Updating test-day milk yield factors for use in genetic evaluations and dairy production systems: a comprehensive review.

Various methods have been proposed to estimate daily yield from partial yields, primarily to deal with unequal milking intervals. This paper offers an exhaustive review of daily milk yields, the foundation of lactation records. Seminal advancements in the late 20th century concentrated on two main adjustment metrics: additive additive correction factors (ACF) and multiplicative correction factors (MCF). An ACF model provides additive adjustments to two times AM or PM milk yield, which then becomes the estimated daily yields, whereas an MCF is a ratio of daily yield to the yield from a single milking. Recent studies highlight the potential of alternative approaches, such as exponential regression and other nonlinear models. Biologically, milk secretion rates are not linear throughout the entire milking interval, influenced by the internal mammary gland pressure. Consequently, nonlinear models are appealing for estimating daily milk yields as well. MCFs and ACFs are typically determined for discrete milking interval classes. Nonetheless, large discrete intervals can introduce systematic biases. A universal solution for deriving continuous correction factors has been proposed, ensuring reduced bias and enhanced daily milk yield estimation accuracy. When leveraging test-day milk yields for genetic evaluations in dairy cattle, two predominant statistical models are employed: lactation and test-day yield models. A lactation model capitalizes on the high heritability of total lactation yields, aligning closely with dairy producers' needs because the total amount of milk production in a lactation directly determines farm revenue. However, a lactation yield model without harnessing all test-day records may ignore vital data about the shapes of lactation curves needed for informed breeding decisions. In contrast, a test-day model emphasizes individual test-day data, accommodating various intervals and recording plans and allowing the estimation of environmental effects on specific test days. In the United States, the patenting of test-day models in 1993 used to restrict the use of test-day models to regional and unofficial evaluations by the patent holders. Estimated test-day milk yields have been used as if they were accurate depictions of actual milk yields, neglecting possible estimation errors. Its potential consequences on subsequent genetic evaluations have not been sufficiently addressed. Moving forward, there are still numerous questions and challenges in this domain.

Does modeling causal relationships improve the accuracy of predicting lactation milk yields?

This study compared 3 correlational (best prediction, linear regression, and feed-forward neural networks) and 2 causal models (recursive structural equation model and recurrent neural networks) for estimating lactation milk yields. The correlational models assumed associations between test-day milk yields (health conditions), while the casual models postulated unidirectional recursive effects between these test-day variables. Wood lactation curves were used to simulate the data and served as a benchmark model. Individual Wood lactation curves provided an excellent parametric interpretation of lactation dynamics, with their prediction accuracies depending on the coverage of the lactation curve dynamics. Best prediction outperformed other models in the absence of mastitis but was suboptimal when mastitis was present and unaccounted for. Recurrent neural networks yielded the highest accuracy when mastitis was present. Although causal models facilitated the inference about the causality underlying lactation, precisely capturing the causal relationships was challenging because the underlying biology was complex. Misspecification of recursive effects in the recursive structural equation model resulted in a loss of accuracy. Hence, modeling causal relationships does not necessarily guarantee improved accuracies. In practice, a parsimonious model is preferred, balancing model complexity and accuracy. In addition to the choice of statistical models, the proper accounting for factors and covariates affecting milk yields is equally crucial.

Estimating test-day milk yields by modeling proportional daily yields: Going beyond linearity.

In the United States, lactation milk yields are not measured directly but are calculated from the test-day milk yields. Still, test-day milk yields are estimated from partial yields obtained from single milkings. Various methods have been proposed to estimate test-day milk yields, primarily to deal with unequal milking intervals dating back to the 1970s and 1980s. The Wiggans model is a de facto method for estimating test-day milk yields in the United States, which was initially proposed for cows milked 3 times daily, assuming a linear relationship between a proportional test-day milk yield and milking interval. However, the linearity assumption did not hold precisely in Holstein cows milked twice daily because of prolonged and uneven milking intervals. The present study reviewed and evaluated the nonlinear models that extended the Wiggans model for estimating daily or test-day milk yields. These nonlinear models, except step functions, demonstrated smaller errors and greater accuracies for estimated test-day milk yields compared with the conventional methods. The nonlinear models offered additional benefits. For example, the locally weighted regression model (e.g., locally estimated scatterplot smoothing) could utilize data information in scalable neighborhoods and weigh observations according to their distance in milking interval time. General additive models provide a flexible, unified framework to model nonlinear predictor variables additively. Another drawback of the conventional methods is a loss of accuracy caused by discretizing milking interval time into large bins while deriving multiplicative correction factors for estimating test-day milk yields. To overcome this problem, we proposed a general approach that allows milk yield correction factors to be derived for every possible milking interval time, resulting in more accurately estimated test-day milk yields. This approach can be applied to any model, including nonparametric models.

Daily milk yield correction factors: What are they?

Cows are typically milked 2 or more times on a test-day, but not all these milkings are sampled and weighed. The initial approach estimated a test-day yield with doubled morning (AM) or evening (PM) yield in the AM-PM milking plans, assuming equal AM and PM milking intervals. However, AM and PM milking intervals can vary, and milk secretion rates may be different between day and night. Statistical methods have been proposed to estimate daily yields in dairy cows, focusing on various yield correction factors in 2 broad categories: additive correction factors (ACF) and multiplicative correction factors (MCF). The ACF are evaluated by the average differences between AM and PM milk yield for various milking interval classes, coupled with other categorical variables. We show that an ACF model is equivalent to a regression model of daily yield on categorical regressor variables, and a continuous variable for AM or PM yield with a fixed regression coefficient of 2.0. Similarly, a linear regression model can be implemented as an ACF model with the regression coefficient for AM or PM yield estimated from the data. The linear regression models improved the accuracy of the estimates compared with the ACF models. The MCF are ratios of daily yield to yield from single milkings, but their statistical interpretations vary. Overall, MCF were more accurate for estimating daily milk yield than ACF. The MCF have biological and statistical challenges. Systematic biases occurred when ACF or MCF were computed on discretized milking interval classes, leading to accuracy loss. An exponential regression model was proposed as an alternative model for estimating daily milk yields, which improved the accuracy. Characterization of ACF and MCF showed how they improved the accuracy compared with doubling AM or PM yield as the daily milk yield. All the methods performed similarly with equal AM and PM milkings. The methods were explicitly described to estimate daily milk yield in AM and PM milking plans. Still, the principles generally apply to cows milked more than 2 times a day and apply similarly to the estimation of daily fat and protein yields with some necessary modifications.

Genomic selection in United States dairy cattle.

The genomic selection program for dairy cattle in the United States has doubled the rate of genetic gain. Since 2010, the average annual increase in net merit has been $85 compared to $40 during the previous 5 years. The number of genotypes has been rapidly increasing both domestically and internationally and reached over 6.5 million in 2022 with 1,134,593 submitted in 2021. Evaluations are calculated for over 50 traits. Feed efficiency (residual feed intake), heifer and cow livability, age at first calving, six health traits, and gestation length have been added in recent years to represent the economic value of selection candidates more accurately; work is underway to develop evaluations for hoof health. Evaluations of animals with newly submitted genotypes are calculated weekly. In April 2019, evaluations were extended to crossbreds; to support that effort, evaluations are initially calculated on an all-breed base and then blended by an estimated breed composition. For animals that are less than 90% of one breed, the evaluation is calculated by weighting contributions of each of the five major dairy breeds evaluated (Ayrshire, Brown Swiss, Guernsey, Holstein, and Jersey) by the breed proportion. Nearly 200,000 animals received blended evaluations in July 2022. Pedigree is augmented by using haplotype matching to discover maternal grandsires and great-grandsires. Haplotype analysis is also used to discover undesirable recessive conditions. In many cases, the causative variant has been identified, and results from a gene test or inclusion on a genotyping chip improves the accuracy of those determinations for the current 27 conditions reported. Recently discovered recessive conditions include neuropathy with splayed forelimbs in Jerseys, early embryonic death in Holsteins, and curly calves in Ayrshires. Techniques have been developed to support rapid searches for parent-progeny relationships and identical genotypes among all likely genotypes, which substantially reduces processing time. Work continues on using sequence data to discover additional informative single nucleotide polymorphisms and to incorporate those previously discovered. Adoption of genotyping by sequencing is expected to improve flexibility of marker selection. The success of the Council on Dairy Cattle Breeding in conducting the genetic evaluation program is the result of close cooperation with industry and research groups, including the United States Department of Agriculture, breed associations, genotyping laboratories, and artificial-insemination organizations.

Statistical Methods Revisited for Estimating Daily Milk Yields: How Well do They Work?

Cost-effective milking plans have been adapted to supplement the standard supervised twice-daily monthly testing scheme since the 1960s. Various methods have been proposed to estimate daily milk yields (DMY), focusing on yield correction factors. The present study evaluated the performance of existing statistical methods, including a recently proposed exponential regression model, for estimating DMY using 10-fold cross-validation in Holstein and Jersey cows. The initial approach doubled the morning (AM) or evening (PM) yield as estimated DMY in AM-PM plans, assuming equal 12-h AM and PM milking intervals. However, in reality, AM milking intervals tended to be longer than PM milking intervals. Additive correction factors (ACF) provided additive adjustments beyond twice AM or PM yields. Hence, an ACF model equivalently assumed a fixed regression coefficient or a multiplier of "2.0" for AM or PM yields. Similarly, a linear regression model was viewed as an ACF model, yet it estimated the regression coefficient for a single milk yield from the data. Multiplicative correction factors (MCF) represented daily to partial milk yield ratios. Hence, multiplying a yield from single milking by an appropriate MCF gave a DMY estimate. The exponential regression model was analogous to an exponential growth function with the yield from single milking as the initial state and the rate of change tuned by a linear function of milking interval. In the present study, all the methods had high precision in the estimates, but they differed considerably in biases. Overall, the MCF and linear regression models had smaller squared biases and greater accuracies for estimating DMY than the ACF models. The exponential regression model had the greatest accuracies and smallest squared biases. Model parameters were compared. Discretized milking interval categories led to a loss of accuracy of the estimates. Characterization of ACF and MCF revealed their similarities and dissimilarities and biases aroused by unequal milking intervals. The present study focused on estimating DMY in AM-PM milking plans. Yet, the methods and relevant principles are generally applicable to cows milked more than two times a day.

Copy number variation analysis reveals variants associated with milk production traits in dairy goats.

Copy number variation (CNV) is a major type of genomic structural variation. We investigated their impacts on goat dairy traits using the CaprineSNP50 array. From 120 samples of five dairy goat breeds, we totally identified 42 CNVs ranging from 56,044 bp to 4,337,625 bp. We found significant associations between two CNVs (CNV5 and CNV25) and two milk production traits (mean of milk fat yield and mean of milk protein yield) after false discovery rate (FDR) correction (P < 0.05). CNV5 overlaps the ADAMTS20 gene, which is involved in the differentiation of mammary cell and plays a crucial role in lactogenic activity of bovine mammary epithelial cells. CNV25 overlaps with PAPPA2, which has been found to be associated with bovine reproduction and milk production traits. Our results revealed that CNVs overlapped with ADAMTS20 and PAPPA2 could be involved in goat dairy traits and function as candidate markers for further genetic selection.

Additional considerations to the use of single-step genomic predictions in a dominance setting.

Recent publications indicate that single-step models are suitable to estimate breeding values, dominance deviations and total genetic values with acceptable quality. Additive single-step methods implicitly extend known number of allele information from genotyped to non-genotyped animals. This theory is well derived in an additive setting. It was recently shown, at least empirically, that this basic strategy can be extended to dominance with reasonable prediction quality. Our study addressed two additional issues. It illustrated the theoretical basis for extension and validated genomic predictions to dominance based on single-step genomic best linear unbiased prediction theory. This development was then extended to include inbreeding into dominance relationships, which is a currently not yet solved issue. Different parametrizations of dominance relationship matrices were proposed. Five dominance single-step inverse matrices were tested and described as C1 , C2 , C3 , C4 and C5 . Genotypes were simulated for a real pig population (n = 11,943 animals). In order to avoid any confounding issues with additive effects, pseudo-records including only dominance deviations and residuals were simulated. SNP effects of heterozygous genotypes were summed up to generate true dominance deviations. We added random noise to those values and used them as phenotypes. Accuracy was defined as correlation between true and predicted dominance deviations. We conducted five replicates and estimated accuracies in three sets: between all (S1 ), non-genotyped (S2 ) and inbred non-genotyped (S3 ) animals. Potential bias was assessed by regressing true dominance deviations on predicted values. Matrices accounting for inbreeding (C3 , C4 and C5 ) best fit. Accuracies were on average 0.77, 0.40 and 0.46 in S1 , S2 and S3 , respectively. In addition, C3 , C4 and C5 scenarios have shown better accuracies than C1 and C2 , and dominance deviations were less biased. Better matrix compatibility (accuracy and bias) was observed by re-scaling diagonal elements to 1 minus the inbreeding coefficient (C5 ).

Genome changes due to artificial selection in U.S. Holstein cattle.

BACKGROUND: The availability of a unique unselected Holstein line since 1964 provided a direct comparison between selected and unselected Holstein genomes whereas large Holstein samples provided unprecedented statistical power for identifying high-confidence SNP effects. Utilizing these unique resources, we aimed to identify genome changes affected by selection since 1964. RESULTS: Direct comparison of genome-wide SNP markers between a Holstein line unselected since 1964 and contemporary Holsteins showed that the 40 years of artificial selection since 1964 resulted in genome landscape changes. Among the regions affected by selection, the regions containing 198 genes with fertility functions had a larger negative correlation than that of all SNPs between the SNP effects on milk yield and daughter pregnancy rate. These results supported the hypothesis that hitchhiking of genetic selection for milk production by negative effects of fertility genes contributed to the unintended declines in fertility since 1964. The genome regions subjected to selection also contained 67 immunity genes, the bovine MHC region of Chr23 with significantly decreased heterozygosity in contemporary Holsteins, and large gene clusters including T-cell receptor and immunoglobulin genes. CONCLUSIONS: This study for the first time provided direct evidence that genetic selection for milk production affected fertility and immunity genes and that the hitchhiking of genetic selection for milk production by negative fertility effects contributed to the fertility declines since 1964, and identified a large number of candidate fertility and immunity genes affected by selection. The results provided novel understanding about genome changes due to artificial selection and their impact on fertility and immunity genes and could facilitate developing genetic methods to reverse the declines in fertility and immunity in Holstein cattle.

Genome-wide copy number variant analysis reveals variants associated with 10 diverse production traits in Holstein cattle.

BACKGROUND: Copy number variation (CNV) is an important type of genetic variation contributing to phenotypic differences among mammals and may serve as an alternative molecular marker to single nucleotide polymorphism (SNP) for genome-wide association study (GWAS). Recently, GWAS analysis using CNV has been applied in livestock, although few studies have focused on Holstein cattle. RESULTS: We describe 191 CNV detected using intensity data from over 700,000 SNP genotypes generated with the BovineHD Genotyping BeadChip (Illumina, San Diego, CA) in 528 Holstein cows. The CNV were used for GWAS analysis of 10 important production traits of 473 cattle related to feed intake, milk quality, and female fertility, as well as 2 composite traits of net merit and productive life. In total, we detected 57 CNV associated (P < 0.05 after false discovery rate correction) with at least one of the 10 phenotypes. Focusing on feed efficiency and intake-related phenotypes of residual feed intake and dry matter intake, we detected a single CNV associated with both traits which overlaps a predicted olfactory receptor gene OR2A2 (LOC787786). Additionally, 2 CNV within the RXFP4 (relaxin/insulin like family peptide receptor 4) and 2 additional olfactory receptor gene regions, respectively, were associated with residual feed intake. The RXFP4 gene encodes a receptor for an orexigenic peptide, insulin-like peptide 5 produced by intestinal L cells, which is expressed by enteric neurons. Olfactory receptors are critical for transmitting the effects of odorants, contributing to the sense of smell, and have been implicated in participating in appetite regulation. CONCLUSIONS: Our results identify CNV for genomic evaluation in Holstein cattle, and provide candidate genes, such as RXFP4, contributing to variation in feed efficiency and feed intake-related traits. These results indicate potential novel targets for manipulating feed intake-related traits of livestock.

Genomic Selection in Dairy Cattle: The USDA Experience.

Genomic selection has revolutionized dairy cattle breeding. Since 2000, assays have been developed to genotype large numbers of single-nucleotide polymorphisms (SNPs) at relatively low cost. The first commercial SNP genotyping chip was released with a set of 54,001 SNPs in December 2007. Over 15,000 genotypes were used to determine which SNPs should be used in genomic evaluation of US dairy cattle. Official USDA genomic evaluations were first released in January 2009 for Holsteins and Jerseys, in August 2009 for Brown Swiss, in April 2013 for Ayrshires, and in April 2016 for Guernseys. Producers have accepted genomic evaluations as accurate indications of a bull's eventual daughter-based evaluation. The integration of DNA marker technology and genomics into the traditional evaluation system has doubled the rate of genetic progress for traits of economic importance, decreased generation interval, increased selection accuracy, reduced previous costs of progeny testing, and allowed identification of recessive lethals.

Optimal Design of Low-Density SNP Arrays for Genomic Prediction: Algorithm and Applications.

Low-density (LD) single nucleotide polymorphism (SNP) arrays provide a cost-effective solution for genomic prediction and selection, but algorithms and computational tools are needed for the optimal design of LD SNP chips. A multiple-objective, local optimization (MOLO) algorithm was developed for design of optimal LD SNP chips that can be imputed accurately to medium-density (MD) or high-density (HD) SNP genotypes for genomic prediction. The objective function facilitates maximization of non-gap map length and system information for the SNP chip, and the latter is computed either as locus-averaged (LASE) or haplotype-averaged Shannon entropy (HASE) and adjusted for uniformity of the SNP distribution. HASE performed better than LASE with ≤1,000 SNPs, but required considerably more computing time. Nevertheless, the differences diminished when >5,000 SNPs were selected. Optimization was accomplished conditionally on the presence of SNPs that were obligated to each chromosome. The frame location of SNPs on a chip can be either uniform (evenly spaced) or non-uniform. For the latter design, a tunable empirical Beta distribution was used to guide location distribution of frame SNPs such that both ends of each chromosome were enriched with SNPs. The SNP distribution on each chromosome was finalized through the objective function that was locally and empirically maximized. This MOLO algorithm was capable of selecting a set of approximately evenly-spaced and highly-informative SNPs, which in turn led to increased imputation accuracy compared with selection solely of evenly-spaced SNPs. Imputation accuracy increased with LD chip size, and imputation error rate was extremely low for chips with ≥3,000 SNPs. Assuming that genotyping or imputation error occurs at random, imputation error rate can be viewed as the upper limit for genomic prediction error. Our results show that about 25% of imputation error rate was propagated to genomic prediction in an Angus population. The utility of this MOLO algorithm was also demonstrated in a real application, in which a 6K SNP panel was optimized conditional on 5,260 obligatory SNP selected based on SNP-trait association in U.S. Holstein animals. With this MOLO algorithm, both imputation error rate and genomic prediction error rate were minimal.

Changes in genetic selection differentials and generation intervals in US Holstein dairy cattle as a result of genomic selection.

Seven years after the introduction of genomic selection in the United States, it is now possible to evaluate the impact of this technology on the population. Selection differential(s) (SD) and generation interval(s) (GI) were characterized in a four-path selection model that included sire(s) of bulls (SB), sire(s) of cows (SC), dam(s) of bulls (DB), and dam(s) of cows (DC). Changes in SD over time were estimated for milk, fat, and protein yield; somatic cell score (SCS); productive life (PL); and daughter pregnancy rate (DPR) for the Holstein breed. In the period following implementation of genomic selection, dramatic reductions were seen in GI, especially the SB and SC paths. The SB GI reduced from ∼7 y to less than 2.5 y, and the DB GI fell from about 4 y to nearly 2.5 y. SD were relatively stable for yield traits, although modest gains were noted in recent years. The most dramatic response to genomic selection was observed for the lowly heritable traits DPR, PL, and SCS. Genetic trends changed from close to zero to large and favorable, resulting in rapid genetic improvement in fertility, lifespan, and health in a breed where these traits eroded over time. These results clearly demonstrate the positive impact of genomic selection in US dairy cattle, even though this technology has only been in use for a short time. Based on the four-path selection model, rates of genetic gain per year increased from ∼50-100% for yield traits and from threefold to fourfold for lowly heritable traits.

Cattle Sex-Specific Recombination and Genetic Control from a Large Pedigree Analysis.

Meiotic recombination is an essential biological process that generates genetic diversity and ensures proper segregation of chromosomes during meiosis. From a large USDA dairy cattle pedigree with over half a million genotyped animals, we extracted 186,927 three-generation families, identified over 8.5 million maternal and paternal recombination events, and constructed sex-specific recombination maps for 59,309 autosomal SNPs. The recombination map spans for 25.5 Morgans in males and 23.2 Morgans in females, for a total studied region of 2,516 Mb (986 kb/cM in males and 1,085 kb/cM in females). The male map is 10% longer than the female map and the sex difference is most pronounced in the subtelomeric regions. We identified 1,792 male and 1,885 female putative recombination hotspots, with 720 hotspots shared between sexes. These hotspots encompass 3% of the genome but account for 25% of the genome-wide recombination events in both sexes. During the past forty years, males showed a decreasing trend in recombination rate that coincided with the artificial selection for milk production. Sex-specific GWAS analyses identified PRDM9 and CPLX1 to have significant effects on genome-wide recombination rate in both sexes. Two novel loci, NEK9 and REC114, were associated with recombination rate in both sexes, whereas three loci, MSH4, SMC3 and CEP55, affected recombination rate in females only. Among the multiple PRDM9 paralogues on the bovine genome, our GWAS of recombination hotspot usage together with linkage analysis identified the PRDM9 paralogue on chromosome 1 to be associated in the U.S. Holstein data. Given the largest sample size ever reported for such studies, our results reveal new insights into the understanding of cattle and mammalian recombination.

Effect of artificial selection on runs of homozygosity in u.s. Holstein cattle.

The intensive selection programs for milk made possible by mass artificial insemination increased the similarity among the genomes of North American (NA) Holsteins tremendously since the 1960s. This migration of elite alleles has caused certain regions of the genome to have runs of homozygosity (ROH) occasionally spanning millions of continuous base pairs at a specific locus. In this study, genome signatures of artificial selection in NA Holsteins born between 1953 and 2008 were identified by comparing changes in ROH between three distinct groups under different selective pressure for milk production. The ROH regions were also used to estimate the inbreeding coefficients. The comparisons of genomic autozygosity between groups selected or unselected since 1964 for milk production revealed significant differences with respect to overall ROH frequency and distribution. These results indicate selection has increased overall autozygosity across the genome, whereas the autozygosity in an unselected line has not changed significantly across most of the chromosomes. In addition, ROH distribution was more variable across the genomes of selected animals in comparison to a more even ROH distribution for unselected animals. Further analysis of genome-wide autozygosity changes and the association between traits and haplotypes identified more than 40 genomic regions under selection on several chromosomes (Chr) including Chr 2, 7, 16 and 20. Many of these selection signatures corresponded to quantitative trait loci for milk, fat, and protein yield previously found in contemporary Holsteins.

Imputation of microsatellite alleles from dense SNP genotypes for parentage verification across multiple Bos taurus and Bos indicus breeds.

To assist cattle producers transition from microsatellite (MS) to single nucleotide polymorphism (SNP) genotyping for parental verification we previously devised an effective and inexpensive method to impute MS alleles from SNP haplotypes. While the reported method was verified with only a limited data set (N = 479) from Brown Swiss, Guernsey, Holstein, and Jersey cattle, some of the MS-SNP haplotype associations were concordant across these phylogenetically diverse breeds. This implied that some haplotypes predate modern breed formation and remain in strong linkage disequilibrium. To expand the utility of MS allele imputation across breeds, MS and SNP data from more than 8000 animals representing 39 breeds (Bos taurus and B. indicus) were used to predict 9410 SNP haplotypes, incorporating an average of 73 SNPs per haplotype, for which alleles from 12 MS markers could be accurately be imputed. Approximately 25% of the MS-SNP haplotypes were present in multiple breeds (N = 2 to 36 breeds). These shared haplotypes allowed for MS imputation in breeds that were not represented in the reference population with only a small increase in Mendelian inheritance inconsistancies. Our reported reference haplotypes can be used for any cattle breed and the reported methods can be applied to any species to aid the transition from MS to SNP genetic markers. While ~91% of the animals with imputed alleles for 12 MS markers had ≤1 Mendelian inheritance conflicts with their parents' reported MS genotypes, this figure was 96% for our reference animals, indicating potential errors in the reported MS genotypes. The workflow we suggest autocorrects for genotyping errors and rare haplotypes, by MS genotyping animals whose imputed MS alleles fail parentage verification, and then incorporating those animals into the reference dataset.

Effect of sample stratification on dairy GWAS results.

BACKGROUND: Artificial insemination and genetic selection are major factors contributing to population stratification in dairy cattle. In this study, we analyzed the effect of sample stratification and the effect of stratification correction on results of a dairy genome-wide association study (GWAS). Three methods for stratification correction were used: the efficient mixed-model association expedited (EMMAX) method accounting for correlation among all individuals, a generalized least squares (GLS) method based on half-sib intraclass correlation, and a principal component analysis (PCA) approach. RESULTS: Historical pedigree data revealed that the 1,654 contemporary cows in the GWAS were all related when traced through approximately 10-15 generations of ancestors. Genome and phenotype stratifications had a striking overlap with the half-sib structure. A large elite half-sib family of cows contributed to the detection of favorable alleles that had low frequencies in the general population and high frequencies in the elite cows and contributed to the detection of X chromosome effects. All three methods for stratification correction reduced the number of significant effects. EMMAX method had the most severe reduction in the number of significant effects, and the PCA method using 20 principal components and GLS had similar significance levels. Removal of the elite cows from the analysis without using stratification correction removed many effects that were also removed by the three methods for stratification correction, indicating that stratification correction could have removed some true effects due to the elite cows. SNP effects with good consensus between different methods and effect size distributions from USDA's Holstein genomic evaluation included the DGAT1-NIBP region of BTA14 for production traits, a SNP 45kb upstream from PIGY on BTA6 and two SNPs in NIBP on BTA14 for protein percentage. However, most of these consensus effects had similar frequencies in the elite and average cows. CONCLUSIONS: Genetic selection and extensive use of artificial insemination contributed to overlapped genome, pedigree and phenotype stratifications. The presence of an elite cluster of cows was related to the detection of rare favorable alleles that had high frequencies in the elite cluster and low frequencies in the remaining cows. Methods for stratification correction could have removed some true effects associated with genetic selection.

Fine mapping of copy number variations on two cattle genome assemblies using high density SNP array.

BACKGROUND: Btau_4.0 and UMD3.1 are two distinct cattle reference genome assemblies. In our previous study using the low density BovineSNP50 array, we reported a copy number variation (CNV) analysis on Btau_4.0 with 521 animals of 21 cattle breeds, yielding 682 CNV regions with a total length of 139.8 megabases. RESULTS: In this study using the high density BovineHD SNP array, we performed high resolution CNV analyses on both Btau_4.0 and UMD3.1 with 674 animals of 27 cattle breeds. We first compared CNV results derived from these two different SNP array platforms on Btau_4.0. With two thirds of the animals shared between studies, on Btau_4.0 we identified 3,346 candidate CNV regions representing 142.7 megabases (~4.70%) of the genome. With a similar total length but 5 times more event counts, the average CNVR length of current Btau_4.0 dataset is significantly shorter than the previous one (42.7 kb vs. 205 kb). Although subsets of these two results overlapped, 64% (91.6 megabases) of current dataset was not present in the previous study. We also performed similar analyses on UMD3.1 using these BovineHD SNP array results. Approximately 50% more and 20% longer CNVs were called on UMD3.1 as compared to those on Btau_4.0. However, a comparable result of CNVRs (3,438 regions with a total length 146.9 megabases) was obtained. We suspect that these results are due to the UMD3.1 assembly's efforts of placing unplaced contigs and removing unmerged alleles. Selected CNVs were further experimentally validated, achieving a 73% PCR validation rate, which is considerably higher than the previous validation rate. About 20-45% of CNV regions overlapped with cattle RefSeq genes and Ensembl genes. Panther and IPA analyses indicated that these genes provide a wide spectrum of biological processes involving immune system, lipid metabolism, cell, organism and system development. CONCLUSION: We present a comprehensive result of cattle CNVs at a higher resolution and sensitivity. We identified over 3,000 candidate CNV regions on both Btau_4.0 and UMD3.1, further compared current datasets with previous results, and examined the impacts of genome assemblies on CNV calling.

Design of a bovine low-density SNP array optimized for imputation.

The Illumina BovineLD BeadChip was designed to support imputation to higher density genotypes in dairy and beef breeds by including single-nucleotide polymorphisms (SNPs) that had a high minor allele frequency as well as uniform spacing across the genome except at the ends of the chromosome where densities were increased. The chip also includes SNPs on the Y chromosome and mitochondrial DNA loci that are useful for determining subspecies classification and certain paternal and maternal breed lineages. The total number of SNPs was 6,909. Accuracy of imputation to Illumina BovineSNP50 genotypes using the BovineLD chip was over 97% for most dairy and beef populations. The BovineLD imputations were about 3 percentage points more accurate than those from the Illumina GoldenGate Bovine3K BeadChip across multiple populations. The improvement was greatest when neither parent was genotyped. The minor allele frequencies were similar across taurine beef and dairy breeds as was the proportion of SNPs that were polymorphic. The new BovineLD chip should facilitate low-cost genomic selection in taurine beef and dairy cattle.

Genome-wide association analysis of thirty one production, health, reproduction and body conformation traits in contemporary U.S. Holstein cows.

BACKGROUND: Genome-wide association analysis is a powerful tool for annotating phenotypic effects on the genome and knowledge of genes and chromosomal regions associated with dairy phenotypes is useful for genome and gene-based selection. Here, we report results of a genome-wide analysis of predicted transmitting ability (PTA) of 31 production, health, reproduction and body conformation traits in contemporary Holstein cows. RESULTS: Genome-wide association analysis identified a number of candidate genes and chromosome regions associated with 31 dairy traits in contemporary U.S. Holstein cows. Highly significant genes and chromosome regions include: BTA13's GNAS region for milk, fat and protein yields; BTA7's INSR region and BTAX's LOC520057 and GRIA3 for daughter pregnancy rate, somatic cell score and productive life; BTA2's LRP1B for somatic cell score; BTA14's DGAT1-NIBP region for fat percentage; BTA1's FKBP2 for protein yields and percentage, BTA26's MGMT and BTA6's PDGFRA for protein percentage; BTA18's 53.9-58.7 Mb region for service-sire and daughter calving ease and service-sire stillbirth; BTA18's PGLYRP1-IGFL1 region for a large number of traits; BTA18's LOC787057 for service-sire stillbirth and daughter calving ease; BTA15's CD82, BTA23's DST and the MOCS1-LRFN2 region for daughter stillbirth; and BTAX's LOC520057 and GRIA3 for daughter pregnancy rate. For body conformation traits, BTA11, BTAX, BTA10, BTA5, and BTA26 had the largest concentrations of SNP effects, and PHKA2 of BTAX and REN of BTA16 had the most significant effects for body size traits. For body shape traits, BTAX, BTA19 and BTA3 were most significant. Udder traits were affected by BTA16, BTA22, BTAX, BTA2, BTA10, BTA11, BTA20, BTA22 and BTA25, teat traits were affected by BTA6, BTA7, BTA9, BTA16, BTA11, BTA26 and BTA17, and feet/legs traits were affected by BTA11, BTA13, BTA18, BTA20, and BTA26. CONCLUSIONS: Genome-wide association analysis identified a number of genes and chromosome regions associated with 31 production, health, reproduction and body conformation traits in contemporary Holstein cows. The results provide useful information for annotating phenotypic effects on the dairy genome and for building consensus of dairy QTL effects.