milk pH indicating that a high pH is related to a less favorable RCT. Curd firmness had a moderate positive genetic (0.48) and phenotypic (0.45) correlation with the protein percentage. Therefore, a high protein percentage is associated with favorable curd firmness. All reported genetic parameters were statistically significantly different from zero. Additional univariate random regression analysis for milk coagulation traits yielded slightly higher average heritabilities of 0.38 and 0.47 for RCT and E₃₀ compared with the heritabilities of the repeatability model.     

An attempt at predicting blood β-hydroxybutyrate from Fourier-transform mid-infrared spectra of milk using multivariate mixed models in Polish dairy cattle

Fourier transform mid-infrared (FT-MIR) spectra of milk are commonly used for phenotyping of traits of interest through links developed between the traits and milk FT-MIR spectra. Predicted traits are then used in genetic analysis for ultimate phenotypic prediction using a single-trait mixed model that account for cows' circumstances at a given test day. Here, this approach is referred to as indirect prediction (IP). Alternatively, FT-MIR spectral variable can be kept multivariate in the form of factor scores in REML and BLUP analyses. These BLUP predictions, including phenotype (predicted factor scores), were converted to single-trait through calibration outputs; this method is referred to as direct prediction (DP). The main aim of this study was to verify whether mixed modeling of milk spectra in the form of factors scores (DP) gives better prediction of blood β-hydroxybutyrate (BHB) than the univariate approach (IP). Models to predict blood BHB from milk spectra were also developed. Two data sets that contained milk FT-MIR spectra and other information on Polish dairy cattle were used in this study. Data set 1 (n = 826) also contained BHB measured in blood samples, whereas data set 2 (n = 158,028) did not contain measured blood values. Part of data set 1 was used to calibrate a prediction model (n = 496) and the remaining part of data set 1 (n = 330) was used to validate the calibration models, as well as to evaluate the DP and IP approaches. Dimensions of FT-MIR spectra in data set 2 were reduced either into 5 or 10 factor scores (DP) or into a single trait (IP) with calibration outputs. The REML estimates for these factor scores were found using WOMBAT. The BLUP values and predicted BHB for observations in the validation set were computed using the REML estimates. Blood BHB predicted from milk FT-MIR spectra by both approaches were regressed on reference blood BHB that had not been used in the model development. Coefficients of determination in cross-validation for untransformed blood BHB were from 0.21 to 0.32, whereas that for the log-transformed BHB were from 0.31 to 0.38. The corresponding estimates in validation were from 0.29 to 0.37 and 0.21 to 0.43, respectively, for untransformed and logarithmic BHB. Contrary to expectation, slightly better predictions of BHB were found when univariate variance structure was used (IP) than when multivariate covariance structures were used (DP). Conclusive remarks on the importance of keeping spectral data in multivariate form for prediction of phenotypes may be found in data sets where the trait of interest has strong relationships with spectral variables.