Multiple quantitative trait loci mapping with cofactors and application of alternative variants of the false discovery rate in an enlarged granddaughter design

Genome-wide association study for milk yield, frame, and udder conformation traits of Gir dairy cattle

Genome-wide association studies (GWAS) are employed to identify genomic regions and candidate genes associated with several traits. The aim of this study was to perform a GWAS to identify causative variants and genes associated with milk yield, frame, and udder conformation traits in Gir dairy cattle. Body conformation traits were classified as "frame," and "udder" traits for this study. After genotyping imputation and quality control 42,105 polymorphisms were available for analyses and 24,489 cows with pedigree information had phenotypes. First, P-value was calculated based on the variance of the prediction error of the SNP-effects on the first iteration. After that, 2 more iterations were performed to carry out the weighted single-step genome-wide association methodology, performed using genomic moving windows defined based on linkage disequilibrium. The significant SNPs and top 10 windows explaining the highest percentage of additive genetic variance were selected and used for QTL and gene annotation. The variants identified in our work overlapped with QTLs from the animal QTL database on chromosomes 1 to 23, except for chromosome 4. The Gir breed is less studied than the Holstein breed and as such the animal QTL database is biased to Holstein results. Hence it is noteworthy that our GWAS had similarities with previously described QTLs. These previously known QTLs were related to milk yield, body height, rump angle, udder width, and udder depth. In total, 5 genes were annotated. Of these genes, FAM13A and CMSS1 had been previously related to bone and carcass weight in cattle.

Single-step genome-wide association analyses for milk urea concentration in Walloon Holstein cows

The aims of this study were to conduct a single-step genome-wide association to identify genomic regions associated with milk urea (MU) and to estimate genetic correlations between MU and milk yield (MY), milk composition (calcium content [CC], fat percentage [FP], protein percentage [PP], and casein percentage [CNP]), and cheese-making properties (CMP; coagulation time [CT], curd firmness after 30 min from rennet addition [a30], and titratable acidity [TA]). The used data have been collected from 2015 to 2020 on 78,073 first-parity (485,218 test-day records) and 48,766 second-parity (284,942 test-day records) Holstein cows distributed in 671 herds in the Walloon Region of Belgium. Data of 565,533 SNP located on 29 BTA of 6,617 animals (1,712 males) were used. Random regression test-day models were used to estimate genetic parameters through the Bayesian Gibbs sampling method. The SNP solutions were estimated using a single-step genomic BLUP approach. The proportion of the total additive genetic variance explained by windows of 50 consecutive SNPs (with an average size of ∼216 kb) was calculated, and the top-3 genomic regions explaining the largest rate of the genetic variance were considered promising regions and used to identify potential candidate genes. Mean (SD) MU was 25.38 (8.02) mg/dL and 25.03 (8.06) mg/dL in the first and second lactation, respectively. Mean heritability estimates for daily MU were 0.21 and 0.23 for the first and second lactation, respectively. The genetic correlations estimated between MU and MY, milk composition, and CMP were quite low (ranged from -0.10 [CC] to 0.10 [TA] and -0.05 [CT] to 0.13 [TA] for the first and second lactations, respectively). The top-3 regions associated with MU were located from 80.61 to 80.74 Mb on BTA6, 103.26 to 103.41 Mb on BTA11, and 1.59 to 2.15 Mb on BTA14. Genes including KCNT1, MROH1, SHARPIN, TSSK5, CPSF1, HSF1, TONSL, DGAT1, SCX, and MAF1 were identified as positional candidate genes for MU. The findings of this study can be used for a better understanding of the genomic architecture underlying MU in Holstein cattle.

Single-step genome-wide association analyses for selected infrared-predicted cheese-making traits in Walloon Holstein cows

This study aimed to perform genome-wide association study to identify genomic regions associated with milk production and cheese-making properties (CMP) in Walloon Holstein cows. The studied traits were milk yield, fat percentage, protein percentage, casein percentage (CNP), calcium content, somatic cell score (SCS), coagulation time, curd firmness after 30 min from rennet addition, and titratable acidity. The used data have been collected from 2014 to 2020 on 78,073 first-parity (485,218 test-day records), 48,766 second-parity (284,942 test-day records), and 21,948 third-parity (105,112 test-day records) Holstein cows distributed in 671 herds in the Walloon Region of Belgium. Data of 565,533 single nucleotide polymorphisms (SNP), located on 29 Bos taurus autosomes (BTA) of 6,617 animals (1,712 males), were used. Random regression test-day models were used to estimate genetic parameters through the Bayesian Gibbs sampling method. The SNP solutions were estimated using a single-step genomic BLUP approach. The proportion of the total additive genetic variance explained by windows of 50 consecutive SNPs (with an average size of ∼216 KB) was calculated, and regions accounting for at least 1.0% of the total additive genetic variance were used to search for positional candidate genes. Heritability estimates for the studied traits ranged from 0.10 (SCS) to 0.53 (CNP), 0.10 (SCS) to 0.50 (CNP), and 0.12 (SCS) to 0.49 (CNP) in the first, second, and third parity, respectively. Genome-wide association analyses identified 6 genomic regions (BTA1, BTA14 [4 regions], and BTA20) associated with the considered traits. Genes including the SLC37A1 (BTA1), SHARPIN, MROH1, DGAT1, FAM83H, TIGD5, MROH6, NAPRT, ADGRB1, GML, LYPD2, JRK (BTA14), and TRIO (BTA20) were identified as positional candidate genes for the studied CMP. The findings of this study help to unravel the genomic background of a cow's ability for cheese production and can be used for the future implementation and use of genomic evaluation to improve the cheese-making traits in Walloon Holstein cows.

Identification of Genetic Effects of ACADVL and IRF6 Genes with Milk Production Traits of Holstein Cattle in China

With the development of high-throughput sequencing, RNA sequencing has been widely used in the identification of candidate genes for complex traits in livestock, and the functional genes and mutations with large genetic effects on milk production traits can provide molecular information for marker-assisted selection to increase the selection accuracy and accelerate genetic gain in dairy cattle. Our previous study on the liver transcriptome of Holstein cows found that acyl-CoA dehydrogenase (ACADVL) and interferon regulatory factor 6 (IRF6) are differentially expressed between dry and peak lactation periods, as well as that they are involved in lipid metabolism and the proliferation and differentiation of mammary epithelial cells. Thus, the two genes were considered candidates for milk traits. Hence, this study further collected 1186 Holstein cows from 110 sire families to investigate their genetic associations with milk yield and composition traits. By resequencing the entire exons and 2000 bp of the 5' and 3' flanking regions of the two genes, we identified eight SNPs in ACADVL and eight SNPs in IRF6. Subsequent single-locus association analyses showed that the eight SNPs in ACADVL were all significantly associated with milk fat yield, fat percentage, and protein yield (p values ≤ 0.0001-0.0414), and the eight SNPs in IRF6 were associated with milk, fat, and protein yields in the first or second lactation (p values ≤ 0.0001-0.0467). Using Haploview 4.2, one haplotype block with eight of the SNPs in ACADVL (D' = 0.99-1.00) and two haplotype blocks in IRF6 with three of the SNPs in each were observed (D' = 0.98-1.00). Similarly, the haplotype combinations of ACADVL were significantly associated with milk yield, fat percentage, fat yield, and protein yield in the two lactations (p values ≤ 0.0001-0.0125), and those of IRF6 were associated with five milk traits (p values ≤ 0.0001-0.0263). Furthermore, with the JASPAR software, it was predicted that the SNPs 19:g.26933503T>C in ACADVL and 16:g.73501985G>A in IRF6 changed the transcription factor binding sites of ZEB1, PLAGL2, and RHOXF1, implying their impacts on the expressions of the corresponding genes. Our findings demonstrated that the ACADVL and IRF6 genes have significant genetic effects on milk yield and composition traits, and the valuable SNPs might be used as genetic markers for genomic selection programs in dairy cattle.

Runs of homocigosity and its association with productive traits in Mexican Holstein cattle

The objective of this study was to describe the runs of homozygosity (ROH) detected in the Mexican Holstein population and to associate them with milk, fat and protein yields, and conformation final score. After imputation and genomic quality control, 4,227 genotyped animals with 100,806 SNPs markers each were used. ROH with a minimum length of 1 Mb and a minimum of 10 SNPs were included in the analysis. One heterozygous SNP marker and five missing genotypes per ROH were allowed. A total of 425,098 ROH were found in the studied population (71.83 ± 10.73 ROH per animal), with an average length and coverage of 4.80 ± 0.77 Mb, and 276.89 Mb, respectively. The average chromosome length covered by ROH was 10.40 ± 3.70 Mb. ROH between 1 and 2 Mb were the most frequent in the population (51.33%) while those between 14 and 16 Mb were the least frequent (1.20%). Long chromosomes showed a larger number of ROH. Chromosomes 10 and 20, had a greater percentage of their length covered by ROH because they presented a largest number of long ROH (>8 Mb). From the total ROH, 17 were detected in 1,847 animals and distributed among different chromosomes, and were associated with milk, fat and protein yield and percentage, and conformation final score. Of the ROH with effects on production traits, the majority were found with a length between 1 and 4 Mb. These results show evidence of genomic regions preserved by genetic selection and associated with the improvement of the productivity and functionality of dairy cattle.

Direct Phenotyping and Principal Component Analysis of Type Traits Implicate Novel QTL in Bovine Mastitis through Genome-Wide Association

Simple Summary

It is well established that the physical conformation of a cow’s udder and teats may influence her susceptibility to mastitis, an inflammatory condition of the udder, which has 25% prevalence in the United States. Our aim was to improve the biological understanding of the genetics underlying mastitis by intensively characterizing cows for udder and teat conformation, including the novel traits of teat width and end shape, and directly associating those phenotypes with high-density genotypes for those exact same cows. We also generated a composite measure that accounts for multiple high-mastitis-risk udder and teat conformations in a single index for risk phenotypes. Using this approach, we identified novel genetic markers associated with udder and teat conformation, which may be good candidates for inclusion in national genetic evaluations for selection of mastitis-resistant cows. Mastitis is the costliest disease facing US dairy producers, and integrating genetic information regarding disease susceptibility into breeding programs may be an efficient way to mitigate economic loss, support the judicious use of antimicrobials, and improve animal welfare.

Abstract

Our objectives were to robustly characterize a cohort of Holstein cows for udder and teat type traits and perform high-density genome-wide association studies for those traits within the same group of animals, thereby improving the accuracy of the phenotypic measurements and genomic association study. Additionally, we sought to identify a novel udder and teat trait composite risk index to determine loci with potential pleiotropic effects related to mastitis. This approach was aimed at improving the biological understanding of the genetic factors influencing mastitis. Cows (N = 471) were genotyped on the Illumina BovineHD777k beadchip and scored for front and rear teat length, width, end shape, and placement; fore udder attachment; udder cleft; udder depth; rear udder height; and rear udder width. We used principal component analysis to create a single composite measure describing type traits previously linked to high odds of developing mastitis within our cohort of cows. Genome-wide associations were performed, and 28 genomic regions were significantly associated (Bonferroni-corrected p < 0.05). Interrogation of these genomic regions revealed a number of biologically plausible genes whicht may contribute to the development of mastitis and whose functions range from regulating cell proliferation to immune system signaling, including ZNF683, DHX9, CUX1, TNNT1, and SPRY1. Genetic investigation of the risk composite trait implicated a novel locus and candidate genes that have potentially pleiotropic effects related to mastitis.

Time- and population-dependent genetic patterns underlie bovine milk somatic cell count

The objective of this study was to determine whether genetic regulation of bovine milk somatic cell count (SCC) varied throughout the course of an individual lactation and to identify quantitative trait loci (QTL) that may differentiate populations of chronically mastitic and robustly healthy cows. Milk SCC has long been a proxy for clinical mastitis diagnosis in management and genetic improvement strategies to control the disease. Cows (n = 471) were genotyped on the Illumina BovineHD 777K BeadChip (Illumina Inc., San Diego, CA), and composite milk samples were collected for SCC at 0-1 d in milk (DIM), 3-5 DIM, 10-14 DIM, 90-110 DIM, and 210-230 DIM, with each time span representing key physiological transitions for the cow. Median lactation somatic cell score (SCS) and area under the SCS curve were calculated from farm test data. A total of 8 genome-wide associations were performed and 167 SNP spanning the genome were significantly associated (false discovery rate <0.05). Of these associated regions, 27 of 48 associated QTL were novel for clinical mastitis or SCC. The linkage disequilibrium block surrounding the associated QTL or a 1-Mb window in the absence of linkage disequilibrium was interrogated for candidate genes, and many of those identified were related to multiple arms of the immune system, including toll-like receptor signaling, macrophage activation, B-cell maturation, T-cell recruitment, and the complement pathway. These genes included EXOC4, BAMBI, ITSN2, IL34, FCN3, CD8A, and CD8B. In addition, we identified populations of robustly healthy (SCS ≤4 from 10-14 DIM until study end), chronically mastitic (SCS >4 from 10-14 DIM until study end), and average cows with fluctuating SCS, and calculated fixation indices to identify regions of the genome differentiating these 3 populations. A total of 12 SNP were identified that showed moderate allelic differentiation (Wright's F statistic, F_ST ≥ 0.4) between the "chronic," "healthy," and "average" populations of cows. Candidate genes in the region surrounding differentiated QTL were related to cell signaling and immune response, such as JAKMIP1 and MADCAM1. The wide range of significantly associated QTL spanning the genome and the diversity of gene functions reinforces that mastitis is a complex trait and suggests that selection based on lactation stage-specific SCS rather than a generalized score may lead to greater success in breeding mastitis-resistant cows.

SNV discovery and functional candidate gene identification for milk composition based on whole genome resequencing of Holstein bulls with extremely high and low breeding values

We have sequenced the whole genomes of eight proven Holstein bulls from the four half-sib or full-sib families with extremely high and low estimated breeding values (EBV) for milk protein percentage (PP) and fat percentage (FP) using Illumina re-sequencing technology. Consequently, 2.3 billion raw reads were obtained with an average effective depth of 8.1×. After single nucleotide variant (SNV) calling, total 10,961,243 SNVs were identified, and 57,451 of them showed opposite fixed sites between the bulls with high and low EBVs within each family (called as common differential SNVs). Next, we annotated the common differential SNVs based on the bovine reference genome, and observed that 45,188 SNVs (78.70%) were located in the intergenic region of genes and merely 11,871 SNVs (20.67%) located within the protein-coding genes. Of them, 13,099 common differential SNVs that were within or close to protein-coding genes with less than 5 kb were chosen for identification of candidate genes for milk compositions in dairy cattle. By integrated analysis of the 2,657 genes with the GO terms and pathways related to protein and fat metabolism, and the known quantitative trait loci (QTLs) for milk protein and fat traits, we identified 17 promising candidate genes: ALG14, ATP2C1, PLD1, C3H1orf85, SNX7, MTHFD2L, CDKN2D, COL5A3, FDX1L, PIN1, FIG4, EXOC7, LASP1, PGS1, SAO, GPLD1 and MGEA5. Our findings provided an important foundation for further study and a prompt for molecular breeding of dairy cattle.

Integrated analysis of long noncoding RNA and mRNA expression profiles reveals the potential role of long noncoding RNA in different bovine lactation stages

Long noncoding RNA (lncRNA) play a critical role in mammary development and breast cancer biology. Despite their important role in the mammary gland, little is known of the roles of lncRNA in bovine lactation, particularly regarding the molecular processes underlying it. To characterize the role of lncRNA in bovine lactation, 4 samples of Holstein cow mammary gland tissue at peak and late lactation stages were examined after biopsy. We then profiled the transcriptome of the mammary gland using RNA sequencing technology. Further, functional lncRNA-mRNA coexpression pairs were constructed to infer the function of lncRNA using a generalized linear model, followed by gene ontology and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. More than 1,000 putative lncRNA were identified, 117 of which were differentially expressed between peak and late lactation stages. Bovine lncRNA were shorter, with fewer exon numbers, and expressed at significantly lower levels than protein-coding genes. Seventy-two differentially expressed (DE) lncRNA were coexpressed with 340 different protein-coding genes. The KEGG pathway analysis showed that target mRNA for DE lncRNA were mainly related to lipid and glucose metabolism, including the peroxisome proliferator-activated receptors and 5' adenosine monophosphate-activated protein kinase signaling pathways. Further bioinformatics and integrative analyses revealed that 12 DE lncRNA potentially played important roles in bovine lactation. Our findings provide a valuable resource for future bovine transcriptome studies, facilitate the understanding of bovine lactation biology, and offer functional information for cattle lactation.

Genome-wide association mapping for type and mammary health traits in French dairy goats identifies a pleiotropic region on chromosome 19 in the Saanen breed

Type traits and mammary health traits are important to dairy ruminant breeding because they influence animal health, milking ability, and longevity, as well as the economic sustainability of farms. The availability of the genomic sequence and a single nucleotide polymorphism chip in goats has opened up new fields of investigation to better understand the genes and mechanisms that underlie such complex traits and to be able to select them. Our objective was to perform a genome-wide association study in dairy goats for 11 type traits and somatic cell count (SCC) as proxies for mastitis resistance. A genome-wide association study was implemented using a daughter design composed of 1,941 Alpine and Saanen goats sired by 20 artificial insemination bucks, genotyped with the Illumina GoatSNP50 BeadChip (Illumina Inc., San Diego, CA). This association study was based on both linkage analyses and linkage disequilibrium using QTLmap software (http://dga7.jouy.inra.fr/qtlmap/) interval mapping was performed with the likelihood ratio test using linear regressions. Breeds were analyzed together and separately. The study highlighted 37 chromosome-wide significant quantitative trait loci (QTL) with linkage analyses and 222 genome-wide significant QTL for linkage disequilibrium, for type and SCC traits in dairy goats. Genomic control of those traits was mostly polygenic and breed-specific, suggesting that within-breed selection would be favored for those traits. Of note, Capra hircus autosome (CHI) 19 appeared to be highly enriched in single nucleotide polymorphisms associated with type and SCC, with 2 highly significant regions in the Saanen breed. One region (33-42 Mb) was significantly associated with SCC and includes candidate genes associated with response to intramammary infections (RARA, STAT3, STAT5A, and STAT5B). Another region of the CHI 19 (24.5-27 Mb) exhibited an adverse pleiotropic effect on milk production (milk, fat yield, and protein yield) and udder traits (udder floor position and rear udder attachment) that agreed with the negative genetic correlations that exist between those 2 groups of traits. These QTL were not found in the Alpine breed. In Alpine, the 2 most significant regions were associated with chest depth on CHI 6 (45.8-46.0 Mb) and CHI 8 (80.7-81.1 Mb). These results will be helpful for goat selection in the future and could lead to identification of causal mutations.

Validating genome-wide associated signals for clinical mastitis in German Holstein cattle

A validation study for six genomic regions previously identified by a genome-wide association study for somatic cell score was conducted with data of clinical mastitis in German Holstein cattle. Out of 10 tested SNPs, five on chromosomes 6, 13 and 19 were significantly associated with clinical mastitis (P < 0.05). Three SNPs on chromosomes 6 and 19 had the same direction of effect as those previously reported in the initial genome-wide association study for somatic cell score. The other two SNPs on chromosome 13 had opposite effects. As well as validating associations within known QTL from previous studies, e.g. chromosomes 6 and 19, novel loci on chromosome 13 were confirmed. Promising candidate genes are, for example: deoxycytidine kinase, immunoglobulin J chain, vitamin D binding protein, forkhead box K2, sodium/hydrogen exchanger 8 and cytoplasmic nuclear factor of activated T-cells 2. Our confirmation study provides additional evidence for the functional role of the linked genomic regions to immune response. This information can be used as a basis for further functional studies for those potential genes.

A genome scan for milk production traits in dairy goats reveals two new mutations in Dgat1 reducing milk fat content

The quantity of milk and milk fat and proteins are particularly important traits in dairy livestock. However, little is known about the regions of the genome that influence these traits in goats. We conducted a genome wide association study in French goats and identified 109 regions associated with dairy traits. For a major region on chromosome 14 closely associated with fat content, the Diacylglycerol O-Acyltransferase 1 (DGAT1) gene turned out to be a functional and positional candidate gene. The caprine reference sequence of this gene was completed and 29 polymorphisms were found in the gene sequence, including two novel exonic mutations: R251L and R396W, leading to substitutions in the protein sequence. The R251L mutation was found in the Saanen breed at a frequency of 3.5% and the R396W mutation both in the Saanen and Alpine breeds at a frequencies of 13% and 7% respectively. The R396W mutation explained 46% of the genetic variance of the trait, and the R251L mutation 6%. Both mutations were associated with a notable decrease in milk fat content. Their causality was then demonstrated by a functional test. These results provide new knowledge on the genetic basis of milk synthesis and will help improve the management of the French dairy goat breeding program.

Single nucleotide polymorphism and haplotype effects associated with somatic cell score in German Holstein cattle

Background

To better understand the genetic determination of udder health, we performed a genome-wide association study (GWAS) on a population of 2354 German Holstein bulls for which daughter yield deviations (DYD) for somatic cell score (SCS) were available. For this study, we used genetic information of 44 576 informative single nucleotide polymorphisms (SNPs) and 11 725 inferred haplotype blocks.

Results

When accounting for the sub-structure of the analyzed population, 16 SNPs and 10 haplotypes in six genomic regions were significant at the Bonferroni threshold of P ≤ 1.14 × 10^-6. The size of the identified regions ranged from 0.05 to 5.62 Mb. Genomic regions on chromosomes 5, 6, 18 and 19 coincided with known QTL affecting SCS, while additional genomic regions were found on chromosomes 13 and X. Of particular interest is the region on chromosome 6 between 85 and 88 Mb, where QTL for mastitis traits and significant SNPs for SCS in different Holstein populations coincide with our results. In all identified regions, except for the region on chromosome X, significant SNPs were present in significant haplotypes. The minor alleles of identified SNPs on chromosomes 18 and 19, and the major alleles of SNPs on chromosomes 6 and X were favorable for a lower SCS. Differences in somatic cell count (SCC) between alternative SNP alleles reached 14 000 cells/mL.

Conclusions

The results support the polygenic nature of the genetic determination of SCS, confirm the importance of previously reported QTL, and provide evidence for the segregation of additional QTL for SCS in Holstein cattle. The small size of the regions identified here will facilitate the search for causal genetic variations that affect gene functions.

Transcriptional profiling of mammary gland in Holstein cows with extremely different milk protein and fat percentage using RNA sequencing

Background

Recently, RNA sequencing (RNA-seq) has rapidly emerged as a major transcriptome profiling system. Elucidation of the bovine mammary gland transcriptome by RNA-seq is essential for identifying candidate genes that contribute to milk composition traits in dairy cattle.

Results

We used massive, parallel, high-throughput, RNA-seq to generate the bovine transcriptome from the mammary glands of four lactating Holstein cows with extremely high and low phenotypic values of milk protein and fat percentage. In total, we obtained 48,967,376–75,572,578 uniquely mapped reads that covered 82.25% of the current annotated transcripts, which represented 15549 mRNA transcripts, across all the four mammary gland samples. Among them, 31 differentially expressed genes (p < 0.05, false discovery rate q < 0.05) between the high and low groups of cows were revealed. Gene ontology and pathway analysis demonstrated that the 31 differently expressed genes were enriched in specific biological processes with regard to protein metabolism, fat metabolism, and mammary gland development (p < 0.05). Integrated analysis of differential gene expression, previously reported quantitative trait loci, and genome-wide association studies indicated that TRIB3, SAA (SAA1, SAA3, and M-SAA3.2), VEGFA, PTHLH, and RPL23A were the most promising candidate genes affecting milk protein and fat percentage.

Conclusions

This study investigated the complexity of the mammary gland transcriptome in dairy cattle using RNA-seq. Integrated analysis of differential gene expression and the reported quantitative trait loci and genome-wide association study data permitted the identification of candidate key genes for milk composition traits.

Short communication: validation of somatic cell score-associated loci identified in a genome-wide association study in German Holstein cattle

Recently, we identified 6 genomic loci affecting daughter yield deviations (DYD) for somatic cell score (SCS) in a genome-wide association study (GWAS) performed with German Holstein bulls. In the current study, we tested if these loci were associated with SCS in cows using their own performance data. The study was performed with 1,412 German Holstein cows, of which 483 were daughters of 71 bulls that had been used in the GWAS. We tested 10 single nucleotide polymorphisms (SNP) representing 6 genomic regions that were associated with DYD for SCS in bulls. All tested SNP were significant in cows. Seven of them, located on Bos taurus autosomes (BTA) 6, 13, and 19, had the same direction of effect as those previously reported in the bull population. The most significant associations were detected on BTA6 and BTA19, accounting for 1.8% of the total genetic variance. The major allele of the 2 SNP on BTA6 and the minor allele of the 2 SNP on BTA19 were favorable for lower SCS. The differences between the homozygous genotype classes were up to 15,000 cells/mL. The verification of SNP associated with SCS in this study provides further evidence for the functional role of the linked genomic regions for immune response and contributes to identification of causative mutations. In particular, SNP with minor frequency of the favorable allele possess high potential to reduce SCS in German Holstein cattle by selection.

A genome-wide association study for somatic cell score using the Illumina high-density bovine beadchip identifies several novel QTL potentially related to mastitis susceptibility

Mastitis is an inflammation-driven disease of the bovine mammary gland that occurs in response to physical damage or infection and is one of the most costly production-related diseases in the dairy industry worldwide. We performed a genome-wide association study (GWAS) to identify genetic loci associated with somatic cell score (SCS), an indicator trait of mammary gland inflammation. A total of 702 Holstein-Friesian bulls were genotyped for 777,962 single nucleotide polymorphisms (SNPs) and associated with SCS phenotypes. The SCS phenotypes were expressed as daughter yield deviations (DYD) based on a large number of progeny performance records. A total of 138 SNPs on 15 different chromosomes reached genome-wide significance (corrected p-value ≤ 0.05) for association with SCS (after correction for multiple testing). We defined 28 distinct QTL regions and a number of candidate genes located in these QTL regions were identified. The most significant association (p-value = 1.70 × 10⁻⁷) was observed on chromosome 6. This QTL had no known genes annotated within it, however, the Ensembl Genome Browser predicted the presence of a small non-coding RNA (a Y RNA gene) in this genomic region. This Y RNA gene was 99% identical to human RNY4. Y RNAs are a rare type of non-coding RNA that were originally discovered due to their association with the autoimmune disease, systemic lupus erythematosus. Examining small-RNA sequencing (RNAseq) data being generated by us in multiple different mastitis-pathogen challenged cell-types has revealed that this Y RNA is expressed (but not differentially expressed) in these cells. Other QTL regions identified in this study also encoded strong candidate genes for mastitis susceptibility. A QTL region on chromosome 13, for example, was found to contain a cluster of β-defensin genes, a gene family with known roles in innate immunity. Due to the increased SNP density, this study also refined the boundaries for several known QTL for SCS and mastitis.

Genome-wide association analysis to identify genotype × environment interaction for milk protein yield and level of somatic cell score as environmental descriptors in German Holsteins

Genotype by environment interaction (G × E) has been widely reported in dairy cattle. If the environment can be measured on a continuous scale, reaction norms can be applied to study G × E. The average herd milk production level has frequently been used as an environmental descriptor because it is influenced by the level of feeding or the feeding regimen. Another important environmental factor is the level of udder health and hygiene, for which the average herd somatic cell count might be a descriptor. In the present study, we conducted a genome-wide association analysis to identify single nucleotide polymorphisms (SNP) that affect intercept and slope of milk protein yield reaction norms when using the average herd test-day solution for somatic cell score as an environmental descriptor. Sire estimates for intercept and slope of the reaction norms were calculated from around 12 million daughter records, using linear reaction norm models. Sires were genotyped for ~54,000 SNP. The sire estimates were used as observations in the association analysis, using 1,797 sires. Significant SNP were confirmed in an independent validation set consisting of 500 sires. A known major gene affecting protein yield was included as a covariable in the statistical model. Sixty (21) SNP were confirmed for intercept with P ≤ 0.01 (P ≤ 0.001) in the validation set, and 28 and 11 SNP, respectively, were confirmed for slope. Most but not all SNP affecting slope also affected intercept. Comparison with an earlier study revealed that SNP affecting slope were, in general, also significant for slope when the environment was modeled by the average herd milk production level, although the two environmental descriptors were poorly correlated.

Using genome-wide association analysis to characterize environmental sensitivity of milk traits in dairy cattle

Genotype-by-environment interaction (GxE) has been widely reported in dairy cattle. One way to analyze GxE is to apply reaction norm models. The first derivative of a reaction norm is the environmental sensitivity (ES). In the present study we conducted a large-scale, genome-wide association analysis to identify single-nucleotide polymorphisms (SNPs) that affect general production (GP) and ES of milk traits in the German Holstein population. Sire estimates for GP and for ES were calculated from approximately 13 million daughter records by the use of linear reaction norm models. The daughters were offspring from 2297 sires. Sires were genotyped for 54k SNPs. The environment was defined as the average milk energy yield performance of the herds at the time during which the daughter observations were recorded. The sire estimates were used as observations in a genome-wide association analysis, using 1797 sires. Significant SNPs were confirmed in an independent validation set (500 sires of the same population). To separate GxE scaling and other GxE effects, the observations were log-transformed in some analyses. Results from the reaction norm model revealed GxE effects. Numerous significant SNPs were validated for both GP and ES. Many SNPs that affect GP also affect ES. We showed that ES of milk traits is a typical quantitative trait, genetically controlled by many genes with small effects and few genes with larger effect. A log-transformation of the observation resulted in a reduced number of validated SNPs for ES, pointing to genes that not only caused scaling GxE effects. The results will have implications for breeding for robustness in dairy cattle.

Genome-wide associations for milk production and somatic cell score in Holstein-Friesian cattle in Ireland

BACKGROUND

Contemporary dairy breeding goals have broadened to include, along with milk production traits, a number of non-production-related traits in an effort to improve the overall functionality of the dairy cow. Increased indirect selection for resistance to mastitis, one of the most important production-related diseases in the dairy sector, via selection for reduced somatic cell count has been part of these broadened goals. A number of genome-wide association studies have identified genetic variants associated with milk production traits and mastitis resistance, however the majority of these studies have been based on animals which were predominantly kept in confinement and fed a concentrate-based diet (i.e. high-input production systems). This genome-wide association study aims to detect associations using genotypic and phenotypic data from Irish Holstein-Friesian cattle fed predominantly grazed grass in a pasture-based production system (low-input).

RESULTS

Significant associations were detected for milk yield, fat yield, protein yield, fat percentage, protein percentage and somatic cell score using separate single-locus, frequentist and multi-locus, Bayesian approaches. These associations were detected using two separate populations of Holstein-Friesian sires and cows. In total, 1,529 and 37 associations were detected in the sires using a single SNP regression and a Bayesian method, respectively. There were 103 associations in common between the sires and cows across all the traits. As well as detecting associations within known QTL regions, a number of novel associations were detected; the most notable of these was a region of chromosome 13 associated with milk yield in the population of Holstein-Friesian sires.

CONCLUSIONS

A total of 276 of novel SNPs were detected in the sires using a single SNP regression approach. Although obvious candidate genes may not be initially forthcoming, this study provides a preliminary framework upon which to identify the causal mechanisms underlying the various milk production traits and somatic cell score. Consequently this will deepen our understanding of how these traits are expressed.

BRCA1: a new candidate gene for bovine mastitis and its association analysis between single nucleotide polymorphisms and milk somatic cell score

Bovine mastitis is a very complex and common disease of dairy cattle and a major source of economic losses to the dairy industry worldwide. In this study, the bovine breast cancer 1, early onset gene (BRCA1) was taken as a candidate gene for mastitis resistance. The main object of this study was to investigate whether the BRCA1 gene was associated with mastitis in cattle. Through DNA sequencing, Polymerase Chain Reaction-Restriction Fragment Length Polymorphism (PCR-RFLP) and Created Restriction Site PCR (CRS-PCR) methods, three SNPs (G22231T, T25025A, and C28300A) were detected and twenty-four combinations of these SNPs were observed. The single SNP and their genetic effects on somatic cell score (SCS) were evaluated and a significant association with SCS was found in C28300A. The mean of genotype EE was significantly lower than those of genotypes EF and FF. The results of combined genotypes analysis of three SNPs showed that BBDDFF genotype with the highest SCS were easily for the mastitis susceptibility, whereas AACCEE genotype with the lowest SCS were favorable for the mastitis resistance. The information provided in the present study will be very useful for improving mastitis resistance in dairy cattle by marker-assisted selection.

Mapping multiple quantitative trait loci under Bayes error control

In mapping of quantitative trait loci (QTLs), performing hypothesis tests of linkage to a phenotype of interest across an entire genome involves multiple comparisons. Furthermore, linkage among loci induces correlation among tests. Under many multiple comparison frameworks, these problems are exacerbated when mapping multiple QTLs. Traditionally, significance thresholds have been subjectively set to control the probability of detecting at least one false positive outcome, although such thresholds are known to result in excessively low power to detect true positive outcomes. Recently, false discovery rate (FDR)-controlling procedures have been developed that yield more power both by relaxing the stringency of the significance threshold and by retaining more power for a given significance threshold. However, these procedures have been shown to perform poorly for mapping QTLs, principally because they ignore recombination fractions between markers. Here, I describe a procedure that accounts for recombination fractions and extends FDR control to include simultaneous control of the false non-discovery rate, i.e. the overall error rate is controlled. This procedure is developed in the Bayesian framework using a direct posterior probability approach. Data-driven significance thresholds are determined by minimizing the expected loss. The procedure is equivalent to jointly maximizing positive and negative predictive values. In the context of mapping QTLs for experimental crosses, the procedure is applicable to mapping main effects, gene-gene interactions and gene-environment interactions.

The number of loci that affect milk production traits in dairy cattle

We have used the results of an experiment mapping quantitative trait loci (QTL) affecting milk yield and composition to estimate the total number of QTL affecting these traits. We did this by estimating the number of segregating QTL within a half-sib daughter design using logic similar to that used to estimate the "false discovery rate" (FDR). In a half-sib daughter design with six sire families we estimate that the average sire was heterozygous for approximately 5 QTL per trait. Also, in most cases only one sire was heterozygous for any one QTL; therefore at least 30 QTL were likely to be segregating for these milk production traits in this Holstein population.

Relaxed significance criteria for linkage analysis

Linkage analysis involves performing significance tests at many loci located throughout the genome. Traditional criteria for declaring a linkage statistically significant have been formulated with the goal of controlling the rate at which any single false positive occurs, called the genomewise error rate (GWER). As complex traits have become the focus of linkage analysis, it is increasingly common to expect that a number of loci are truly linked to the trait. This is especially true in mapping quantitative trait loci (QTL), where sometimes dozens of QTL may exist. Therefore, alternatives to the strict goal of preventing any single false positive have recently been explored, such as the false discovery rate (FDR) criterion. Here, we characterize some of the challenges that arise when defining relaxed significance criteria that allow for at least one false positive linkage to occur. In particular, we show that the FDR suffers from several problems when applied to linkage analysis of a single trait. We therefore conclude that the general applicability of FDR for declaring significant linkages in the analysis of a single trait is dubious. Instead, we propose a significance criterion that is more relaxed than the traditional GWER, but does not appear to suffer from the problems of the FDR. A generalized version of the GWER is proposed, called GWERk, that allows one to provide a more liberal balance between true positives and false positives at no additional cost in computation or assumptions.