Genome-wide association and expression quantitative trait loci in cattle reveals common genes regulating mammalian fertility

Most genetic variants associated with fertility in mammals fall in non-coding regions of the genome and it is unclear how these variants affect fertility. Here we use genome-wide association summary statistics for Heifer puberty (pubertal or not at 600 days) from 27,707 Bos indicus, Bos taurus and crossbred cattle; multi-trait GWAS signals from 2119 indicine cattle for four fertility traits, including days to calving, age at first calving, pregnancy status, and foetus age in weeks (assessed by rectal palpation of the foetus); and expression quantitative trait locus for whole blood from 489 indicine cattle, to identify 87 putatively functional genes affecting cattle fertility. Our analysis reveals a significant overlap between the set of cattle and previously reported human fertility-related genes, impling the existence of a shared pool of genes that regulate fertility in mammals. These findings are crucial for developing approaches to improve fertility in cattle and potentially other mammals.

Concurrently mapping quantitative trait loci associations from multiple subspecies within hybrid populations

Many of the world's agriculturally important plant and animal populations consist of hybrids of subspecies. Cattle in tropical and sub-tropical regions for example, originate from two subspecies, Bos taurus indicus (Bos indicus) and Bos taurus taurus (Bos taurus). Methods to derive the underlying genetic architecture for these two subspecies are essential to develop accurate genomic predictions in these hybrid populations. We propose a novel method to achieve this. First, we use haplotypes to assign SNP alleles to ancestral subspecies of origin in a multi-breed and multi-subspecies population. Then we use a BayesR framework to allow SNP alleles originating from the different subspecies differing effects. Applying this method in a composite population of B. indicus and B. taurus hybrids, our results show that there are underlying genomic differences between the two subspecies, and these effects are not identified in multi-breed genomic evaluations that do not account for subspecies of origin effects. The method slightly improved the accuracy of genomic prediction. More significantly, by allocating SNP alleles to ancestral subspecies of origin, we were able to identify four SNP with high posterior probabilities of inclusion that have not been previously associated with cattle fertility and were close to genes associated with fertility in other species. These results show that haplotypes can be used to trace subspecies of origin through the genome of this hybrid population and, in conjunction with our novel Bayesian analysis, subspecies SNP allele allocation can be used to increase the accuracy of QTL association mapping in genetically diverse populations.

Multi-breed genomic evaluation for tropical beef cattle when no pedigree information is available

BACKGROUND

It has been challenging to implement genomic selection in multi-breed tropical beef cattle populations. If commercial (often crossbred) animals could be used in the reference population for these genomic evaluations, this could allow for very large reference populations. In tropical beef systems, such animals often have no pedigree information. Here we investigate potential models for such data, using marker heterozygosity (to model heterosis) and breed composition derived from genetic markers, as covariates in the model. Models treated breed effects as either fixed or random, and included genomic best linear unbiased prediction (GBLUP) and BayesR. A tropically-adapted beef cattle dataset of 29,391 purebred, crossbred and composite commercial animals was used to evaluate the models.

RESULTS

Treating breed effects as random, in an approach analogous to genetic groups allowed partitioning of the genetic variance into within-breed and across breed-components (even with a large number of breeds), and estimation of within-breed and across-breed genomic estimated breeding values (GEBV). We demonstrate that moderately-accurate (0.30-0.43) GEBV can be calculated using these models. Treating breed effects as random gave more accurate GEBV than treating breed as fixed. A simple GBLUP model where no breed effects were fitted gave the same accuracy (and correlations of GEBV very close to 1) as a model where GEBV for within-breed and the GEBV for (random) across-breed effects were included. When GEBV were predicted for herds with no data in the reference population, BayesR resulted in the highest accuracy, with 3% accuracy improvement averaged across traits, especially when the validation population was less related to the reference population. Estimates of heterosis from our models were in line with previous estimates from beef cattle. A method for estimating the number of effective breed comparisons for each breed combination accumulated across contemporary groups is presented.

CONCLUSIONS

When no pedigree is available, breed composition and heterosis for inclusion in multi-breed genomic evaluation can be estimated from genotypes. When GEBV were predicted for herds with no data in the reference population, BayesR resulted in the highest accuracy.

<i>Corrigendum to</i>: Production and performance of commercial beef breeding females in northern Australia. 4. Factors influencing the occurrence of lactating cows becoming pregnant within 4 months of calving

<sec> Context Sound reproductive efficiency is a key determinant for the overall productivity of a beef breeding business. For beef breeding herds to obtain high levels of reproductive productivity, breeding females need to efficiently become pregnant while lactating. </sec> <sec> Aims This study aimed to determine and quantify the major factors associated with lactating cows becoming pregnant within 4months of calving (P4M) in commercial beef breeding herds of northern Australia. </sec> <sec> Methods A prospective epidemiological study was conducted using 78 commercial northern Australian beef breeding herds and involved 78&#x2009;000 cattle that were monitored for 3&#x2013;4years. A multivariable model-building process was employed to scrutinise the resulting dataset to identify what herd-management practices, and nutritional, environmental and individual cow factors were major determinants of lactating cows becoming pregnant within 4months of calving (P4M) and to estimate their magnitudes of effect. </sec> <sec> Key results Overall, 41.6% of cows per production year were successful for P4M. Country type was strongly associated with 65.4%, 57.5%, 61.8% and 16.4% P4M for the Southern Forest, Central Forest, Northern Downs, and Northern Forest respectively. Between-year variability ranged between 3.3 and 11.7 percentage points. Cows calving in December&#x2013;January (61%) had a substantially higher occurrence of P4M than did cows calving between July and September (15%). The difference in P4M when comparing availability of wet-season pasture protein and phosphorus was 12.7 and 20.3 percentage points respectively. Modelling of the impact of group seroprevalence and management group prevalence of recent infection with several infectious diseases was estimated, with a large negative association between group bovine viral diarrhoea seroprevalence and P4M suggested. </sec> <sec> Conclusions This study further demonstrated the substantial impact that environment, herd management practices, nutrition and disease factors can have on the reproductive performance of females. </sec> <sec> Implications To optimise the performance of females (through increasing the occurrence of cows contributing calves in consecutive years) under commercial conditions in northern Australia, herd managers should focus on maximising the proportion of cows within a herd calving at the desired time of the year, ensuring that any nutritional deficiencies and herd health issues are managed, and that cows are managed such that they are of good body condition score at the time of calving. </sec>

High frequency of delayed milk delivery to neonates in tropical beef herds

Beef-calf mortality rates across tropical and subtropical Australia are high, with sub-optimal nutrition in pregnant cows being the primary risk. The nutritional deficiencies associated with calf mortality are the same as those associated with reduced milk yields. Although the highest mortality risk occurs during neonatal life, the role of inadequate milk delivery to beef neonates is not well established. This study investigated the frequency of low milk delivery in tropically adapted neonatal calves and the time for their dams to initiate full lactation in five management groups of Brahman and Droughtmaster calving cows in the dry tropics of northern Queensland, Australia. Change in calf weight in the days following birth was the primary measure of milk uptake. Plasma globulin concentration was used to indicate colostrum uptake. Across management groups, data were available on 250 calves for regression analysis of average daily gain vs. globulin and on 78 for plotting calf growth profiles. Calves had one of two growth profiles, either with immediate high growth from birth (day one) or with high growth delayed until day three. The frequency of delayed growth calves (with inadequate milk intake to gain at least 0.5 kg by day three after birth) was on average 30% across management groups, with management groups ranging 25%-50%. The frequency of calves growing ≤0.2 kg/day to day three was 15%-37%, depending on management group. The frequency of calves growing ≤0.2 kg/day to day five was 7%-20%, depending on management group. Calf globulin explained only 25% of the variation in calf average daily gain. Our study shows that a third of tropically adapted calves may experience a three-day delay to initiation of full lactation by their dams. Although study conditions were relatively benign, any additional risks with milk delivery, such as those that occur widely in tropical and subtropical northern Australia, would place such calves at risk of dehydration and mortality. Calf plasma globulin should not be used as a standalone measure of adequacy of neonatal milk delivery, especially when comparing across herds. This study demonstrates a fundamental problem of high frequency in northern Australia. The underlying risks for delayed milk delivery should be considered in the quest for practical solutions to reduce tropically adapted beef-calf mortalities.

Adaptive sampling during sequencing reveals the origins of the bovine reproductive tract microbiome across reproductive stages and sexes

Cattle enterprises are one of the major livestock production systems globally and are forecasted to have stable growth in the next decade. To facilitate sustainable live weight production, optimal reproductive performance is essential. Microbial colonisation in the reproductive tract has been demonstrated as one of the factors contributing to bovine reproductive performance. Studies also implied that reproductive metagenomes are different at each stage of the estrous cycle. This study applied Oxford Nanopore Technologies’ adaptive long-read sequencing to profile the bovine reproductive microbiome collected from tropical cattle in northern Queensland, Australia. The microbiome samples were collected from cattle of different sexes, reproductive status and locations to provide a comprehensive view of the bovine reproductive microbiome in northern Australian cattle. Ascomycota, Firmicutes and Proteobacteria were abundant phyla identified in the bovine reproductive metagenomes of Australian cattle regardless of sexes, reproductive status and location. The species level taxonomical investigation suggested that gastrointestinal metagenome and the surrounding environment were potentially the origins of the bovine reproductive metagenome. Functional profiles further affirmed this implication, revealing that the reproductive metagenomes of the prepubertal and postpartum animals were dominated by microorganisms that catabolise dietary polysaccharides as an energy substrate while that of the pregnant animals had the function of harvesting energy from aromatic compounds. Bovine reproductive metagenome investigations can be employed to trace the origins of abnormal metagenomes, which is beneficial for disease prevention and control. Additionally, our results demonstrated different reproductive metagenome diversities between cattle from two different locations. The variation in diversity within one location can serve as the indicator of abnormal reproductive metagenome, but between locations inferences cannot be made. We suggest establishing localised metagenomic indices that can be used to infer abnormal reproductive metagenomes which contribute to abortion or sub-fertility.

Environmental variation effects fertility in tropical beef cattle

The northern Australia beef cattle industry operates in harsh environmental conditions which consistently suppress female fertility. To better understand the environmental effect on cattle raised extensively in northern Australia, new environmental descriptors were defined for 54 commercial herds located across the region. Three fertility traits, based on the presence of a corpus luteum at 600 d of age, indicating puberty, (CL Presence, n = 25,176), heifer pregnancy (n = 20,989) and first lactation pregnancy (n = 10,072) were recorded. Temperature, humidity, and rainfall were obtained from publicly available data based on herd location. Being pubertal at 600 d (i.e. CL Presence) increased the likelihood of success at heifer pregnancy and first lactation pregnancy (P < 0.05), underscoring the importance of early puberty in reproductive success. A temperature humidity index (THI) of 65–70 had a significant (P < 0.05) negative effect on first lactation pregnancy rate, heifer pregnancy and puberty at 600 d of age. Area under the curve of daily THI was significant (P < 0.05) and reduced the likelihood of pregnancy at first lactation and puberty at 600 days. Deviation from long-term average rainfall was not significant (P < 0.05) for any trait. Average daily weight gain had a significant and positive relationship (P < 0.05) for heifer and first lactation pregnancy. The results indicate that chronic or cumulative heat load is more determinantal to reproductive performance than acute heat stress. The reason for the lack of a clear relationship between acute heat stress and reproductive performance is unclear but may be partially explained by peak THI and peak nutrition coinciding at the same time. Sufficient evidence was found to justify the use of average daily weight gain and chronic heat load as descriptors to define an environmental gradient.

Prevalence of Tritrichomonas foetus in beef bulls slaughtered at two abattoirs in northern Australia

Bovine trichomoniasis, caused by the protozoal parasite Tritrichomonas foetus, is a highly contagious venereal disease characterised by early pregnancy loss, abortion and pyometra. Persistently infected bulls and cows are the primary reservoirs of infection in infected herds. This research investigated the prevalence of T. foetus infection in bulls from properties located across northern Australia and New South Wales. Preputial samples were collected from 606 bulls at slaughter and tested for T. foetus using the VetMAX‐Gold Trich Detection Kit (Thermo Fisher Scientific). The apparent prevalence of T. foetus infection varied between regions, with northern regions in the Northern Territory, Queensland and Western Australia showing a prevalence of 15.4%, 13.8% and 11.4%, respectively. There was some evidence of an association between infection and postcode (P = 0.06) and increasing bull age (P = 0.054). This study confirms that T. foetus infection is likely to be present in many beef breeding herds and contributing to lower than expected reproductive performance, particularly across northern Australia.

Pregnancy rate per cycle is heritable and reduces with cycle in naturally mated tropically adapted beef cows

Pregnancy in cattle is the outcome of the complex process of initiation of cycling, fertilization, maternal recognition of pregnancy and foeto-placental development. Though much is known about initiation of cycling and associated risk factors, there are virtually no data on pregnancy rate per cycle for naturally mated cattle, especially for extensively managed, tropically adapted genotypes, which this study aimed to determine. Tropical composite (Bos indicus and African Sanga crosses with Bos taurus) and Brahman cattle (n = 2,181) of known pedigree in four-year groups at four sites were mated annually for 84 days. Body condition, ovarian function, pregnancies, calving and lactation were monitored through six full reproductive cycles using 4-8 weekly ultrasound of the reproductive tract outside the calving period and daily monitoring during calving. From this, dates of commencement of cycling and conception in each year were estimated for each animal, enabling calculation of established pregnancy for consecutive 21-day periods while cycling and of pregnancies within four months of calving while lactating (P4M). Pregnancy per 21-day period (cycle) during mating for cycling animals averaged 63%, 71%, 41% and 28% in four consecutive cycles. Pregnant per cycle was 2%-11% higher in tropical composites than in Brahmans. The only other consistently significant risk to becoming pregnant was if cycling commenced later than three weeks before mating commenced. P4M averaged 62% and was lower for cows in sub-optimal body condition and in first-parity and later-calving cows. Pregnant per cycle was moderately heritable (~20%), while heritability was moderate to high (33%) for P4M. Selection for pregnant per cycle could be achieved indirectly by selection for P4M, a trait that is readily measured.

Breed-adjusted genomic relationship matrices as a method to account for population stratification in multibreed populations of tropically adapted beef heifers

Context Beef cattle breeds in Australia can broadly be broken up into two subspecies, namely, Bos indicus and Bos taurus. Due to the time since divergence between the subspecies, it is likely that mutations affecting quantitative traits have developed independently in each. Aims We hypothesise that this will affect the prediction accuracy of genomic selection of admixed and composite populations that include both ancestral subspecies. Our study investigates methods to quantify population stratification in a multibreed population of tropically adapted heifers, with the aim of improving prediction accuracy of genomic selection for reproductive maturity score. Methods We used genotypes and reproductive maturity phenotypes from 3695 tropically adapted heifers from three purebred populations, namely, Brahman, Santa Gertrudis and Droughtmaster. Two of these breeds, Santa Gertrudis and Droughtmaster, are stabilised composites of varying B. indicus × B. taurus ancestry, and the third breed, Brahman, has predominately B. indicus ancestry. Genotypes were imputed to three marker-panel densities and population stratification was accounted for in genomic relationship matrices by using breed-specific allele frequencies when calculating the genomic relationships among animals. Prediction accuracy and bias were determined using a five-fold cross validation of randomly selected multibreed cohorts. Key Results Our results showed that the use of breed-adjusted genomic relationship matrices did not improve either prediction accuracy or bias for a lowly heritable trait such as reproductive maturity score. However, using breed-adjusted genomic relationship matrices allowed the capture of a higher proportion of additive genetic effects when estimating variance components. Conclusions These findings suggest that, despite seeing no improvement in prediction accuracy, it may still be beneficial to use breed-adjusted genomic relationship matrices in multibreed populations to improve the estimation of variance components. Implications As such, genomic evaluations using breed-adjusted genomic relationship matrices may be beneficial in multibreed populations.

Use of whole-genome sequence data and novel genomic selection strategies to improve selection for age at puberty in tropically-adapted beef heifers

Background

In tropically-adapted beef heifers, application of genomic prediction for age at puberty has been limited due to low prediction accuracies. Our aim was to investigate novel methods of pre-selecting whole-genome sequence (WGS) variants and alternative analysis methodologies; including genomic best linear unbiased prediction (GBLUP) with multiple genomic relationship matrices (MGRM) and Bayesian (BayesR) analyses, to determine if prediction accuracy for age at puberty can be improved.

Methods

Genotypes and phenotypes were obtained from two research herds. In total, 868 Brahman and 960 Tropical Composite heifers were recorded in the first population and 3695 Brahman, Santa Gertrudis and Droughtmaster heifers were recorded in the second population. Genotypes were imputed to 23 million whole-genome sequence variants. Eight strategies were used to pre-select variants from genome-wide association study (GWAS) results using conditional or joint (COJO) analyses. Pre-selected variants were included in three models, GBLUP with a single genomic relationship matrix (SGRM), GBLUP MGRM and BayesR. Five-way cross-validation was used to test the effect of marker panel density (6 K, 50 K and 800 K), analysis model, and inclusion of pre-selected WGS variants on prediction accuracy.

Results

In all tested scenarios, prediction accuracies for age at puberty were highest in BayesR analyses. The addition of pre-selected WGS variants had little effect on the accuracy of prediction when BayesR was used. The inclusion of WGS variants that were pre-selected using a meta-analysis with COJO analyses by chromosome, fitted in a MGRM model, had the highest prediction accuracies in the GBLUP analyses, regardless of marker density. When the low-density (6 K) panel was used, the prediction accuracy of GBLUP was equal (0.42) to that with the high-density panel when only six additional sequence variants (identified using meta-analysis COJO by chromosome) were included.

Conclusions

While BayesR consistently outperforms other methods in terms of prediction accuracies, reasonable improvements in accuracy can be achieved when using GBLUP and low-density panels with the inclusion of a relatively small number of highly relevant WGS variants.

Multivariate genomic predictions for age at puberty in tropically adapted beef heifers

Heifers that have an earlier age at puberty often have greater lifetime productivity. Age at puberty is moderately heritable so selection should effectively reduce the number of days to puberty, and improve heifer productivity and profitability as a result. However, recording age at puberty is intensive, requiring repeat ovarian scanning to determine age at first corpus luteum (AGECL). Genomic selection has been proposed as a strategy to select for earlier age at puberty; however, large reference populations of cows with AGECL records and genotypes would be required to generate accurate GEBV for this trait. Reproductive maturity score (RMS) is a proxy trait for age at puberty for implementation in northern Australia beef herds, where large scale recording of AGECL is not feasible. RMS assigns a score of 0 to 5 from a single ovarian scan to describe ovarian maturity at ~600 d. Here we use multivariate genomic prediction to evaluate the value of a large RMS data set to improve accuracy of GEBV for age at puberty (AGECL). There were 882 Brahman and 990 Tropical Composite heifers with AGECL phenotypes, and an independent set of 974 Brahman, 1,798 Santa Gertrudis, and 910 Droughtmaster heifers with RMS phenotypes. All animals had 728,785 real or imputed SNP genotypes. The correlation of AGECL and RMS (h2 = 0.23) was estimated as -0.83 using the genomic information. This result also demonstrates that using genomic information it is possible to estimate genetic correlations between traits collected on different animals in different herds, with minimal or unknown pedigree linkage between them. Inclusion of heifers with RMS in the multi-trait model improved the accuracy of genomic evaluations for AGECL. Accuracy of RMS GEBV generally did not improve by adding heifers with AGECL phenotypes into the reference population. These results suggest that RMS and AGECL may be used together in a multi-trait prediction model to increase the accuracy of prediction for age at puberty in tropically adapted beef cattle.