Economic weights for genetic improvement of lactation persistency and milk yield

This study aimed to establish a criterion for measuring the relative weight of lactation persistency (the ratio of yield at 280 d in milk to peak yield) in restricted selection index for the improvement of net merit comprising 3-parity total yield and total lactation persistency. The restricted selection index was compared with selection based on first-lactation total milk yield (I₁), the first-two-lactation total yield (I₂), and first-three-lactation total yield (I₃). Results show that genetic response in net merit due to selection on restricted selection index could be greater than, equal to, or less than that due to the unrestricted index depending upon the relative weight of lactation persistency and the restriction level imposed. When the relative weight of total lactation persistency is equal to the criterion, the restricted selection index is equal to the selection method compared (I₁, I₂, or I₃). The restricted selection index yielded a greater response when the relative weight of total lactation persistency was above the criterion, but a lower response when it was below the criterion. The criterion varied depending upon the restriction level (c) imposed and the selection criteria compared. A curvilinear relationship (concave curve) exists between the criterion and the restricted level. The criterion increases as the restriction level deviates in either direction from 1.5. Without prior information of the economic weight of lactation persistency, the imposition of the restriction level of 1.5 on lactation persistency would maximize change in net merit. The procedure presented allows for simultaneous modification of multi-parity lactation curves.     

Efficiency of different selection criteria for persistency and lactation milk yield

A conversion formula was developed to convert the genetic covariance matrices of daily yields and of random regression coefficients between 305-d and 335-d production periods under a random regression test day model. Five selection criteria were compared in terms of genetic improvement in persistency and lactation milk: 1) lactation estimated breeding value (EBVL), 2) P6 = 279Sigma(i=65) (D280 - Di), 3) ratio of daily estimated breeding value (EBV)(r280/65 = D280/D65), 4) ratio of partial lactation EBV (P280/65 = D66 approximately 28/D5 approximately 65), and 5) differential daily EBV (d65-280 = D65 - D280), where Di refers to EBV at days in milk (DIM) i. Fundamental differences among these 5 selection criteria were interpreted conceptually with a graph. Persistency, defined as k = (delta G65 - delta G280)/215, was the average daily rate of decline in selection gain from DIM 65 to 280, which is free from the effect of lactation milk on the rate of decline. Parameter k provides an objective measure of persistency, which increases when k < 0 and decreases when k > 0. Of the 5 selection criteria compared, d65-280 and P6 achieved greater persistency at the expense of genetic gain in lactation milk, whereas selection based on EBVL achieved the highest response in lactation milk, but was coupled with greatest decline in persistency. Selection on P280/65 or r280/65 improved both lactation milk and persistency and, thus, is recommended for simultaneous improvement of these 2 economically important traits. Further study of the relative economic values of persistency and lactation milk in order to combine both traits into an index for selection decision is warranted.     

Modifying the lactation curve to improve lactation milk and persistency

Daily, stage and lactation estimated breeding values (EBV) and the shape of the lactation curve for each cow are controlled by a unique set of random (genetic) regression coefficients under a test day model, thus providing a basis for genetic improvement of these characteristics. Three selection procedures were developed for simultaneous improvement of total lactation milk and persistency: 1) index selection based on daily EBV, 2) index selection based on stage EBV, and 3) index selection based on random regression (RR) coefficients. A numerical example was given to demonstrate the computation of indexes based on stage EBV and based on RR coefficients. A conversion equation was derived to convert between genetic changes in EBV and RR coefficients. Index selection based on daily EBV would require the finding of 305 weighting factors for a lactation period of 305 d, making it impractical to determine the weighting factors on a daily basis. Alternatively, a lactation period was partitioned into a few stages to facilitate the construction of index selection based on stage EBV and index selection based on RR coefficients. These selection procedures make use of the annual genetic gains routinely computed in national genetic evaluations to restrict the genetic gains between different lactation stages to achieve the desired curve. When there is no prior knowledge of annual genetic gains, the proportional restriction of genetic gains between stages may be used. In summary, this study provides a simple means of modifying the lactation curve by manipulating genetic changes in different lactation stages at a pre-specified rate.     

Genetic evaluation of lactation persistency for five breeds of dairy cattle

Cows with high lactation persistency tend to produce less milk than expected at the beginning of lactation and more than expected at the end. Best prediction of lactation persistency is calculated as a function of trait-specific standard lactation curves and linear regressions of test-day deviations on days in milk. Because regression coefficients are deviations from a tipping point selected to make yield and lactation persistency phenotypically uncorrelated it should be possible to use 305-d actual yield and lactation persistency to predict yield for lactations with later endpoints. The objectives of this study were to calculate (co)variance components and breeding values for best predictions of lactation persistency of milk (PM), fat (PF), protein (PP), and somatic cell score (PSCS) in breeds other than Holstein, and to demonstrate the calculation of prediction equations for 400-d actual milk yield. Data included lactations from Ayrshire, Brown Swiss, Guernsey (GU), Jersey (JE), and Milking Shorthorn (MS) cows calving since 1997. The number of sires evaluated ranged from 86 (MS) to 3,192 (JE), and mean sire estimated breeding value for PM ranged from 0.001 (Ayrshire) to 0.10 (Brown Swiss); mean estimated breeding value for PSCS ranged from −0.01 (MS) to −0.043 (JE). Heritabilities were generally highest for PM (0.09 to 0.15) and lowest for PSCS (0.03 to 0.06), with PF and PP having intermediate values (0.07 to 0.13). Repeatabilities varied considerably between breeds, ranging from 0.08 (PSCS in GU, JE, and MS) to 0.28 (PM in GU). Genetic correlations of PM, PF, and PP with PSCS were moderate and favorable (negative), indicating that increasing lactation persistency of yield traits is associated with decreases in lactation persistency of SCS, as expected. Genetic correlations among yield and lactation persistency were low to moderate and ranged from −0.55 (PP in GU) to 0.40 (PP in MS). Prediction equations for 400-d milk yield were calculated for each breed by regression of both 305-d yield and 305-d yield and lactation persistency on 400-d yield. Goodness-of-fit was very good for both models, but the addition of lactation persistency to the model significantly improved fit in all cases. Routine genetic evaluations for lactation persistency, as well as the development of prediction equations for several lactation end-points, may provide producers with tools to better manage their herds.     

Genetic markers for lactation persistency in primiparous Australian dairy cows

Good performance in extended lactations of dairy cattle may have a beneficial effect on food costs, health, and fertility. Because data for extended lactation performance is scarce, lactation persistency has been suggested as a suitable selection criterion. Persistency phenotypes were calculated in several ways: P1 was yield relative to an approximate peak, P2 was the slope after peak production, and P3 was a measure derived to be phenotypically uncorrelated to yield and calculated as a function of linear regressions on test-day deviations of days in milk. Phenotypes P1, P2, and P3 were calculated for sires as solutions estimated from a random regression model fitted to milk yield. Because total milk yield, calculated as the sum of daily sire solutions, was correlated to P1 and P2 (r = 0.30 and 0.35 for P1 and P2, respectively), P1 and P2 were also adjusted for milk yield (P1adj and P2adj, respectively). To find genomic regions associated with the persistency phenotypes, we used a discovery population of 743 Holstein bulls proven before 2005 and 2 validation data sets of 357 Holstein bulls proven after 2005 and 294 Jersey sires. Two strategies were used to search for genomic regions associated with persistency: 1) persistency solutions were regressed on each single nucleotide polymorphism (SNP) in turn and 2) a genomic selection method (BayesA) was used where all SNP were fitted simultaneously. False discovery rates in the validation data were high (>66% in Holsteins and >77% in Jerseys). However, there were 2 genomic regions on chromosome 6 that validated in both breeds, including a cluster of 6 SNP at around 13.5 to 23.7 Mbp and another cluster of 5 SNP (70.4 to 75.6 Mbp). A third cluster validated in both breeds on chromosome 26 (0.33 to 1.46 Mbp). Validating SNP effects across 2 breeds is unlikely to happen by chance even when false discovery rates within each breed are high. However, marker-assisted selection on these selected SNP may not be the best way to improve this trait because the average variation explained by validated SNP was only 1 to 2%. Genomic selection could be a better alternative. Correlations between genomic breeding values predicted using all SNP simultaneously and estimated breeding values based on progeny test were twice as high as the equivalent correlations between estimated breeding values and parent average. Persistency is a good candidate for genomic selection because the trait is expressed late in lactation.     

Development of an optimal index to improve lactation yield and persistency with the least selection intensity

It is important to have improvement in both lactation milk yield and persistency. Modification of the lactation curve requires severe restrictions on selection criteria designed to simultaneously improve both milk yield and persistency. As a result, manipulating the lactation curve for improved persistency requires higher selection intensity than unrestricted selection based on 305-d estimated breeding value (EBV). Our study showed that for a given restriction imposed on both milk EBV and persistency, it is possible to derive different indexes to achieve this selection constraint with different degrees of selection intensity. Of the class of indexes that meets the same restriction, it is preferable to choose the index that requires the least selection intensity because it is easier to achieve the selection goal with the use of an index that requires a lower selection intensity than a higher selection intensity. An optimal index based on random regression (RR) coefficients was developed to achieve the prespecified stage genetic gains with the lowest selection intensity. A conversion equation was derived to convert the selection index based on RR coefficients to the selection index based on stage EBV with the lowest selection intensity. A numerical example is provided to demonstrate the procedures developed compared with conventional selection based on 305-d milk EBV.     

Maximization of lactation milk production without decreasing persistency

This study treats each daily estimated breeding value (EBV) of the lactation as a separate trait to modify the lactation curve on a daily basis. Six selection strategies for improving lactation milk without decreasing persistency were compared: 1) index I_R1, subject to the restriction of equal genetic gains at days in milk (DIM) 60 and 280, 2) I_R2, subject to the restriction of zero gain at DIM 60, 3) desired gains index I_d, designed to increase lactation milk without altering the lactation curve, 4) index I_u, comprising lactation EBV and persistency without standardization, 5) index I_w, consisting of lactation EBV (EBV_L) and persistency with standardization, and 6) conventional selection on EBV_L and used as a basis for comparison. Of the 6 selection strategies compared, I_R2 yielded the greatest persistency, but achieved the smallest response in EBV_L, suggesting that it is impractical to increase persistency by inhibiting change in the peak yield. Index I_u showed the same response in lactation milk as conventional selection on EBV_L, but resulted in the same decreased persistency. Although both I_R1 and I_d achieved constant persistency, the former produced a greater lactation response (669 kg EBV) than the latter (560 kg EBV). Thus, I_R1 is a viable strategy for improving EBV_L while holding persistency constant. None of the 6 selection strategies excelled in both lactation milk and persistency. Index I_w appears to be a reasonable choice for improving both traits, although responses would depend on the relative economic importance of the 2 traits. Differential responses between I_u and I_w emphasize the need to weight the EBV of different traits by the inverse of their standard deviations in index construction when the EBV vary widely in variance. The general formula developed here provides a useful genetic means of modifying the lactation curve by restricting differential genetic gains among different days of the lactation.     

Genetic modification of the lactation curve by bending the eigenvectors of the additive genetic random regression coefficient matrix

The eigenvectors of the additive genetic random regression covariance (K) matrix contribute differentially to different parts of the lactation curve in response to genetic selection. It is, therefore, important to examine the genetic response patterns from the individual eigenvectors of the matrix K for the modification of the shape of the lactation curve. This study demonstrated a general methodology for imposing differential restrictions on different eigenvectors according to their effects on the shape of the lactation curve. A numerical example is given to illustrate the derivation and implementation of this procedure. Theoretically and experimentally, manipulating individual eigenvectors based on their individual effects on the shape of the lactation curve is more important than manipulating the joint effect of all the eigenvectors of K on the lactation curve. This described procedure provides a useful tool for simultaneous improvement of milk production and lactation persistency by modifying the shape of the lactation curve.     

Effects of extended voluntary waiting period from calving until first insemination on body condition,milk yield,and lactation persistency

A 1-yr calving interval (CInt) is usually associated with maximized milk output, due to the calving-related peak in milk yield. Extending CInt could benefit cow health and production efficiency due to fewer transition periods per unit of time. Extending CInt can affect lactation performance by fewer days dry per year, delayed pregnancy effect on milk yield, and greater milk solid yield in late lactation. This study first investigated the effects of 3 different voluntary waiting periods (VWP) from calving until first insemination on body weight, body condition, milk yield, and lactation persistency. Second, individual cow characteristics in early lactation were identified that contributed to milk yield and persistency of cows with different VWP. Holstein-Friesian dairy cows (n = 154) within 1 herd were blocked for parity, calving season, and expected milk yield. Cows were randomly assigned within the blocks to 1 of 3 VWP (50, 125, or 200 d: VWP50, VWP125, or VWP200, respectively) and monitored through 1 complete lactation and the first 6 wk of the subsequent lactation, or until culling. Minimum and mean CInt (384 vs. 452 vs. 501 d for VWP50 vs. VWP125 vs. VWP200) increased with increasing VWP, but maximum CInt was equal for the 3 VWP. Fat- and protein-corrected milk yield (FPCM) was analyzed weekly. Milk yield and FPCM were also expressed per day of CInt, to compare yields of cows with different VWP. Persistency was determined between d 100 and d 200 of the lactation, as well as between d 100 and dry-off. Values are presented as least squares means ± standard error of the mean. During the first 44 wk of lactation, VWP did not affect FPCM yield in both primiparous and multiparous cows. The VWP did not affect milk yield per day of CInt. The VWP did not affect FPCM yield per day of calving interval for primiparous cows. Multiparous cows in VWP125 had FPCM yield per day of CInt similar to that of VWP50. Multiparous cows in VWP200 had lower FPCM yield per day of CInt compared with VWP50 (27.2 vs. 30.4 kg/d). During the last 6 wk before dry-off, cows in VWP125 had lower yield compared with cows in VWP50, which could benefit their udder health in the dry period and after calving. Persistency was better for cows in VWP200 compared with cows in VWP50 (?0.05 vs. ?0.07 kg/d). Body weight was not different among VWP groups. Multiparous cows in VWP200 had a higher body condition score in the last 3 mo before dry-off and the first 6 wk of the next lactation, compared with multiparous cows in VWP125 and VWP50. The VWP could be extended from 50 d to 125 d without an effect on daily yield per day of calving interval. Extending VWP until 200 d for primiparous cows did not affect their daily milk yield, but multiparous cows with a 200-d VWP had a reduced milk yield per day of calving interval and an increased body condition in late lactation and the subsequent lactation, compared with multiparous cows with a 50-d VWP.     

Estimation of genetic parameters for milk yield across lactations in mixed-breed dairy goats

Currently, breeding values for dairy goats in the United Kingdom are not estimated and selection is based only on phenotypes. Several studies from other countries have applied various methodologies to estimate breeding values for milk yield in dairy goats. However, most of the previous analyses were based on relatively small data sets, which might have affected the accuracy of the parameter estimates. The objective of this study was to estimate genetic parameters for milk yield in crossbred dairy goats in lactations 1 to 4. The research was based on data provided by 2 commercial goat farms in the United Kingdom comprising 390,482 milk yield records on 13,591 dairy goats kidding between 1987 and 2012. The population was created by crossing 3 breeds: Alpine, Saanen, and Toggenburg. In each generation, the best-performing animals were selected for breeding and, as a result, a synthetic breed was created. The pedigree file contained 28,184 individuals, of which 2,414 were founders. The data set contained test-day records of milk yield, lactation number, farm, age at kidding, and year and season of kidding. Data on milk composition was unavailable. Covariance components were estimated with the average information REML algorithm in the ASReml package (VSN International Ltd., Hemel Hempstead, UK). A random regression animal model for milk yield with fixed effects of herd test day, year-season, and age at kidding was used. Heritability was the highest at 200 and 250 d in milk (DIM), reaching 0.45 in the first lactation and between 0.34 and 0.25 in subsequent lactations. After 300 DIM, the heritability started decreasing to 0.23 and 0.10 at 400 DIM in the first and subsequent lactations, respectively. Genetic correlation between milk yield in the first and subsequent lactations was between 0.16 and 0.88. This study found that milk yields in first and subsequent lactations are highly correlated, both at the genetic and phenotypic level. Estimates of heritability for milk yield were higher than most of the values reported in the literature, although they were in the range reported in this species. This should facilitate genetic improvement for the population studied as part of a broader multi-trait breeding program.     

Genetic parameters for milk production and persistency for Danish Holsteins estimated in random regression models using REML

(Co)variance components for milk, fat, and protein yield of 8075 first-parity Danish Holsteins (DH) were estimated in random regression models by REML. For all analyses, the fixed part of the model was held constant, whereas four different functions were applied to model the additive genetic effect and the permanent environment effect. Homogeneous residual variance was assumed throughout lactation. Univariate models were compared using a minimum of -2 ln(restricted likelihood) as the criterion for best fit. Heritabilities as a function of time were calculated from the estimated curve parameters from univariate analyses. Independent of the function applied and the trait in question, heritabilities were lowest in the beginning of the lactation. Heritabilities for persistency of fat yield were slightly higher than heritabilities for persistency of milk and protein yield. Genetic correlations between persistency and 305-d production were higher for protein and milk yield than for fat yield. Bivariate analyses between the production traits were carried out in sire models using the models with the best 3-parameter curve fit in the univariate analyses. Correlations between traits were calculated from covariance components for curve parameters estimated in bivariate analyses. Genetic correlations between milk and protein yield were higher than between milk and fat yield.     

Relationships between milk culture results and milk yield in Norwegian dairy cattle

Associations between test-day milk yield and positive milk cultures for Staphylococcus aureus, Streptococcus spp., and other mastitis pathogens or a negative milk culture for mastitis pathogens were assessed in quarter milk samples from randomly sampled cows selected without regard to current or previous udder health status. Staphylococcus aureus was dichotomized according to sparse (≤1,500 cfu/mL of milk) or rich (>1,500 cfu/mL of milk) growth of the bacteria. Quarter milk samples were obtained on 1 to 4 occasions from 2,740 cows in 354 Norwegian dairy herds, resulting in a total of 3,430 samplings. Measures of test-day milk yield were obtained monthly and related to 3,547 microbiological diagnoses at the cow level. Mixed model linear regression models incorporating an autoregressive covariance structure accounting for repeated test-day milk yields within cow and random effects at the herd and sample level were used to quantify the effect of positive milk cultures on test-day milk yields. Identical models were run separately for first-parity, second-parity, and third-parity or older cows. Fixed effects were days in milk, the natural logarithm of days in milk, sparse and rich growth of Staph. aureus (1/0), Streptococcus spp. (1/0), other mastitis pathogens (1/0), calving season, time of test-day milk yields relative to time of microbiological diagnosis (test day relative to time of diagnosis), and the interaction terms between microbiological diagnosis and test day relative to time of diagnosis. The models were run with the logarithmically transformed composite milk somatic cell count excluded and included. Rich growth of Staph. aureus was associated with decreased production levels in first-parity cows. An interaction between rich growth of Staph. aureus and test day relative to time of diagnosis also predicted a decline in milk production in third-parity or older cows. Interaction between sparse growth of Staph. aureus and test day relative to time of diagnosis predicted declining test-day milk yields in first-parity cows. Sparse growth of Staph. aureus was associated with high milk yields in third-parity or older cows after including the logarithmically transformed composite milk somatic cell count in the model, which illustrates that lower production levels are related to elevated somatic cell counts in high-producing cows. The same association with test-day milk yield was found among Streptococcus spp.-positive pluriparous cows.     

Genetic evaluation and best prediction of lactation persistency

Cows with high persistency tend to produce less milk than expected at the beginning of lactation and more than expected at the end. Best prediction of persistency was calculated as a function of a trait-specific standard lactation curve and a linear regression of test-day deviations on days in milk. Regression coefficients were deviations from a balance point to make yield and persistency phenotypically uncorrelated. The objectives of this study were to calculate (co)variance components and breeding values for best predictions of persistency of milk (PM), fat (PF), protein (PP), and SCS (PSCS) in Holstein cows. Data included 8,682,138 lactations from 4,375,938 cows calving since 1997, and 39,354 sires were evaluated. Sire estimated breeding values (EBV) for PM, PF, and PP were similar and ranged from −0.70 to 0.75 for PM; EBV for PSCS ranged from −0.37 to 0.28. Regressions of sire EBV on birth year were near zero (<0.003) but positive for PM, PF, and PP, and negative for PSCS. Genetic correlations of PM, PF, and PP with PSCS were moderate and favorable, indicating that increasing SCS decreases yield traits, as expected. Genetic correlations among yield and persistency were low to moderate and ranged from −0.09 (PSCS) to 0.18 (PF). This definition of persistency may be more useful than those used in test-day models, which are often correlated with yield. Routine genetic evaluations for persistency are feasible and may allow for improved predictions of yield traits. As calving intervals increase, persistency may have greater value.     

Short communication: Genetic association of variations in the osteopontin gene (SPP1) with lactation persistency in dairy cattle

Improving lactation persistency (LP) in dairy cattle has a beneficial effect on animal health and fertility and herd productivity. A complex trait, LP not only reflects the cow's ability to maintain milk secretion activity after the lactation peak but is also a function of the postcalving development of the mammary gland and, later on, of tissue remodeling as lactation declines. This decline is a consequence of an imbalance between cell proliferation and cell removal. In a previous study, single nucleotide polymorphisms were identified in the osteopontin (OPN) gene, SPP1. Osteopontin is a multifaceted protein that plays an important role in immune regulation and tissue remodeling. Because OPN is involved in involution, it might also have an effect on LP. The objective of the present study was to evaluate whether LP could be influenced by genetic variations in the SPP1 gene. This association with LP was analyzed in the population of 578 bulls characterized in a previous study. The population mean of estimated breeding value (EBV) for LP was 100.95 ± 5.06 units. Allele and genotype association analyses were performed by comparing the frequencies of the different genotypes and alleles with EBV for LP for the respective lactation using logistic regression. The EBV for LP at the first lactation (LP1), second lactation (LP2), and third lactation (LP3) and for overall lactation (OLP) are reported for the genotypes SPP1c.-1301G>A, SPP1c.-1251C>T, SPP1c.-430G>A, and SPP1c.*40A>C. The first single nucleotide polymorphism, SPP1c.-1301G>A, affected LP1, LP2, LP3, and OLP. Analysis of the estimated average allele substitution effects also confirmed that G is a favorable allele for LP, given the gain observed over LP1, LP2, LP3, and OLP. Differences in EBV for LP were observed between animals with different haplotypes at LP1, LP2, LP3, and OLP. Contrast analysis for OLP revealed that mean EBV is greater for block H1 (101.34 ± 0.30) than for animals that do not have H1 (98.20 ± 0.77). The gain with block H1 (GCGA) suggests the presence of the favorable allele G (first position in the block: SPP1c.-1301G). The pleiotropic roles of OPN position it at the crossroads of immune regulation, tissue remodeling, and involution. From a genetic perspective, data from the present study suggest OPN as a candidate gene associated with LP for dairy cows.     

Selection for milk production and persistency using eigenvectors of the random regression coefficient matrix

The purpose of this study was to investigate the relationships of the eigenvectors of the additive genetic random regression coefficient matrix (K) to selection responses and to determine how many eigenvectors are necessary in the breeding goal to explain the variation. The construction of various eigenvector indexes was based on the K matrix estimated from test-day records of Japanese Holstein cattle. The first (leading) eigenvector index produced constant responses for each day of lactation, indicating that the first eigenvector is responsible for scaling the lactation curve without altering its shape. Daily genetic responses to the second eigenvector index increased linearly as DIM increased. Genetic responses to the third eigenvector index were negative in mid-lactation but were positive in early and late lactation (concave curve). Genetic responses to the fourth and fifth eigenvector indexes hovered around zero across the lactation. The results suggest that both second and third eigenvectors account for the change in the shape of the lactation curve and there is little utility of the fourth and fifth eigenvectors in improving lactation milk or persistency. When the goal is to increase lactation milk yield alone, the index based on the first eigenvector produced a similar response to the index based on all 5 eigenvectors. When the goal is to improve both lactation milk yield and persistency, the index based on the first 3 eigenvectors achieved more than 99.9% of the genetic response to an index based on all 5 eigenvectors. The advantage of an eigenvector index over conventional selection based on total lactation milk yield increases with increasing economic weight assigned to persistency.     

Association between somatic cell count during the first lactation and the cumulative milk yield of cows in Irish dairy herds

Reduced potential milk yield is an important component of mastitis costs in dairy cows. The first aim of this study was to assess associations between somatic cell count (SCC) during the first lactation, and cumulative milk yield over the first lactation and subsequent lifetime of cows in Irish dairy herds. The second aim was to assess the association between SCC at 5 to 30 d in milk during parity 1 (SCC1), and SCC over the entire first lactation for cows in Irish dairy herds. The data set studied included records from 51,483 cows in 5,900 herds. Somatic cell count throughout the first lactation was summarized using the geometric mean and variance of SCC. Data were analyzed using linear models that included random effects to account for the lack of independence between observations, and herd-level variation in coefficients. Models were developed in a Bayesian framework and parameters were estimated from 10,000 Markov chain Monte Carlo simulations. The final models were a good fit to the data. A 1-unit increase in mean natural logarithm SCC over the first lactation was associated with a median decrease in first lactation and lifetime milk yield of 135 and 1,663 kg, respectively. A 1-unit increase in the variance of natural logarithm SCC over the first lactation was associated with a median decrease in lifetime milk yield of 719 kg. To demonstrate the context of lifetime milk yield results, microsimulation was used to model the trajectory of individual cows and evaluate the expected outcomes for particular changes in herd-level geometric mean SCC over the first lactation. A 75% certainty of savings of at least €199/heifer in the herd was detected if herd-level geometric mean SCC over the first lactation was reduced from ≥120,000 to ≤72,000 cells/mL. The association between SCC1 and SCC over the remainder of the first lactation was highly herd dependent, indicating that control measures for heifer mastitis should be preferentially targeted on an individual-herd basis toward either the pre- and peripartum period, or the lactating period, to optimize the lifetime milk yield of dairy cows.     

Effect of mobility score on milk yield and activity in dairy cattle

Previous studies have indicated that lame cows have reduced milk yield both before and after they are treated for lameness. One explanation for the reduction in yield before treatment is delay to treatment; that is, cows have impaired mobility for some time before they are treated. The aim of this study was to test this hypothesis by investigating temporal associations between change in milk yield and change in mobility score. Mobility score (MS, on a scale from 0 to 3), milk yield, treatments for lameness, and cow activity were recorded on 312 cows in a dairy herd in Somerset, UK, for 1 yr. The MS was scored every 2 wk and compared with daily yield and activity (mean log steps/h) averaged over the previous 16 d. Approximately 52% of MS changed within 14 d, usually by 1 unit. Overall, milk yields of cows with MS 1 were greater than those of cows with other scores. Cows with MS 2 and 3 produced 0.7 (95% confidence interval: 0.35–0.97) and 1.6 (0.98–2.23) kg less milk/d, respectively, compared with cows with MS 1. In addition, cows with MS 1 were significantly more active than cows with MS 0, 2, or 3. Cows with MS 2 and 3 were 0.02 (0.01–0.03) and 0.03 (0.01–0.05) mean log steps less active than cows with MS 1. Six to 8 wk before nonlame cows became MS 2 or 3, their daily milk yield decreased by a mean (95% CI) of 0.5 kg (0.12–0.47) and 0.9 kg (0.16–1.65) respectively. Daily yield remained lower by 0.42 kg (0.09–0.75) for 4 wk after cows with MS 2 had recovered. The activity of cows was significantly less (0.01 mean log steps) with increasing MS; the associations between activity and parity (means 0.03–0.11) and month of lactation (means 0.03–0.36) were quantitatively larger. Results from a multistate model indicated that once cows were lame they remained lame or became lame again despite treatment. In conclusion, cows’ milk production started to decline before their mobility was visibly impaired.     

Phenotypic relationships of common health disorders in dairy cows to lactation persistency estimated from daily milk weights

The objective of this study was to investigate the phenotypic relationship between common health disorders in dairy cows and lactation persistency, uncorrelated with 305-d yield. The relationships with peak yield and days in milk (DIM) at peak were also studied. Daily milk weights and treatment incidence records of 991 Holstein lactations from experimental dairy herds at Virginia Tech and Pennsylvania State University were used. Persistency was calculated as a function of daily yield deviations from standard lactation curves, developed separately for first (FL) and later lactations (LL), and deviations of DIM around reference dates: 128 for FL and 125 for LL. Days in milk at peak and peak yield were computed for each lactation by using Wood's function. The disease traits studied were mastitis (MAST) only during the first 100 d (MAST1), only after 100 DIM (MAST2), both before and after 100 DIM (MAST12), and at any stage of lactation (MAST1/2), as well as metritis, displaced abomasums, lameness, and metabolic diseases. Each disease was defined as a binary trait, distinguishing between lactations with at least one incidence (1) and lactations with no incidences (0). The relationships of diseases to persistency, DIM at peak, and peak yield were investigated separately for FL and LL for all disease traits except MAST12, which was investigated across parities. The relationships of persistency to probability of the diseases in the same lactation and in the next lactation were examined using odds ratios from a logistic regression model. Metritis and displaced abomasums in FL and LL were associated with significantly higher persistencies. Metabolic diseases and MAST1 in LL were significantly related to higher persistencies. The relationships of MAST2 in both FL and LL, and MAST12 across parities were significant but negative. Cows affected by MAST tended to have less persistent lactations. Most of the diseases had a significant impact on DIM at peak in LL. In LL, metritis, metabolic diseases, and displaced abomasums tended to significantly delay DIM at peak. Mastitis only after 100 DIM was associated with significantly earlier DIM at peak in LL. Increasing persistency was associated with low MAST2 and MAST1/2 in primiparous cows. None of the diseases studied was significantly related to persistency of the previous lactation.     

Selection on yield and fitness traits when culling holsteins during the first three lactations

Emphasis by dairy producers on various yield and fitness traits when culling cows was documented for US Holstein calvings since 1982. Least squares differences between cows retained for additional parities and those culled were estimated for milk, fat, and protein yields; somatic cell score (SCS); days open (DO); dystocia score (DS), final score (FS), and 14 type traits. Compared with cows culled during first lactation, superiority for first-parity milk yield was 569 to 1,175 kg for cows with 2 lactations, 642 to 1,283 kg for cows with ≥2 lactations, 710 to 1,350 kg for cows with 3 lactations, and 663 to 1,331 kg for cows with ≥4 lactations. Cows retained for ≥2 lactations had first-parity SCS that were 0.34 to 0.62 lower (more favorable) than those of cows culled during first lactation; first-parity SCS for cows retained for 3 or ≥4 lactations were even more favorable than those of cows with 1 or 2 lactations. The negative genetic relationship between yield and fertility contributed to increased DO as selection for higher milk yield persisted across time despite considerable preference for early conception when culling cows. In 1982, cows retained in the herd for 2, 3, and ≥4 lactations conceived earlier during first lactation (19, 17, and 23 fewer DO, respectively) than those culled during first lactation; those differences had increased to 34, 41, and 52 fewer DO by 2000. Although DS has a negative relationship with survival, first-parity DS were only slightly lower (by 0.10 to 0.14) for survivors than for cows culled during first lactation. Cows retained for ≥2 lactations had greater first-parity FS by 1.4 to 1.9 points than those culled during first lactation. On a standardized basis, the most intense selection during first lactation was for milk and protein yields with less for fat (74 to 86% of that for milk), DO (18 to 74%), FS (22 to 38%), SCS (19 to 37%), and DS (7 to 15%). Producers continued to emphasize the same traits when culling during second and third lactations. Trait priority by producers during culling could aid in setting trait emphasis when selecting bulls for progeny test and could also be useful in developing software for index-based culling guides.     

Deconstructing milk yield and composition during lactation using biologically based lactation models

A recently developed biological model of lactation described changes in daily milk yield throughout lactation as the result of 3 processes, secretory cell differentiation, cell death, and secretion rate per cell. This paper extends the model to describe the production of milk components (fat, protein, lactose, and water) throughout lactation by replacing milk secretion rate of the original model with the secretion rates of the four components. The milk component model approach was used to examine the relationship between milk yield and the major determinants of its production, using the secretion of milk components throughout lactation. Newly derived models were tested on 461 lactations from a single Holstein herd and used to estimate variability of secretion rates throughout lactation. Because the pattern of cell numbers throughout lactation is not precisely known, an alternative pattern of cell numbers was modeled and the concomitant change in secretion rates outlined. Fat secretion rate was the most variable, as measured by its weekly coefficient of variation throughout lactation. Secretion rates of lactose and water were nearly constant throughout lactation and highly correlated (0.94). Fat and protein secretion rates also were well correlated (0.53). The known biochemistry of milk component production related well to the secretion rate observations derived from the model. Lactose secretion rate and numbers of active secretory cells primarily determined daily milk yield.