Genome-wide association study of age at puberty and its (co)variances with fertility and stature in growing and lactating Holstein-Friesian dairy cattle

Reproductive performance is a key determinant of cow longevity in a pasture-based, seasonal dairy system. Unfortunately, direct fertility phenotypes such as intercalving interval or pregnancy rate tend to have low heritabilities and occur relatively late in an animal's life. In contrast, age at puberty (AGEP) is a moderately heritable, early-in-life trait that may be estimated using an animal's age at first measured elevation in blood plasma progesterone (AGEP4) concentrations. Understanding the genetic architecture of AGEP4 in addition to genetic relationships between AGEP4 and fertility traits in lactating cows is important, as is its relationship with body size in the growing animal. Thus, the objectives of this research were 3-fold. First, to estimate the genetic and phenotypic (co)variances between AGEP4 and subsequent fertility during first and second lactations. Second, to quantify the associations between AGEP4 and height, length, and BW measured when animals were approximately 11 mo old (standard deviation = 0.5). Third, to identify genomic regions that are likely to be associated with variation in AGEP4. We measured AGEP4, height, length, and BW in approximately 5,000 Holstein-Friesian or Holstein-Friesian × Jersey crossbred yearling heifers across 54 pasture-based herds managed in seasonal calving farm systems. We also obtained calving rate (CR42, success or failure to calve within the first 42 d of the seasonal calving period), breeding rate (PB21, success or failure to be presented for breeding within the first 21 d of the seasonal breeding period) and pregnancy rate (PR42, success or failure to become pregnant within the first 42 d of the seasonal breeding period) phenotypes from their first and second lactations. The animals were genotyped using the Weatherby's Versa 50K SNP array (Illumina, San Diego, CA). The estimated heritabilities of AGEP4, height, length, and BW were 0.34 (90% credibility interval [CRI]: 0.30, 0.37), 0.28 (90% CRI: 0.25, 0.31), 0.21 (90% CRI: 0.18, 0.23), and 0.33 (90% CRI: 0.30, 0.36), respectively. In contrast, the heritabilities of CR42, PB21 and PR42 were all <0.05 in both first and second lactations. The genetic correlations between AGEP4 and these fertility traits were generally moderate, ranging from 0.11 to 0.60, whereas genetic correlations between AGEP4 and yearling body-conformation traits ranged from 0.02 to 0.28. Our GWAS highlighted a genomic window on chromosome 5 that was strongly associated with variation in AGEP4. We also identified 4 regions, located on chromosomes 14, 6, 1, and 11 (in order of decreasing importance), that exhibited suggestive associations with AGEP4. Our results show that AGEP4 is a reasonable predictor of estimated breeding values for fertility traits in lactating cows. Although the GWAS provided insights into genetic mechanisms underpinning AGEP4, further work is required to test genomic predictions of fertility that use this information.

Genome-wide association study of anogenital distance and its (co)variances with fertility in growing and lactating Holstein-Friesian dairy cattle

Anogenital distance (AGD) is a moderately heritable trait that can be measured at a young age that may provide an opportunity to indirectly select for improved fertility in dairy cattle. In this study, we characterized AGD and its genetic and phenotypic relationships with a range of body stature and fertility traits. We measured AGD, shoulder height, body length, and body weight in a population of 5,010 Holstein-Friesian and Holstein-Friesian × Jersey crossbred heifers at approximately 11 mo of age (AGD1). These animals were born in 2018 across 54 seasonal calving, pasture-based dairy herds. A second measure of AGD was collected in a subset of herds (n = 17; 1,956 animals) when the animals averaged 29 mo of age (AGD2). Fertility measures included age at puberty (AGEP), then time of calving, breeding, and pregnancy during the first and second lactations. We constructed binary traits reflecting the animal's ability to calve during the first 42 d of their herd's seasonal calving period (CR42), be presented for breeding during the first 21 d of the seasonal breeding period (PB21) and become pregnant during the first 42 d of the seasonal breeding period (PR42). The posterior mean of sampled heritabilities for AGD1 was 0.23, with 90% of samples falling within a credibility interval (90% CRI) of 0.20 to 0.26, whereas the heritability of AGD2 was 0.29 (90% CRI 0.24 to 0.34). The relationship between AGD1 and AGD2 was highly positive, with a genetic correlation of 0.89 (90% CRI 0.82 to 0.94). Using a GWAS analysis of 2,460 genomic windows based on 50k genotype data, we detected a region on chromosome 20 that was highly associated with variation in AGD1, and a second region on chromosome 13 that was moderately associated with variation in AGD1. We did not detect any genomic regions associated with AGD2 which was measured in fewer animals. The genetic correlation between AGD1 and AGEP was 0.10 (90% CRI 0.00 to 0.19), whereas the genetic correlation between AGD2 and AGEP was 0.30 (90% CRI 0.15 to 0.44). The timing of calving, breeding, and pregnancy (CR42, PB21, and PR42) during first or second lactations exhibited moderate genetic relationships with AGD1 (0.19 to 0.52) and AGD2 (0.46 to 0.63). Genetic correlations between AGD and body stature traits were weak (≤0.16). We conclude that AGD is a moderately heritable trait, which may have value as an early-in-life genetic predictor for reproductive success during lactation.

Comparison of methods for deriving phenotypes from incomplete observation data with an application to age at puberty in dairy cattle

BACKGROUND

Many phenotypes in animal breeding are derived from incomplete measures, especially if they are challenging or expensive to measure precisely. Examples include time-dependent traits such as reproductive status, or lifespan. Incomplete measures for these traits result in phenotypes that are subject to left-, interval- and right-censoring, where phenotypes are only known to fall below an upper bound, between a lower and upper bound, or above a lower bound respectively. Here we compare three methods for deriving phenotypes from incomplete data using age at first elevation (> 1 ng/mL) in blood plasma progesterone (AGEP4), which generally coincides with onset of puberty, as an example trait.

METHODS

We produced AGEP4 phenotypes from three blood samples collected at about 30-day intervals from approximately 5,000 Holstein-Friesian or Holstein-Friesian × Jersey cross-bred dairy heifers managed in 54 seasonal-calving, pasture-based herds in New Zealand. We used these actual data to simulate 7 different visit scenarios, increasing the extent of censoring by disregarding data from one or two of the three visits. Three methods for deriving phenotypes from these data were explored: 1) ordinal categorical variables which were analysed using categorical threshold analysis; 2) continuous variables, with a penalty of 31 d assigned to right-censored phenotypes; and 3) continuous variables, sampled from within a lower and upper bound using a data augmentation approach.

RESULTS

Credibility intervals for heritability estimations overlapped across all methods and visit scenarios, but estimated heritabilities tended to be higher when left censoring was reduced. For sires with at least 5 daughters, the correlations between estimated breeding values (EBVs) from our three-visit scenario and each reduced data scenario varied by method, ranging from 0.65 to 0.95. The estimated breed effects also varied by method, but breed differences were smaller as phenotype censoring increased.

CONCLUSION

Our results indicate that using some methods, phenotypes derived from one observation per offspring for a time-dependent trait such as AGEP4 may provide comparable sire rankings to three observations per offspring. This has implications for the design of large-scale phenotyping initiatives where animal breeders aim to estimate variance parameters and estimated breeding values (EBVs) for phenotypes that are challenging to measure or prohibitively expensive.

Genetic and phenotypic relationships between ewe reproductive performance and wool and growth traits in Uruguayan Ultrafine Merino sheep

This study reports genetic parameters for yearling and adult wool and growth traits, and ewe reproductive performance. Data were sourced from an Uruguayan Merino flock involved in a long-term selection program focused on reduced fibre diameter (FD), and increased clean fleece weight (CFW) and live weight (LW). Pedigree and performance data from approximately 5,700 mixed-sex yearling lambs and 2,000 mixed-age ewes born between 1999 and 2019 were analysed. The number of records ranged from 1,267 to 5,738 for yearling traits, and from 1,931 to 7,079 for ewe productive and reproductive performance. Data on yearling and adult wool traits, LW and body condition score (BCS), yearling eye muscle area (Y_EMA), and fat thickness (Y_FAT), and several reproduction traits were analysed. The genetic relationships between FD and reproduction traits were not different from zero. Moderate unfavourable genetic correlations were found between adult CFW and ewe lifetime reproduction traits (-0.34 ± 0.08 and -0.33 ± 0.09 for total number of lambs weaned and total lamb live weight at weaning, respectively). There were moderate to strong positive genetic correlations between yearling LW and all reproduction traits other than ewe rearing ability (-0.08 ± 0.11) and pregnancy rate (0.18 ± 0.08). The genetic correlations between Y_EMA and reproduction traits were positive and ranged from 0.15 to 0.49. Moderate unfavourable genetic correlations were observed between yearling FD and Y_FAT and between adult FD and BCS at mating (0.31 ± 0.12 and 0.23 ± 0.07, respectively). The genetic correlations between adult fleece weight and ewe BCS at different stages of the cycle were negative, but generally not different from zero. This study shows that selection for reduced FD is unlikely to have any effect on reproduction traits. Selection for increased yearling LW and Y_EMA will improve ewe reproductive performance. On the other hand, selection for increased adult CFW will reduce ewe reproductive performance, whereas selection for reduced FD will negatively impact on body fat levels. Although unfavourable genetic relationships between wool traits and both FAT and ewe reproductive performance existed, simultaneous improvements in the traits would occur using appropriately designed indexes.

Genomic Regions Associated with Wool, Growth and Reproduction Traits in Uruguayan Merino Sheep

The aim of this study was to identify genomic regions and genes associated with the fiber diameter (FD), clean fleece weight (CFW), live weight (LW), body condition score (BCS), pregnancy rate (PR) and lambing potential (LP) of Uruguayan Merino sheep. Phenotypic records of approximately 2000 mixed-age ewes were obtained from a Merino nucleus flock. Genome-wide association studies were performed utilizing single-step Bayesian analysis. For wool traits, a total of 35 genomic windows surpassed the significance threshold (PVE ≥ 0.25%). The proportion of the total additive genetic variance explained by those windows was 4.85 and 9.06% for FD and CFW, respectively. There were 42 windows significantly associated with LWM, which collectively explained 43.2% of the additive genetic variance. For BCS, 22 relevant windows accounted for more than 40% of the additive genetic variance, whereas for the reproduction traits, 53 genomic windows (24 and 29 for PR and LP, respectively) reached the suggestive threshold of 0.25% of the PVE. Within the top 10 windows for each trait, we identified several genes showing potential associations with the wool (e.g., IGF-1, TGFB2R, PRKCA), live weight (e.g., CAST, LAP3, MED28, HERC6), body condition score (e.g., CDH10, TMC2, SIRPA, CPXM1) or reproduction traits (e.g., ADCY1, LEPR, GHR, LPAR2) of the mixed-age ewes.

A method to obtain exact single-step GBLUP for non-genotyped descendants when the genomic relationship matrix of ancestors is not available

BACKGROUND

Single-step genomic best linear unbiased prediction (GBLUP) involves a joint analysis of individuals with genotype information, and their ancestors, descendants, or contemporaries, without recorded genotypes. Livestock applications typically represent populations with fewer individuals with genotypes relative to the number not genotyped. Most breeding programmes are structured, consisting of a nucleus tier in which selection drives genetic gains that are propagated through descendants that represent parents in multiplier and commercial tiers. In some cases, the genotypes in the nucleus tier are proprietary to a breeding company, and not publicly available for a whole industry analysis. Bayesian inference involves combining a defined description of prior information with new information to generate a posterior distribution that contains all available information on parameters of interest. A natural extension of Bayesian analysis would be to use information from the posterior distribution to define the prior distribution in a subsequent analysis.

METHODS

We derive the mixed model equations for inference on breeding values for non genotyped individuals in that subset of the population that is of current interest, using only data on the performance of current individuals and their immediate pedigree, along with prior information defined from the posterior distribution of an external BLUP or single-step GBLUP analysis of the ancestors of the current population.

DISCUSSION

Identical estimates of breeding values and their prediction error covariances for current animals of interest in the multiplier or commercial tier can be obtained without requiring neither the genomic relationship matrix nor genotypes of any of their ancestors in the nucleus tier, as can be obtained from a single analysis using pedigree, performance, and genomic information from all tiers. The Bayesian analysis of the current population does not require explicit information on unselected genotyped animals in the external population.

Estimating Heritabilities and Breeding Values From Censored Phenotypes Using a Data Augmentation Approach

Time-dependent traits are often subject to censorship, where instead of precise phenotypes, only a lower and/or upper bound can be established for some of the individuals. Censorship reduces the precision of phenotypes but can represent compromise between measurement cost and animal ethics considerations. This compromise is particularly relevant for genetic evaluation because phenotyping initiatives often involve thousands of individuals. This research aimed to: 1) demonstrate a data augmentation approach for analysing censored phenotypes, and 2) quantify the implications of phenotype censorship on estimation of heritabilities and predictions of breeding values. First, we simulated uncensored phenotypes, representing fine-scale “age at puberty” for each individual in a population of some 5,000 animals across 50 herds. Analysis of these uncensored phenotypes provided a gold-standard control. We then produced seven “test” phenotypes by superimposing varying degrees of left, interval, and/or right censorship, as if herds were measured on only one, two or three occasions, with a binary measure categorized for animals at each visit (either pre or post pubertal). We demonstrated that our estimates of heritabilities and predictions of breeding values obtained using a data augmentation approach were remarkably robust to phenotype censorship. Our results have important practical implications for measuring time-dependent traits for genetic evaluation. More specifically, we suggest that data collection can be designed with relatively infrequent repeated measures, thereby reducing costs and increasing feasibility across large numbers of animals.

Pathology of the peripheral neuropathy Charcot-Marie-Tooth disease type 4H in Holstein Friesian cattle with a splice site mutation in FGD4

Charcot-Marie-Tooth disease (CMT) is a hereditary sensory and motor peripheral neuropathy that is one of the most common inherited neurological diseases of humans and may be caused by mutations in a number of different genes. The subtype Charcot-Marie-Tooth disease type 4H (CMT4H) is caused by homozygous mutations in the FGD4 (FYVE, RhoGEF, and PH domain-containing 4) gene. A previous genome-wide association study involving 130,783 dairy cows found 6 novel variants, one of which was a homozygous splice site mutation in the FGD4 gene. Descendants of carriers were genotyped to identify 9 homozygous Holstein Friesian calves that were raised to maturity, of which 5 were euthanized and sampled for histopathology and electron microscopy at 2 and 2.5 years of age. Three control Holstein Friesian animals were raised with the calves and euthanized at the same time points. No macroscopic lesions consistent with CMT4H were seen at necropsy. Microscopically, peripheral nerves were hypercellular due to hyperplasia of S100-positive Schwann cells, and there was onion bulb formation, axonal degeneration with demyelination, and increased thickness of the endoneurium. On electron microscopy, decreased axonal density, onion bulb formations, myelin outfoldings, and increased numbers of mitochondria were present. These changes are consistent with those described in mouse models and humans with CMT4H, making these cattle a potential large animal model for CMT.

Erratum to "Cumulative dairy cow genetic change from selection and crossbreeding over the last 2 decades in New Zealand closely aligns to model-based predictions published in 2000" (JDS Commun. 2:51-54)

[This corrects the article DOI: 10.3168/jdsc.2020-0043.].

Pregnancy status predicted using milk mid-infrared spectra from dairy cattle

Accurate and timely pregnancy diagnosis is an important component of effective herd management in dairy cattle. Predicting pregnancy from Fourier-transform mid-infrared (FT-MIR) spectroscopy data is of particular interest because the data are often already available from routine milk testing. The purpose of this study was to evaluate how well pregnancy status could be predicted in a large data set of 1,161,436 FT-MIR milk spectra records from 863,982 mixed-breed pasture-based New Zealand dairy cattle managed within seasonal calving systems. Three strategies were assessed for defining the nonpregnant cows when partitioning the records according to pregnancy status in the training population. Two of these used records for cows with a subsequent calving only, whereas the third also included records for cows without a subsequent calving. For each partitioning strategy, partial least squares discriminant analysis models were developed, whereby spectra from all the cows in 80% of herds were used to train the models, and predictions on cows in the remaining herds were used for validation. A separate data set was also used as a secondary validation, whereby pregnancy diagnosis had been assigned according to the presence of pregnancy-associated glycoproteins (PAG) in the milk samples. We examined different ways of accounting for stage of lactation in the prediction models, either by including it as an effect in the prediction model, or by pre-adjusting spectra before fitting the model. For a subset of strategies, we also assessed prediction accuracies from deep learning approaches, utilizing either the raw spectra or images of spectra. Across all strategies, prediction accuracies were highest for models using the unadjusted spectra as model predictors. Strategies for cows with a subsequent calving performed well in herd-independent validation with sensitivities above 0.79, specificities above 0.91 and area under the receiver operating characteristic curve (AUC) values over 0.91. However, for these strategies, the specificity to predict nonpregnant cows in the external PAG data set was poor (0.002-0.04). The best performing models were those that included records for cows without a subsequent calving, and used unadjusted spectra and days in milk as predictors, with consistent results observed across the training, herd-independent validation and PAG data sets. For the partial least squares discriminant analysis model, sensitivity was 0.71, specificity was 0.54 and AUC values were 0.68 in the PAG data set; and for an image-based deep learning model, the sensitivity was 0.74, specificity was 0.52 and the AUC value was 0.69. Our results demonstrate that in pasture-based seasonal calving herds, confounding between pregnancy status and spectral changes associated with stage of lactation can inflate prediction accuracies. When the effect of this confounding was reduced, prediction accuracies were not sufficiently high enough to use as a sole indicator of pregnancy status.

Non-additive QTL mapping of lactation traits in 124,000 cattle reveals novel recessive loci

BACKGROUND

Deleterious recessive conditions have been primarily studied in the context of Mendelian diseases. Recently, several deleterious recessive mutations with large effects were discovered via non-additive genome-wide association studies (GWAS) of quantitative growth and developmental traits in cattle, which showed that quantitative traits can be used as proxies of genetic disorders when such traits are indicative of whole-animal health status. We reasoned that lactation traits in cattle might also reflect genetic disorders, given the increased energy demands of lactation and the substantial stresses imposed on the animal. In this study, we screened more than 124,000 cows for recessive effects based on lactation traits.

RESULTS

We discovered five novel quantitative trait loci (QTL) that are associated with large recessive impacts on three milk yield traits, with these loci presenting missense variants in the DOCK8, IL4R, KIAA0556, and SLC25A4 genes or premature stop variants in the ITGAL, LRCH4, and RBM34 genes, as candidate causal mutations. For two milk composition traits, we identified several previously reported additive QTL that display small dominance effects. By contrasting results from milk yield and milk composition phenotypes, we note differing genetic architectures. Compared to milk composition phenotypes, milk yield phenotypes had lower heritabilities and were associated with fewer additive QTL but had a higher non-additive genetic variance and were associated with a higher proportion of loci exhibiting dominance.

CONCLUSIONS

We identified large-effect recessive QTL which are segregating at surprisingly high frequencies in cattle. We speculate that the differences in genetic architecture between milk yield and milk composition phenotypes derive from underlying dissimilarities in the cellular and molecular representation of these traits, with yield phenotypes acting as a better proxy of underlying biological disorders through presentation of a larger number of major recessive impacts.

An inherited night blindness in Wiltshire sheep

Twelve cases of adult-onset blindness were identified in a flock of 130 polled Wiltshire sheep in New Zealand over a 3-year period. Affected sheep developed night blindness between 2 and 3 years of age, which progressed to complete blindness by 4 to 5 years of age. Fundic examination findings included progressive tapetal hyperreflectivity and attenuation of retinal blood vessels. Histologically, the retinas had a selective loss of rod photoreceptors with initial preservation of cone photoreceptors. Retinal degeneration was not accompanied by any other ocular or central nervous system abnormalities, and pedigree analysis suggested an inherited basis for the disease. Mating an affected Wiltshire ram to 2 affected Wiltshire ewes resulted in 6 progeny that all developed retinal degeneration by 2 years of age, while mating of the same affected ram to 6 unaffected ewes resulted in 8 unaffected progeny, consistent with autosomal recessive inheritance. Homozygosity mapping of 5 affected Wiltshire sheep and 1 unaffected Wiltshire sheep using an OvineSNP50 Genotyping BeadChip revealed an identical-by-descent region on chromosome 5, but none of the genes within this region were considered plausible candidate genes. Whole-genome sequencing of 2 affected sheep did not reveal any significant mutations in any of the genes associated with retinitis pigmentosa in humans or progressive retinal atrophy in dogs. Inherited progressive retinal degeneration affecting rod photoreceptors has not been previously reported in sheep, but this disease has several similarities to inherited retinal dystrophies in other species.

HandyCNV: Standardized Summary, Annotation, Comparison, and Visualization of Copy Number Variant, Copy Number Variation Region, and Runs of Homozygosity

Detection of CNVs (copy number variants) and ROH (runs of homozygosity) from SNP (single nucleotide polymorphism) genotyping data is often required in genomic studies. The post-analysis of CNV and ROH generally involves many steps, potentially across multiple computing platforms, which requires the researchers to be familiar with many different tools. In order to get around this problem and improve research efficiency, we present an R package that integrates the summarization, annotation, map conversion, comparison and visualization functions involved in studies of CNV and ROH. This one-stop post-analysis system is standardized, comprehensive, reproducible, timesaving, and user-friendly for researchers in humans and most diploid livestock species.

Sequence-based genome-wide association study of individual milk mid-infrared wavenumbers in mixed-breed dairy cattle

BACKGROUND

Fourier-transform mid-infrared (FT-MIR) spectroscopy provides a high-throughput and inexpensive method for predicting milk composition and other novel traits from milk samples. While there have been many genome-wide association studies (GWAS) conducted on FT-MIR predicted traits, there have been few GWAS for individual FT-MIR wavenumbers. Using imputed whole-genome sequence for 38,085 mixed-breed New Zealand dairy cattle, we conducted GWAS on 895 individual FT-MIR wavenumber phenotypes, and assessed the value of these direct phenotypes for identifying candidate causal genes and variants, and improving our understanding of the physico-chemical properties of milk.

RESULTS

Separate GWAS conducted for each of 895 individual FT-MIR wavenumber phenotypes, identified 450 1-Mbp genomic regions with significant FT-MIR wavenumber QTL, compared to 246 1-Mbp genomic regions with QTL identified for FT-MIR predicted milk composition traits. Use of mammary RNA-seq data and gene annotation information identified 38 co-localized and co-segregating expression QTL (eQTL), and 31 protein-sequence mutations for FT-MIR wavenumber phenotypes, the latter including a null mutation in the ABO gene that has a potential role in changing milk oligosaccharide profiles. For the candidate causative genes implicated in these analyses, we examined the strength of association between relevant loci and each wavenumber across the mid-infrared spectrum. This revealed shared association patterns for groups of genomically-distant loci, highlighting clusters of loci linked through their biological roles in lactation and their presumed impacts on the chemical composition of milk.

CONCLUSIONS

This study demonstrates the utility of FT-MIR wavenumber phenotypes for improving our understanding of milk composition, presenting a larger number of QTL and putative causative genes and variants than found from FT-MIR predicted composition traits. Examining patterns of significance across the mid-infrared spectrum for loci of interest further highlighted commonalities of association, which likely reflects the physico-chemical properties of milk constituents.

Animal medical genetics: a historical perspective on more than 50 years of research into genetic disorders of animals at Massey University

Over the last 50 years, there have been major advances in knowledge and technology regarding genetic diseases, and the subsequent ability to control them in a cost-effective manner. This review traces these advances through research into genetic diseases of animals at Massey University (Palmerston North, NZ), and briefly discusses the disorders investigated during that time, with additional detail for disorders of major importance such as bovine α-mannosidosis, ovine ceroid-lipofuscinosis, canine mucopolysaccharidosis IIIA and feline hyperchylomicronaemia. The overall research has made a significant contribution to veterinary medicine, has provided new biological knowledge and advanced our understanding of similar disorders in human patients, including testing various specific therapies prior to human clinical trials.

Non-additive association analysis using proxy phenotypes identifies novel cattle syndromes

Mammalian species carry ~100 loss-of-function variants per individual^1,2, where ~1-5 of these impact essential genes and cause embryonic lethality or severe disease when homozygous³. The functions of the remainder are more difficult to resolve, although the assumption is that these variants impact fitness in less manifest ways. Here we report one of the largest sequence-resolution screens of cattle to date, targeting discovery and validation of non-additive effects in 130,725 animals. We highlight six novel recessive loci with impacts generally exceeding the largest-effect variants identified from additive genome-wide association studies, presenting analogs of human diseases and hitherto-unrecognized disorders. These loci present compelling missense (PLCD4, MTRF1 and DPF2), premature stop (MUS81) and splice-disrupting (GALNT2 and FGD4) mutations, together explaining substantial proportions of inbreeding depression. These results demonstrate that the frequency distribution of deleterious alleles segregating in selected species can afford sufficient power to directly map novel disorders, presenting selection opportunities to minimize the incidence of genetic disease.

Mutations in the Kinesin-2 Motor KIF3B Cause an Autosomal-Dominant Ciliopathy

Kinesin-2 enables ciliary assembly and maintenance as an anterograde intraflagellar transport (IFT) motor. Molecular motor activity is driven by a heterotrimeric complex comprised of KIF3A and KIF3B or KIF3C plus one non-motor subunit, KIFAP3. Using exome sequencing, we identified heterozygous KIF3B variants in two unrelated families with hallmark ciliopathy phenotypes. In the first family, the proband presents with hepatic fibrosis, retinitis pigmentosa, and postaxial polydactyly; he harbors a de novo c.748G>C (p.Glu250Gln) variant affecting the kinesin motor domain encoded by KIF3B. The second family is a six-generation pedigree affected predominantly by retinitis pigmentosa. Affected individuals carry a heterozygous c.1568T>C (p.Leu523Pro) KIF3B variant segregating in an autosomal-dominant pattern. We observed a significant increase in primary cilia length in vitro in the context of either of the two mutations while variant KIF3B proteins retained stability indistinguishable from wild type. Furthermore, we tested the effects of KIF3B mutant mRNA expression in the developing zebrafish retina. In the presence of either missense variant, rhodopsin was sequestered to the photoreceptor rod inner segment layer with a concomitant increase in photoreceptor cilia length. Notably, impaired rhodopsin trafficking is also characteristic of recessive KIF3B models as exemplified by an early-onset, autosomal-recessive, progressive retinal degeneration in Bengal cats; we identified a c.1000G>A (p.Ala334Thr) KIF3B variant by genome-wide association study and whole-genome sequencing. Together, our genetic, cell-based, and in vivo modeling data delineate an autosomal-dominant syndromic retinal ciliopathy in humans and suggest that multiple KIF3B pathomechanisms can impair kinesin-driven ciliary transport in the photoreceptor.

The evolving role of Fourier-transform mid-infrared spectroscopy in genetic improvement of dairy cattle

Over the last 100 years, significant advances have been made in the characterisation of milk composition for dairy cattle improvement programs. Technological progress has enabled a shift from labour intensive, on-farm collection and processing of samples that assess yield and fat levels in milk, to large-scale processing of samples through centralised laboratories, with the scope extended to include quantification of other traits. Fourier-transform mid-infrared (FT-MIR) spectroscopy has had a significant role in the transformation of milk composition phenotyping, with spectral-based predictions of major milk components already being widely used in milk payment and animal evaluation systems globally. Increasingly, there is interest in analysing the individual FT-MIR wavenumbers, and in utilising the FT-MIR data to predict other novel traits of importance to breeding programs. This includes traits related to the nutritional value of milk, the processability of milk into products such as cheese, and traits relevant to animal health and the environment. The ability to successfully incorporate these traits into breeding programs is dependent on the heritability of the FT-MIR predicted traits, and the genetic correlations between the FT-MIR predicted and actual trait values. Linking FT-MIR predicted traits to the underlying mutations responsible for their variation can be difficult because the phenotypic expression of these traits are a function of a diverse range of molecular and biological mechanisms that can obscure their genetic basis. The individual FT-MIR wavenumbers give insights into the chemical composition of milk and provide an additional layer of granularity that may assist with establishing causal links between the genome and observed phenotypes. Additionally, there are other molecular phenotypes such as those related to the metabolome, chromatin accessibility, and RNA editing that could improve our understanding of the underlying biological systems controlling traits of interest. Here we review topics of importance to phenotyping and genetic applications of FT-MIR spectra datasets, and discuss opportunities for consolidating FT-MIR datasets with other genomic and molecular data sources to improve future dairy cattle breeding programs.

Accuracy of genomic prediction of shell quality in a White Leghorn line

Several genomic methods were applied for predicting shell quality traits recorded at 4 different hen ages in a White Leghorn line. The accuracies of genomic prediction of single-step GBLUP and single-trait Bayes B were compared with predictions of breeding values based on pedigree-BLUP under single-trait or multitrait models. Breaking strength (BS) and dynamic stiffness (Kdyn) measurements were collected on 18,524 birds from 3 consecutive generations, of which 4,164 animals also had genotypes from an Affymetrix 50K panel containing 49,591 SNPs after quality control edits. All traits had low to moderate heritability, ranging from 0.17 for BS to 0.34 for Kdyn. The highest accuracies of prediction were obtained for the multitrait single-step model. The use of marker information resulted in higher prediction accuracies than pedigree-based models for almost all traits. A genome-wide association study based on a Bayes B model was conducted to detect regions explaining the largest proportion of genetic variance. Across all 8 shell quality traits analyzed, 7 regions each explaining over 2% of genetic variance and 54 regions each explaining over 1% of genetic variance were identified. The windows explaining a large proportion of genetic variance overlapped with several potential candidate genes with biological functions linked to shell formation. A multitrait repeatability model using a single-step method is recommended for genomic evaluation of shell quality in layer chickens.

Sire selection and genetic improvement of dairy cattle assuming pure market competition

Breeding companies and farmers rely on selection indices to identify sires they expect to improve production system profitability. Such indices combine estimates of genetic merit for individual traits with corresponding economic values that are fixed based on circumstances at a particular time. Perfect market competition has also been proposed as an economic basis to value alternative sires. The objective of this study was to propose an economic model and develop it for the evaluation of dairy sires. The pure competition model (PUC) was used to evaluate the relationship between a profitability index based on the PUC approach versus the traditional selection index approach for 330 dairy sires comprising Holstein-Friesians, Jerseys, and Ayrshires. The correlation between these 2 selection indices was only 0.56, indicating that the conventional selection index did not correlate well with an index based on the PUC model. In particular, the higher ranking bulls were overvalued using the conventional selection index. Our study concluded that the use of fixed economic values is problematic for the delivery of consistent rankings in selection indices. In contrast, sire rankings based on PUC are more reliable because the sires are evaluated on the basis of efficiency gains rather than production while accounting for market prices and marginal values of dairy outputs over time.

Accuracies of direct genomic breeding values for birth and weaning weights of registered Charolais cattle in Mexico

Context Genomic prediction is now routinely used in many livestock species to rank individuals based on genomic breeding values (GEBV). Aims This study reports the first assessment aimed to evaluate the accuracy of direct GEBV for birth (BW) and weaning (WW) weights of registered Charolais cattle in Mexico. Methods The population assessed included 823 animals genotyped with an array of 77000 single nucleotide polymorphisms. Genomic prediction used genomic best linear unbiased prediction (GBLUP), Bayes C (BC), and single-step Bayesian regression (SSBR) methods in comparison with a pedigree-based BLUP method. Key results Our results show that the genomic prediction methods provided low and similar accuracies to BLUP. The prediction accuracy of GBLUP and BC were identical at 0.31 for BW and 0.29 for WW, similar to BLUP. Prediction accuracies of SSBR for BW and WW were up to 4% higher than those by BLUP. Conclusions Genomic prediction is feasible under current conditions, and provides a slight improvement using SSBR. Implications Some limitations on reference population size and structure were identified and need to be addressed to obtain more accurate predictions in liveweight traits under the prevalent cattle breeding conditions of Mexico.

Fine mapping of genomic regions associated with female fertility in Nellore beef cattle based on sequence variants from segregating sires

Background

Impaired fertility in cattle limits the efficiency of livestock production systems. Unraveling the genetic architecture of fertility traits would facilitate their improvement by selection. In this study, we characterized SNP chip haplotypes at QTL blocks then used whole-genome sequencing to fine map genomic regions associated with reproduction in a population of Nellore (Bos indicus) heifers.

Methods

The dataset comprised of 1337 heifers genotyped using a GeneSeek® Genomic Profiler panel (74677 SNPs), representing the daughters from 78 sires. After performing marker quality control, 64800 SNPs were retained. Haplotypes carried by each sire at six previously identified QTL on BTAs 5, 14 and 18 for heifer pregnancy and BTAs 8, 11 and 22 for antral follicle count were constructed using findhap software. The significance of the contrasts between the effects of every two paternally-inherited haplotype alleles were used to identify sires that were heterozygous at each QTL. Whole-genome sequencing data localized to the haplotypes from six sires and 20 other ancestors were used to identify sequence variants that were concordant with the haplotype contrasts. Enrichment analyses were applied to these variants using KEGG and MeSH libraries.

Results

A total of six (BTA 5), six (BTA 14) and five (BTA 18) sires were heterozygous for heifer pregnancy QTL whereas six (BTA 8), fourteen (BTA 11), and five (BTA 22) sires were heterozygous for number of antral follicles’ QTL. Due to inadequate representation of many haplotype alleles in the sequenced animals, fine mapping analysis could only be reliably performed for the QTL on BTA 5 and 14, which had 641 and 3733 concordant candidate sequence variants, respectively. The KEGG “Circadian rhythm” and “Neurotrophin signaling pathway” were significantly associated with the genes in the QTL on BTA 5 whereas 32 MeSH terms were associated with the QTL on BTA 14. Among the concordant sequence variants, 0.2% and 0.3% were classified as missense variants for BTAs 5 and 14, respectively, highlighting the genes MTERF2, RTMB, ENSBTAG00000037306 (miRNA), ENSBTAG00000040351, PRKDC, and RGS20. The potential causal mutations found in the present study were associated with biological processes such as oocyte maturation, embryo development, placenta development and response to reproductive hormones.

Conclusions

The identification of heterozygous sires by positionally phasing SNP chip data and contrasting haplotype effects for previously detected QTL can be used for fine mapping to identify potential causal mutations and candidate genes. Genomic variants on genes MTERF2, RTBC, miRNA ENSBTAG00000037306, ENSBTAG00000040351, PRKDC, and RGS20, which are known to have influence on reproductive biological processes, were detected.

Unraveling genomic associations with feed efficiency and body weight traits in chickens through an integrative approach

Background

Feed efficiency and growth rate have been targets for selection to improve chicken production. The incorporation of genomic tools may help to accelerate selection. We genotyped 529 individuals using a high-density SNP chip (600 K, Affymetrix®) to estimate genomic heritability of performance traits and to identify genomic regions and their positional candidate genes associated with performance traits in a Brazilian F₂ Chicken Resource population. Regions exhibiting selection signatures and a SNP dataset from resequencing were integrated with the genomic regions identified using the chip to refine the list of positional candidate genes and identify potential causative mutations.

Results

Feed intake (FI), feed conversion ratio (FC), feed efficiency (FE) and weight gain (WG) exhibited low genomic heritability values (i.e. from 0.0002 to 0.13), while body weight at hatch (BW1), 35 days-of-age (BW35), and 41 days-of-age (BW41) exhibited high genomic heritability values (i.e. from 0.60 to 0.73) in this F₂ population. Twenty unique 1-Mb genomic windows were associated with BW1, BW35 or BW41, located on GGA1–4, 6–7, 10, 14, 24, 27 and 28. Thirty-eight positional candidate genes were identified within these windows, and three of them overlapped with selection signature regions. Thirteen predicted deleterious and three high impact sequence SNPs in these QTL regions were annotated in 11 positional candidate genes related to osteogenesis, skeletal muscle development, growth, energy metabolism and lipid metabolism, which may be associated with body weight in chickens.

Conclusions

The use of a high-density SNP array to identify QTL which were integrated with whole genome sequence signatures of selection allowed the identification of candidate genes and candidate causal variants. One novel QTL was detected providing additional information to understand the genetic architecture of body weight traits. We identified QTL for body weight traits, which were also associated with fatness in the same population. Our findings form a basis for further functional studies to elucidate the role of specific genes in regulating body weight and fat deposition in chickens, generating useful information for poultry breeding programs.

Strategies for noise reduction and standardization of milk mid-infrared spectra from dairy cattle

The use of Fourier-transform mid-infrared (FTIR) spectroscopy is of interest to the dairy industry worldwide for predicting milk composition and other novel traits that are difficult or expensive to measure directly. Although there are many valuable applications for FTIR spectra, noise from differences in spectral responses between instruments is problematic because it reduces prediction accuracy if ignored. The purpose of this study was to develop strategies to reduce the impact of noise and to compare methods for standardizing FTIR spectra in order to reduce between-instrument variability in multiple-instrument networks. Noise levels in bands of the infrared spectrum caused by the water content of milk were characterized, and a method for identifying and removing outliers was developed. Two standardization methods were assessed and compared: piecewise direct standardization (PDS), which related spectra on a primary instrument to spectra on 5 other (secondary) instruments using identical milk-based reference samples (n = 918) analyzed across the 6 instruments; and retroactive percentile standardization (RPS), whereby percentiles of observed spectra from routine milk test samples (n = 2,044,094) were used to map and exploit primary- and secondary-instrument relationships. Different applications of each method were studied to determine the optimal way to implement each method across time. Industry-standard predictions of milk components from 2,044,094 spectra records were regressed against predictions from spectra before and after standardization using PDS or RPS. The PDS approach resulted in an overall decrease in root mean square error between industry-standard predictions and predictions from spectra from 0.190 to 0.071 g/100 mL for fat, from 0.129 to 0.055 g/100 mL for protein, and from 0.143 to 0.088 g/100 mL for lactose. Reductions in prediction error for RPS were similar but less consistent than those for PDS across time, but similar reductions were achieved when PDS coefficients were updated monthly and separate primary instruments were assigned for the North and South Islands of New Zealand. We demonstrated that the PDS approach is the most consistent method to reduce prediction errors across time. We also showed that the RPS approach is sensitive to shifts in milk composition but can be used to reduce prediction errors, provided that secondary-instrument spectra are standardized to a primary instrument with samples of broadly equivalent milk composition. Appropriate implementation of either of these approaches will improve the quality of predictions based on FTIR spectra for various downstream applications.

β-Mannosidosis in German Shepherd Dogs

A neurological disease was investigated in 3 German Shepherd pups from the same litter that failed to grow normally, appeared stiff, were reluctant to move, and were deaf. They developed intermittent seizures and ataxia and had proprioceptive defects. Histopathology showed severe vacuolation of neurons, astrocytes in nervous tissue, renal tubular epithelial cells, and macrophages in nervous tissue, spleen, and liver. Vacuoles appeared empty with no storage material stained by periodic acid-Schiff (PAS) or Sudan black stains, leading to a diagnosis of a lysosomal storage disease and in particular an oligosaccharidosis. Biochemical and genomic studies showed that this was β-mannosidosis, not previously diagnosed in dogs. A c.560T>A transition in exon 4 of the MANBA gene was found, which segregated in these and other family members in a manner consistent with it being the causative mutation of an autosomal recessive disease. This mutation led to substitution of isoleucine to asparagine at position 187 of the 885 amino acid enzyme, a change expected to have functional significance.

Identification of recombination hotspots and quantitative trait loci for recombination rate in layer chickens

Background

The frequency of recombination events varies across the genome and between individuals, which may be related to some genomic features. The objective of this study was to assess the frequency of recombination events and to identify QTL (quantitative trait loci) for recombination rate in two purebred layer chicken lines.

Methods

A total of 1200 white-egg layers (WL) were genotyped with 580 K SNPs and 5108 brown-egg layers (BL) were genotyped with 42 K SNPs (single nucleotide polymorphisms). Recombination events were identified within half-sib families and both the number of recombination events and the recombination rate was calculated within each 0.5 Mb window of the genome. The 10% of windows with the highest recombination rate on each chromosome were considered to be recombination hotspots. A BayesB model was used separately for each line to identify genomic regions associated with the genome-wide number of recombination event per meiosis. Regions that explained more than 0.8% of genetic variance of recombination rate were considered to harbor QTL.

Results

Heritability of recombination rate was estimated at 0.17 in WL and 0.16 in BL. On average, 11.3 and 23.2 recombination events were detected per individual across the genome in 1301 and 9292 meioses in the WL and BL, respectively. The estimated recombination rates differed significantly between the lines, which could be due to differences in inbreeding levels, and haplotype structures. Dams had about 5% to 20% higher recombination rates per meiosis than sires in both lines. Recombination rate per 0.5 Mb window had a strong negative correlation with chromosome size and a strong positive correlation with GC content and with CpG island density across the genome in both lines. Different QTL for recombination rate were identified in the two lines. There were 190 and 199 non-overlapping recombination hotspots detected in WL and BL respectively, 28 of which were common to both lines.

Conclusions

Differences in the recombination rates, hotspot locations, and QTL regions associated with genome-wide recombination were observed between lines, indicating the breed-specific feature of detected recombination events and the control of recombination events is a complex polygenic trait.

Electronic supplementary material

The online version of this article (10.1186/s40104-019-0332-y) contains supplementary material, which is available to authorized users.

A certain invariance property of BLUE in a whole-genome regression context

A curious result from mixed linear models applied to genome-wide association studies was expanded. In particular, a model in which one or more markers are considered as fixed but are allowed to contribute to the covariance structure by treating such markers as random as well was examined. The best linear unbiased estimator of marker effects is invariant with respect to whether those markers are employed in constructing a genomic relationship matrix or are ignored, provided marker effects are uncorrelated with those not being tested. Also, the implications of regarding some marker effects as fixed when, in fact, these possess a non-trivial covariance structure with those declared as random were examined.

Variance component estimates for post-thaw sperm variables measured by computer assisted semen analyzer for inbred and non-inbred Hereford bulls

Inbreeding has been shown to have negative effects on animal reproductive performance. It, however, has not been documented as to the extent that inbreeding impacts sperm cell characteristics especially after cells have undergone cryopreservation. In the present study, therefore, there was evaluation of sperm motion and size characteristics from two inbred lines and non-inbred Hereford bulls using Computer Assisted Sperm Analysis (CASA). Pre-frozen semen samples were obtained from Line 1, Prospector, and a non-inbred Hereford bulls. Average inbreeding of each group was 27%, 45%, and 0%, for Line 1, Prospector, and non-inbred groups, respectively. For each bull, two straws (0.5 ml) were evaluated, within each straw two slides with ten fields per slide analyzed with the CASA. Cells were analyzed by mixed model analysis using a model that contained line as fixed effect and individual bull, straw, slide, and fields are random effects. No significant differences between lines of bulls were found for motility or progressive motility. For the head elongation (ELON), and head area (AREA) significant differences existed between lines. Mean separation indicated that Line 1 group had a larger head area and the heads were not as elongated as the Prospector and non-inbred group cells. The lack of significant differences for motility and progressive motility across lines was surprising and indicates there was no inbreeding depression for post-thaw semen characteristics. One explanation for this could be that indirect selection pressure was placed on these characteristics as inbreeding was increased.

Integration of genome wide association studies and whole genome sequencing provides novel insights into fat deposition in chicken

Excessive fat deposition is a negative factor for poultry production because it reduces feed efficiency, increases the cost of meat production and is a health concern for consumers. We genotyped 497 birds from a Brazilian F₂ Chicken Resource Population, using a high-density SNP array (600 K), to estimate the genomic heritability of fat deposition related traits and to identify genomic regions and positional candidate genes (PCGs) associated with these traits. Selection signature regions, haplotype blocks and SNP data from a previous whole genome sequencing study in the founders of this chicken F2 population were used to refine the list of PCGs and to identify potential causative SNPs. We obtained high genomic heritabilities (0.43-0.56) and identified 22 unique QTLs for abdominal fat and carcass fat content traits. These QTLs harbored 26 PCGs involved in biological processes such as fat cell differentiation, insulin and triglyceride levels, and lipid biosynthetic process. Three of these 26 PCGs were located within haplotype blocks there were associated with fat traits, five overlapped with selection signature regions, and 12 contained predicted deleterious variants. The identified QTLs, PCGs and potentially causative SNPs provide new insights into the genetic control of fat deposition and can lead to improved accuracy of selection to reduce excessive fat deposition in chickens.

Meta-analysis of genome-wide association studies for cattle stature identifies common genes that regulate body size in mammals

Stature is affected by many polymorphisms of small effect in humans ¹ . In contrast, variation in dogs, even within breeds, has been suggested to be largely due to variants in a small number of genes^2,3. Here we use data from cattle to compare the genetic architecture of stature to those in humans and dogs. We conducted a meta-analysis for stature using 58,265 cattle from 17 populations with 25.4 million imputed whole-genome sequence variants. Results showed that the genetic architecture of stature in cattle is similar to that in humans, as the lead variants in 163 significantly associated genomic regions (P < 5 × 10^-8) explained at most 13.8% of the phenotypic variance. Most of these variants were noncoding, including variants that were also expression quantitative trait loci (eQTLs) and in ChIP-seq peaks. There was significant overlap in loci for stature with humans and dogs, suggesting that a set of common genes regulates body size in mammals.

Network Analysis Reveals Putative Genes Affecting Meat Quality in Angus Cattle

Improvements in eating satisfaction will benefit consumers and should increase beef demand which is of interest to the beef industry. Tenderness, juiciness, and flavor are major determinants of the palatability of beef and are often used to reflect eating satisfaction. Carcass qualities are used as indicator traits for meat quality, with higher quality grade carcasses expected to relate to more tender and palatable meat. However, meat quality is a complex concept determined by many component traits making interpretation of genome-wide association studies (GWAS) on any one component challenging to interpret. Recent approaches combining traditional GWAS with gene network interactions theory could be more efficient in dissecting the genetic architecture of complex traits. Phenotypic measures of 23 traits reflecting carcass characteristics, components of meat quality, along with mineral and peptide concentrations were used along with Illumina 54k bovine SNP genotypes to derive an annotated gene network associated with meat quality in 2,110 Angus beef cattle. The efficient mixed model association (EMMAX) approach in combination with a genomic relationship matrix was used to directly estimate the associations between 54k SNP genotypes and each of the 23 component traits. Genomic correlated regions were identified by partial correlations which were further used along with an information theory algorithm to derive gene network clusters. Correlated SNP across 23 component traits were subjected to network scoring and visualization software to identify significant SNP. Significant pathways implicated in the meat quality complex through GO term enrichment analysis included angiogenesis, inflammation, transmembrane transporter activity, and receptor activity. These results suggest that network analysis using partial correlations and annotation of significant SNP can reveal the genetic architecture of complex traits and provide novel information regarding biological mechanisms and genes that lead to complex phenotypes, like meat quality, and the nutritional and healthfulness value of beef. Improvements in genome annotation and knowledge of gene function will contribute to more comprehensive analyses that will advance our ability to dissect the complex architecture of complex traits.

Genotype imputation in a tropical crossbred dairy cattle population

The objective of this study was to investigate different strategies for genotype imputation in a population of crossbred Girolando (Gyr × Holstein) dairy cattle. The data set consisted of 478 Girolando, 583 Gyr, and 1,198 Holstein sires genotyped at high density with the Illumina BovineHD (Illumina, San Diego, CA) panel, which includes ∼777K markers. The accuracy of imputation from low (20K) and medium densities (50K and 70K) to the HD panel density and from low to 50K density were investigated. Seven scenarios using different reference populations (RPop) considering Girolando, Gyr, and Holstein breeds separately or combinations of animals of these breeds were tested for imputing genotypes of 166 randomly chosen Girolando animals. The population genotype imputation were performed using FImpute. Imputation accuracy was measured as the correlation between observed and imputed genotypes (CORR) and also as the proportion of genotypes that were imputed correctly (CR). This is the first paper on imputation accuracy in a Girolando population. The sample-specific imputation accuracies ranged from 0.38 to 0.97 (CORR) and from 0.49 to 0.96 (CR) imputing from low and medium densities to HD, and 0.41 to 0.95 (CORR) and from 0.50 to 0.94 (CR) for imputation from 20K to 50K. The CORR_anim exceeded 0.96 (for 50K and 70K panels) when only Girolando animals were included in RPop (S1). We found smaller CORR_anim when Gyr (S2) was used instead of Holstein (S3) as RPop. The same behavior was observed between S4 (Gyr + Girolando) and S5 (Holstein + Girolando) because the target animals were more related to the Holstein population than to the Gyr population. The highest imputation accuracies were observed for scenarios including Girolando animals in the reference population, whereas using only Gyr animals resulted in low imputation accuracies, suggesting that the haplotypes segregating in the Girolando population had a greater effect on accuracy than the purebred haplotypes. All chromosomes had similar imputation accuracies (CORR_snp) within each scenario. Crossbred animals (Girolando) must be included in the reference population to provide the best imputation accuracies.

Efficient strategies for leave-one-out cross validation for genomic best linear unbiased prediction

Background

A random multiple-regression model that simultaneously fit all allele substitution effects for additive markers or haplotypes as uncorrelated random effects was proposed for Best Linear Unbiased Prediction, using whole-genome data. Leave-one-out cross validation can be used to quantify the predictive ability of a statistical model.

Methods

Naive application of Leave-one-out cross validation is computationally intensive because the training and validation analyses need to be repeated n times, once for each observation. Efficient Leave-one-out cross validation strategies are presented here, requiring little more effort than a single analysis.

Results

Efficient Leave-one-out cross validation strategies is 786 times faster than the naive application for a simulated dataset with 1,000 observations and 10,000 markers and 99 times faster with 1,000 observations and 100 markers. These efficiencies relative to the naive approach using the same model will increase with increases in the number of observations.

Conclusions

Efficient Leave-one-out cross validation strategies are presented here, requiring little more effort than a single analysis.

Single nucleotide variants and InDels identified from whole-genome re-sequencing of Guzerat, Gyr, Girolando and Holstein cattle breeds

Whole-genome re-sequencing, alignment and annotation analyses were undertaken for 12 sires representing four important cattle breeds in Brazil: Guzerat (multi-purpose), Gyr, Girolando and Holstein (dairy production). A total of approximately 4.3 billion reads from an Illumina HiSeq 2000 sequencer generated for each animal 10.7 to 16.4-fold genome coverage. A total of 27,441,279 single nucleotide variations (SNVs) and 3,828,041 insertions/deletions (InDels) were detected in the samples, of which 2,557,670 SNVs and 883,219 InDels were novel. The submission of these genetic variants to the dbSNP database significantly increased the number of known variants, particularly for the indicine genome. The concordance rate between genotypes obtained using the Bovine HD BeadChip array and the same variants identified by sequencing was about 99.05%. The annotation of variants identified numerous non-synonymous SNVs and frameshift InDels which could affect phenotypic variation. Functional enrichment analysis was performed and revealed that variants in the olfactory transduction pathway was over represented in all four cattle breeds, while the ECM-receptor interaction pathway was over represented in Girolando and Guzerat breeds, the ABC transporters pathway was over represented only in Holstein breed, and the metabolic pathways was over represented only in Gyr breed. The genetic variants discovered here provide a rich resource to help identify potential genomic markers and their associated molecular mechanisms that impact economically important traits for Gyr, Girolando, Guzerat and Holstein breeding programs.

Computational strategies for alternative single-step Bayesian regression models with large numbers of genotyped and non-genotyped animals

Background

Two types of models have been used for single-step genomic prediction and genome-wide association studies that include phenotypes from both genotyped animals and their non-genotyped relatives. The two types are breeding value models (BVM) that fit breeding values explicitly and marker effects models (MEM) that express the breeding values in terms of the effects of observed or imputed genotypes. MEM can accommodate a wider class of analyses, including variable selection or mixture model analyses. The order of the equations that need to be solved and the inverses required in their construction vary widely, and thus the computational effort required depends upon the size of the pedigree, the number of genotyped animals and the number of loci.

Theory

We present computational strategies to avoid storing large, dense blocks of the MME that involve imputed genotypes. Furthermore, we present a hybrid model that fits a MEM for animals with observed genotypes and a BVM for those without genotypes. The hybrid model is computationally attractive for pedigree files containing millions of animals with a large proportion of those being genotyped.

Application

We demonstrate the practicality on both the original MEM and the hybrid model using real data with 6,179,960 animals in the pedigree with 4,934,101 phenotypes and 31,453 animals genotyped at 40,214 informative loci. To complete a single-trait analysis on a desk-top computer with four graphics cards required about 3 h using the hybrid model to obtain both preconditioned conjugate gradient solutions and 42,000 Markov chain Monte-Carlo (MCMC) samples of breeding values, which allowed making inferences from posterior means, variances and covariances. The MCMC sampling required one quarter of the effort when the hybrid model was used compared to the published MEM.

Conclusions

We present a hybrid model that fits a MEM for animals with genotypes and a BVM for those without genotypes. Its practicality and considerable reduction in computing effort was demonstrated. This model can readily be extended to accommodate multiple traits, multiple breeds, maternal effects, and additional random effects such as polygenic residual effects.

Epistatic interactions associated with fatty acid concentrations of beef from angus sired beef cattle

BACKGROUND

Consumers are becoming increasingly conscientious about the nutritional value of their food. Consumption of some fatty acids has been associated with human health traits such as blood pressure and cardiovascular disease. Therefore, it is important to investigate genetic variation in content of fatty acids present in meat. Previously publications reported regions of the cattle genome that are additively associated with variation in fatty acid content. This study evaluated epistatic interactions, which could account for additional genetic variation in fatty acid content.

RESULTS

Epistatic interactions for 44 fatty acid traits in a population of Angus beef cattle were evaluated with EpiSNPmpi. False discovery rate (FDR) was controlled at 5 % and was limited to well-represented genotypic combinations. Epistatic interactions were detected for 37 triacylglyceride (TAG), 36 phospholipid (PL) fatty acid traits, and three weight traits. A total of 6,181, 7,168, and 0 significant epistatic interactions (FDR < 0.05, 50-animals per genotype combination) were associated with Triacylglyceride fatty acids, Phospholipid fatty acids, and weight traits respectively and most were additive-by-additive interactions. A large number of interactions occurred in potential regions of regulatory control along the chromosomes where genes related to fatty acid metabolism reside.

CONCLUSIONS

Many fatty acids were associated with epistatic interactions. Despite a large number of significant interactions, there are a limited number of genomic locations that harbored these interactions. While larger population sizes are needed to accurately validate and quantify these epistatic interactions, the current findings point towards additional genetic variance that can be accounted for within these fatty acid traits.

An efficient exact method to obtain GBLUP and single-step GBLUP when the genomic relationship matrix is singular

Background

The mixed linear model employed for genomic best linear unbiased prediction (GBLUP) includes the breeding value for each animal as a random effect that has a mean of zero and a covariance matrix proportional to the genomic relationship matrix (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg), where the inverse of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg is required to set up the usual mixed model equations (MME). When only some animals have genomic information, genomic predictions can be obtained by an extension known as single-step GBLUP, where the covariance matrix of breeding values is constructed by combining the pedigree-based additive relationship matrix with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg. The inverse of the combined relationship matrix can be obtained efficiently, provided \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg can be inverted. In some livestock species, however, the number \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$N_{g}$$\end{document}Ng of animals with genomic information exceeds the number of marker covariates used to compute \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg, and this results in a singular \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg. For such a case, an efficient and exact method to obtain GBLUP and single-step GBLUP is presented here.

Results

Exact methods are already available to obtain GBLUP when \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg is singular, but these require working with large dense matrices. Another approach is to modify \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg to make it nonsingular by adding a small value to all its diagonals or regressing it towards the pedigree-based relationship matrix. This, however, results in the inverse of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg being dense and difficult to compute as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$N_{g}$$\end{document}Ng grows. The approach presented here recognizes that the number r of linearly independent genomic breeding values cannot exceed the number of marker covariates, and the mixed linear model used here for genomic prediction only fits these r linearly independent breeding values as random effects.

Conclusions

The exact method presented here was compared to Apy-GBLUP and to Apy single-step GBLUP, both of which are approximate methods that use a modified \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mathbf {G}}_{gg}$$\end{document}Ggg that has a sparse inverse which can be computed efficiently. In a small numerical example, predictions from the exact approach and Apy were almost identical, but the MME from Apy had a condition number about 1000 times larger than that from the exact approach, indicating ill-conditioning of the MME from Apy. The practical application of exact SSGBLUP is not more difficult than implementation of Apy.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-016-0260-7) contains supplementary material, which is available to authorized users.

An Upper Bound for Accuracy of Prediction Using GBLUP

This study aims at characterizing the asymptotic behavior of genomic prediction R² as the size of the reference population increases for common or rare QTL alleles through simulations. Haplotypes derived from whole-genome sequence of 85 Caucasian individuals from the 1,000 Genomes Project were used to simulate random mating in a population of 10,000 individuals for at least 100 generations to create the LD structure in humans for a large number of individuals. To reduce computational demands, only SNPs within a 0.1M region of each of the first 5 chromosomes were used in simulations, and therefore, the total genome length simulated was 0.5M. When the genome length is 30M, to get the same genomic prediction R² as with a 0.5M genome would require a reference population 60 fold larger. Three scenarios were considered varying in minor allele frequency distributions of markers and QTL, for h² = 0.8 resembling height in humans. Total number of markers was 4,200 and QTL were 70 for each scenario. In this study, we considered the prediction accuracy in terms of an estimability problem, and thereby provided an upper bound for reliability of prediction, and thus, for prediction R². Genomic prediction methods GBLUP, BayesB and BayesC were compared. Our results imply that for human height variable selection methods BayesB and BayesC applied to a 30M genome have no advantage over GBLUP when the size of reference population was small (<6,000 individuals), but are superior as more individuals are included in the reference population. All methods become asymptotically equivalent in terms of prediction R², which approaches genomic heritability when the size of the reference population reaches 480,000 individuals.

Deriving Gene Networks from SNP Associated with Triacylglycerol and Phospholipid Fatty Acid Fractions from Ribeyes of Angus Cattle

The fatty acid profile of beef is a complex trait that can benefit from gene-interaction network analysis to understand relationships among loci that contribute to phenotypic variation. Phenotypic measures of fatty acid profile from triacylglycerol and phospholipid fractions of longissimus muscle, pedigree information, and Illumina 54 k bovine SNP genotypes were utilized to derive an annotated gene network associated with fatty acid composition in 1,833 Angus beef cattle. The Bayes-B statistical model was utilized to perform a genome wide association study to estimate associations between 54 k SNP genotypes and 39 individual fatty acid phenotypes within each fraction. Posterior means of the effects were estimated for each of the 54 k SNP and for the collective effects of all the SNP in every 1-Mb genomic window in terms of the proportion of genetic variance explained by the window. Windows that explained the largest proportions of genetic variance for individual lipids were found in the triacylglycerol fraction. There was almost no overlap in the genomic regions explaining variance between the triacylglycerol and phospholipid fractions. Partial correlations were used to identify correlated regions of the genome for the set of largest 1 Mb windows that explained up to 35% genetic variation in either fatty acid fraction. SNP were allocated to windows based on the bovine UMD3.1 assembly. Gene network clusters were generated utilizing a partial correlation and information theory algorithm. Results were used in conjunction with network scoring and visualization software to analyze correlated SNP across 39 fatty acid phenotypes to identify SNP of significance. Significant pathways implicated in fatty acid metabolism through GO term enrichment analysis included homeostasis of number of cells, homeostatic process, coenzyme/cofactor activity, and immunoglobulin. These results suggest different metabolic pathways regulate the development of different types of lipids found in bovine muscle tissues. Network analysis using partial correlations and annotation of significant SNPs can yield information about the genetic architecture of complex traits.

Mixture models detect large effect QTL better than GBLUP and result in more accurate and persistent predictions

BACKGROUND

Accurate evaluation of SNP effects is important for genome wide association studies and for genomic prediction. The genetic architecture of quantitative traits differs widely, with some traits exhibiting few if any quantitative trait loci (QTL) with large effects, while other traits have one or several easily detectable QTL with large effects.

METHODS

Body weight in broilers and egg weight in layers are two examples of traits that have QTL of large effect. A commonly used method for genome wide association studies is to fit a mixture model such as BayesB that assumes some known proportion of SNP effects are zero. In contrast, the most commonly used method for genomic prediction is known as GBLUP, which involves fitting an animal model to phenotypic data with the variance-covariance or genomic relationship matrix among the animals being determined by genome wide SNP genotypes. Genotypes at each SNP are typically weighted equally in determining the genomic relationship matrix for GBLUP. We used the equivalent marker effects model formulation of GBLUP for this study. We compare these two classes of models using egg weight data collected over 8 generations from 2,324 animals genotyped with a 42 K SNP panel.

RESULTS

Using data from the first 7 generations, both BayesB and GBLUP found the largest QTL in a similar well-recognized QTL region, but this QTL was estimated to account for 24 % of genetic variation with BayesB and less than 1 % with GBLUP. When predicting phenotypes in generation 8 BayesB accounted for 36 % of the phenotypic variation and GBLUP for 25 %. When using only data from any one generation, the same QTL was identified with BayesB in all but one generation but never with GBLUP. Predictions of phenotypes in generations 2 to 7 based on only 295 animals from generation 1 accounted for 10 % phenotypic variation with BayesB but only 6 % with GBLUP. Predicting phenotype using only the marker effects in the 1 Mb region that accounted for the largest effect on egg weight from generation 1 data alone accounted for almost 8 % variation using BayesB but had no predictive power with GBLUP.

CONCLUSIONS

In conclusion, In the presence of large effect QTL, BayesB did a better job of QTL detection and its genomic predictions were more accurate and persistent than those from GBLUP.

Genetic parameters and genetic correlations among triacylglycerol and phospholipid fractions in Angus cattle

The objective of this study was to estimate genetic parameters for intramuscular fatty acids from triacylglycerol (TAG) and phospholipid (PL) fractions in beef LM tissue. Longissimus muscle samples were obtained from 1,833 Angus cattle to determine the intramuscular fatty acid composition for 31 lipids and lipid classes from TAG and PL fractions and were classified by structure into saturated (SFA), monounsaturated (MUFA), polyunsaturated (PUFA), omega-3 (n-3), and omega-6 (n-6) fatty acids. An atherogenic index (AI) was also determined as a measure of the unsaturated fatty acid to SFA ratio. Restricted maximum likelihood methods combined with pedigree data were used to estimate variance components with the WOMBAT software package. Heritability estimates ranged from 0.00 to 0.63 for the major classes of fatty acids. Heritability estimates differed between the TAG and PL fractions, with higher estimates for TAG up to 0.64 and lower estimates for PL that ranged from 0.00 to 0.14. Phenotypic and genetic correlations among individual fatty acids were determined for the TAG fraction as well as among carcass traits, including rib eye area, numerical marbling score, yield grade, ether fat, and Warner-Bratzler shear force value. Strong negative or positive genetic correlations were observed among individual fatty acids in the TAG fraction, which ranged from -0.99 to 0.97 ( < 0.05). Moderate correlations between carcass traits and fatty acids from the TAG fraction ranged from -0.43 to 0.32 ( < 0.05). These results indicate that fatty acids prominent in beef tissues show significant genetic variation as well as genetic relationships with carcass traits.

Accuracy of prediction of simulated polygenic phenotypes and their underlying quantitative trait loci genotypes using real or imputed whole-genome markers in cattle

BACKGROUND

More accurate genomic predictions are expected when the effects of QTL (quantitative trait loci) are predicted from markers in close physical proximity to the QTL. The objective of this study was to quantify to what extent whole-genome methods using 50 K or imputed 770 K SNPs (single nucleotide polymorphisms) could predict single or multiple QTL genotypes based on SNPs in close proximity to those QTL.

METHODS

Phenotypes with a heritability of 1 were simulated for 2677 Hereford animals genotyped with the BovineSNP50 BeadChip. Genotypes for the high-density 770 K SNP panel were imputed using Beagle software. Various Bayesian regression methods were used to predict single QTL or a trait influenced by 42 such QTL. We quantified to what extent these predictions were based on SNPs in close proximity to the QTL by comparing whole-genome predictions to local predictions based on estimates of the effects of variable numbers of SNPs i.e. ±1, ±2, ±5, ±10, ±50 or ±100 that flanked the QTL.

RESULTS

Prediction accuracies based on local SNPs using whole-genome training for single QTL with the 50 K SNP panel and BayesC0 ranged from 0.49 (±1 SNP) to 0.75 (±100 SNPs). The minimum number of local SNPs for an accurate prediction is ±10 SNPs. Prediction accuracies that were based on local SNPs only were higher than those based on whole-genome SNPs for both 50 K and 770 K SNP panels. For the 770 K SNP panel, prediction accuracies were higher than 0.70 and varied little i.e. between 0.73 (±1 SNP) and 0.77 (±5 SNPs). For the summed 42 QTL, prediction accuracies were generally higher than for single QTL regardless of the number of local SNPs. For QTL with low minor allele frequency (MAF) compared to QTL with high MAF, prediction accuracies increased as the number of SNPs around the QTL increased.

CONCLUSIONS

These results suggest that with both 50 K and imputed 770 K SNP genotypes the level of linkage disequilibrium is sufficient to predict single and multiple QTL. However, prediction accuracies are eroded through spuriously estimated effects of SNPs that are distant from the QTL. Prediction accuracies were higher with the 770 K than with the 50 K SNP panel.

A fast and efficient Gibbs sampler for BayesB in whole-genome analyses

Background

In whole-genome analyses, the number p of marker covariates is often much larger than the number n of observations. Bayesian multiple regression models are widely used in genomic selection to address this problem of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$p\gg n.$$\end{document}p≫n. The primary difference between these models is the prior assumed for the effects of the covariates. Usually in the BayesB method, a Metropolis–Hastings (MH) algorithm is used to jointly sample the marker effect and the locus-specific variance, which may make BayesB computationally intensive. In this paper, we show how the Gibbs sampler without the MH algorithm can be used for the BayesB method.

Results

We consider three different versions of the Gibbs sampler to sample the marker effect and locus-specific variance for each locus. Among the Gibbs samplers that were considered, the most efficient sampler is about 2.1 times as efficient as the MH algorithm proposed by Meuwissen et al. and 1.7 times as efficient as that proposed by Habier et al.

Conclusions

The three Gibbs samplers presented here were twice as efficient as Metropolis–Hastings samplers and gave virtually the same results.

Response and inbreeding from a genomic selection experiment in layer chickens

Background

Genomic selection (GS) using estimated breeding values (GS-EBV) based on dense marker data is a promising approach for genetic improvement. A simulation study was undertaken to illustrate the opportunities offered by GS for designing breeding programs. It consisted of a selection program for a sex-limited trait in layer chickens, which was developed by deterministic predictions under different scenarios. Later, one of the possible schemes was implemented in a real population of layer chicken.

Methods

In the simulation, the aim was to double the response to selection per year by reducing the generation interval by 50 %, while maintaining the same rate of inbreeding per year. We found that GS with retraining could achieve the set objectives while requiring 75 % fewer reared birds and 82 % fewer phenotyped birds per year. A multi-trait GS scenario was subsequently implemented in a real population of brown egg laying hens. The population was split into two sub-lines, one was submitted to conventional phenotypic selection, and one was selected based on genomic prediction. At the end of the 3-year experiment, the two sub-lines were compared for multiple performance traits that are relevant for commercial egg production.

Results

Birds that were selected based on genomic prediction outperformed those that were submitted to conventional selection for most of the 16 traits that were included in the index used for selection. However, although the two programs were designed to achieve the same rate of inbreeding per year, the realized inbreeding per year assessed from pedigree was higher in the genomic selected line than in the conventionally selected line.

Conclusions

The results demonstrate that GS is a promising alternative to conventional breeding for genetic improvement of layer chickens.