Accuracies of genomic breeding values in American Angus beef cattle using K-means clustering for cross-validation

Background

Genomic selection is a recently developed technology that is beginning to revolutionize animal breeding. The objective of this study was to estimate marker effects to derive prediction equations for direct genomic values for 16 routinely recorded traits of American Angus beef cattle and quantify corresponding accuracies of prediction.

Methods

Deregressed estimated breeding values were used as observations in a weighted analysis to derive direct genomic values for 3570 sires genotyped using the Illumina BovineSNP50 BeadChip. These bulls were clustered into five groups using K-means clustering on pedigree estimates of additive genetic relationships between animals, with the aim of increasing within-group and decreasing between-group relationships. All five combinations of four groups were used for model training, with cross-validation performed in the group not used in training. Bivariate animal models were used for each trait to estimate the genetic correlation between deregressed estimated breeding values and direct genomic values.

Results

Accuracies of direct genomic values ranged from 0.22 to 0.69 for the studied traits, with an average of 0.44. Predictions were more accurate when animals within the validation group were more closely related to animals in the training set. When training and validation sets were formed by random allocation, the accuracies of direct genomic values ranged from 0.38 to 0.85, with an average of 0.65, reflecting the greater relationship between animals in training and validation. The accuracies of direct genomic values obtained from training on older animals and validating in younger animals were intermediate to the accuracies obtained from K-means clustering and random clustering for most traits. The genetic correlation between deregressed estimated breeding values and direct genomic values ranged from 0.15 to 0.80 for the traits studied.

Conclusions

These results suggest that genomic estimates of genetic merit can be produced in beef cattle at a young age but the recurrent inclusion of genotyped sires in retraining analyses will be necessary to routinely produce for the industry the direct genomic values with the highest accuracy.     

Breeding value prediction for production traits in layer chickens using pedigree or genomic relationships in a reduced animal model

Background

Genomic selection involves breeding value estimation of selection candidates based on high-density SNP genotypes. To quantify the potential benefit of genomic selection, accuracies of estimated breeding values (EBV) obtained with different methods using pedigree or high-density SNP genotypes were evaluated and compared in a commercial layer chicken breeding line.

Methods

The following traits were analyzed: egg production, egg weight, egg color, shell strength, age at sexual maturity, body weight, albumen height, and yolk weight. Predictions appropriate for early or late selection were compared. A total of 2,708 birds were genotyped for 23,356 segregating SNP, including 1,563 females with records. Phenotypes on relatives without genotypes were incorporated in the analysis (in total 13,049 production records).The data were analyzed with a Reduced Animal Model using a relationship matrix based on pedigree data or on marker genotypes and with a Bayesian method using model averaging. Using a validation set that consisted of individuals from the generation following training, these methods were compared by correlating EBV with phenotypes corrected for fixed effects, selecting the top 30 individuals based on EBV and evaluating their mean phenotype, and by regressing phenotypes on EBV.

Results

Using high-density SNP genotypes increased accuracies of EBV up to two-fold for selection at an early age and by up to 88% for selection at a later age. Accuracy increases at an early age can be mostly attributed to improved estimates of parental EBV for shell quality and egg production, while for other egg quality traits it is mostly due to improved estimates of Mendelian sampling effects. A relatively small number of markers was sufficient to explain most of the genetic variation for egg weight and body weight.     

Evaluation of coagulation activation after Rhinovirus infection in patients with asthma and healthy control subjects: an observational study

Background

Asthma exacerbations are frequently triggered by rhinovirus infections. Both asthma and respiratory tract infection can activate haemostasis. Therefore we hypothesized that experimental rhinovirus-16 infection and asthmatic airway inflammation act in synergy on the haemostatic balance.

Methods

28 patients (14 patients with mild allergic asthma and 14 healthy non-allergic controls) were infected with low-dose rhinovirus type 16. Venous plasma and bronchoalveolar lavage fluid (BAL fluid) were obtained before and 6 days after infection to evaluate markers of coagulation activation, thrombin-antithrombin complexes, von Willebrand factor, plasmin-antiplasmin complexes, plasminogen activator inhibitor type-1, endogenous thrombin potential and tissue factor-exposing microparticles by fibrin generation test, in plasma and/or BAL fluid. Data were analysed by nonparametric tests (Wilcoxon, Mann Whitney and Spearman correlation).

Results

13 patients with mild asthma (6 females, 19-29 y) and 11 healthy controls (10 females, 19-31 y) had a documented Rhinovirus-16 infection. Rhinovirus-16 challenge resulted in a shortening of the fibrin generation test in BAL fluid of asthma patients (t = -1: 706 s vs. t = 6: 498 s; p = 0.02), but not of controls (t = -1: 693 s vs. t = 6: 636 s; p = 0.65). The fold change in tissue factor-exposing microparticles in BAL fluid inversely correlated with the fold changes in eosinophil cationic protein and myeloperoxidase in BAL fluid after virus infection (r = -0.517 and -0.528 resp., both p = 0.01).Rhinovirus-16 challenge led to increased plasminogen activator inhibitor type-1 levels in plasma in patients with asthma (26.0 ng/mL vs. 11.5 ng/mL in healthy controls, p = 0.04). Rhinovirus-16 load in BAL showed a linear correlation with the fold change in endogenous thrombin potential, plasmin-antiplasmin complexes and plasminogen activator inhibitor type-1.

Conclusions

Experimental rhinovirus infection induces procoagulant changes in the airways of patients with asthma through increased activity of tissue factor-exposing microparticles. These microparticle-associated procoagulant changes are associated with both neutrophilic and eosinophilic inflammation. Systemic activation of haemostasis increases with Rhinoviral load.

Trial registration

This trial was registered at the Dutch trial registry (http://www.trialregister.nl): NTR1677.     

The effect of using approximate gametic variance covariance matrices on marker assisted selection by BLUP

Under additive inheritance, the Henderson mixed model equations (HMME) provide an efficient approach to obtaining genetic evaluations by marker assisted best linear unbiased prediction (MABLUP) given pedigree relationships, trait and marker data. For large pedigrees with many missing markers, however, it is not feasible to calculate the exact gametic variance covariance matrix required to construct HMME. The objective of this study was to investigate the consequences of using approximate gametic variance covariance matrices on response to selection by MABLUP. Two methods were used to generate approximate variance covariance matrices. The first method (Method A) completely discards the marker information for individuals with an unknown linkage phase between two flanking markers. The second method (Method B) makes use of the marker information at only the most polymorphic marker locus for individuals with an unknown linkage phase. Data sets were simulated with and without missing marker data for flanking markers with 2, 4, 6, 8 or 12 alleles. Several missing marker data patterns were considered. The genetic variability explained by marked quantitative trait loci (MQTL) was modeled with one or two MQTL of equal effect. Response to selection by MABLUP using Method A or Method B were compared with that obtained by MABLUP using the exact genetic variance covariance matrix, which was estimated using 15 000 samples from the conditional distribution of genotypic values given the observed marker data. For the simulated conditions, the superiority of MABLUP over BLUP based only on pedigree relationships and trait data varied between 0.1% and 13.5% for Method A, between 1.7% and 23.8% for Method B, and between 7.6% and 28.9% for the exact method. The relative performance of the methods under investigation was not affected by the number of MQTL in the model.     

Local spread of classical swine fever upon virus introduction into The Netherlands: Mapping of areas at high risk

Background

In the recent past, the introduction of Classical Swine Fever Virus (CSFV) followed by between-herd spread has given rise to a number of large epidemics in The Netherlands and Belgium. Both these countries are pork-exporting countries. Particularly important in these epidemics has been the occurrence of substantial "neighborhood transmission" from herd to herd in the presence of base-line control measures prescribed by EU legislation. Here we propose a calculation procedure to map out "high-risk areas" for local between-herd spread of CSFV as a tool to support decision making on prevention and control of CSFV outbreaks. In this procedure the identification of such areas is based on an estimated inter-herd distance dependent probability of neighborhood transmission or "local transmission". Using this distance-dependent probability, we derive a threshold value for the local density of herds. In areas with local herd density above threshold, local transmission alone can already lead to epidemic spread, whereas in below-threshold areas this is not the case. The first type of area is termed 'high-risk' for spread of CSFV, while the latter type is termed 'low-risk'.

Results

As we show for the case of The Netherlands, once the distance-dependent probability of local transmission has been estimated from CSFV outbreak data, it is possible to produce a map of the country in which areas of high-risk herds and of low-risk herds are identified. We made these maps even more informative by estimating border zones between the two types of areas. In these border zones the risk of local transmission of infection to a nearby high-risk area exceeds a certain level.

Conclusion

The risk maps provide an easily understandable visualization of the spatial heterogeneities in transmission risk. They serve as a tool for area-specific designs of control strategies, and possibly also for spatial planning of areas where livestock farming is allowed. Similar risk maps can in principle be constructed for other highly-transmissible livestock infections that spread via neighborhood transmission.

     

Microsatellite variation in Drosophila melanogaster and Drosophila simulans: a reciprocal test of the ascertainment bias hypothesis

Interspecific comparisons of microsatellite loci have repeatedly shown that the loci are longer and more variable in the species from which they are derived (the focal species) than are homologous loci in other (nonfocal) species. There is debate as to whether this is due to directional evolution or to an ascertainment bias during the cloning and locus selection processes. This study tests these hypotheses by performing a reciprocal study. Eighteen perfect dinucleotide microsatellite loci identified from a Drosophila simulans library screen and 18 previously identified in an identical Drosophila melanogaster library screen were used to survey natural populations of each species. No difference between focal and nonfocal species was observed for mean PCR fragment length. However, heterozygosity and number of alleles were significantly higher in the focal species than in the nonfocal species. The most common allele in the Zimbabwe population of both species was sequenced for 31 of the 36 loci. The length of the longest stretch of perfect repeat units is, on average, longer in the focal species than in the non-focal species. There is a positive correlation between the length of the longest stretch of perfect repeats and heterozygosity. The difference in heterozygosity can thus be explained by a reduction in the length of the longest stretch of perfect repeats in the nonfocal species. Furthermore, flanking-sequence length difference was noted between the two species at 58% of the loci sequenced. These data do not support the predictions of the directional-evolution hypothesis; however, consistent with the ascertainment bias hypothesis, the lower variability in nonfocal species is an artifact of the microsatellite cloning and isolation process. Our results also suggest that the magnitude of ascertainment bias for repeat unit length is a function of the microsatellite size distribution in the genomes of different species.      

Nitrogen dynamics of a boreal black spruce wildfire chronosequence

This study examined the nitrogen (N) dynamics of a black spruce (Picea mariana (Mill.) BSP)-dominated chronosequence in Manitoba, Canada. The seven sites studied each contained separate well- and poorly drained stands, originated from stand-killing wildfires, and were between 3 and 151 years old. Our goals were to (i) measure total N concentration ([N]) of all biomass components and major soil horizons; (ii) compare N content and select vegetation N cycle processes among the stands; and (iii) examine relationships between ecosystem C and N cycling for these stands. Vegetation [N] varied significantly by tissue type, species, soil drainage, and stand age; woody debris [N] increased with decay state and decreased with debris size. Soil [N] declined with horizon depth but did not vary with stand age. Total (live + dead) biomass N content ranged from 18.4 to 99.7 g N m⁻² in the well-drained stands and 37.8–154.6 g N m⁻² in the poorly drained stands. Mean soil N content (380.6 g N m⁻²) was unaffected by stand age. Annual vegetation N requirement (5.9 and 8.4 g N m⁻² yr⁻¹ in the middle-aged well- and poorly drained stands, respectively) was dominated by trees and fine roots in the well-drained stands, and bryophytes in the poorly drained stands. Fraction N retranslocated was significantly higher in deciduous than evergreen tree species, and in older than younger stands. Nitrogen use efficiency (NUE) was significantly lower in bryophytes than in trees, and in deciduous than in evergreen trees. Tree NUE increased with stand age, but overall stand NUE was roughly constant (∼ ∼150 g g⁻¹ N) across the entire chronosequence.     

Total cost estimation for implementing genome-enabled selection in a multi-level swine production system

Background

Determining an animal’s genetic merit using genomic information can improve estimated breeding value (EBV) accuracy; however, the magnitude of the accuracy improvement must be large enough to recover the costs associated with implementing genome-enabled selection. One way to reduce costs is to genotype nucleus herd selection candidates using a low-density chip and to use high-density chip genotyping for animals that are used as parents in the nucleus breeding herd. The objective of this study was to develop a tool to estimate the cost structure associated with incorporating genome-enabled selection into multi-level commercial breeding programs.

Results

For the purpose of this deterministic study, it was assumed that a commercial pig is created from a terminal line sire and a dam that is a cross between two maternal lines. It was also assumed that all male and female selection candidates from the 1000 sow maternal line nucleus herds were genotyped at low density and all animals used for breeding at high density. With the assumptions used in this analysis, it was estimated that genome-enabled selection costs for a maternal line would be approximately US$0.082 per weaned pig in the commercial production system. A total of US$0.164 per weaned pig is needed to incorporate genome-enabled selection into the two maternal lines. Similarly, for a 600 sow terminal line nucleus herd and genotyping only male selection candidates with the low-density panel, the cost per weaned pig in the commercial herd was estimated to be US$0.044. This means that US$0.21 per weaned pig produced at the commercial level and sired by boars obtained from the nucleus herd breeding program needs to be added to the genetic merit value in order to break even on the additional cost required when genome-enabled selection is used in both maternal lines and the terminal line.

Conclusions

By modifying the input values, such as herd size and genotyping strategy, a flexible spreadsheet tool developed from this work can be used to estimate the additional costs associated with genome-enabled selection. This tool will aid breeders in estimating the economic viability of incorporating genome-enabled selection into their specific breeding program.