Nina Melzer

Including non-additive genetic effects in Bayesian methods for the prediction of genetic values based on genome-wide markers

BackgroundMolecular marker information is a common source to draw inferences about the relationship between genetic and phenotypic variation. Genetic effects are often modelled as additively acting marker allele effects. The true mode of biological action can, of course, be different from this plain assumption. One possibility to better understand the genetic architecture of complex traits is to include intra-locus (dominance) and inter-locus (epistasis) interaction of alleles as well as the additive genetic effects when fitting a model to a trait. Several Bayesian MCMC approaches exist for the genome-wide estimation of genetic effects with high accuracy of genetic value prediction. Including pairwise interaction for thousands of loci would probably go beyond the scope of such a sampling algorithm because then millions of effects are to be estimated simultaneously leading to months of computation time. Alternative solving strategies are required when epistasis is studied.MethodsWe extended a fast Bayesian method (fBayesB), which was previously proposed for a purely additive model, to include non-additive effects. The fBayesB approach was used to estimate genetic effects on the basis of simulated datasets. Different scenarios were simulated to study the loss of accuracy of prediction, if epistatic effects were not simulated but modelled and vice versa.ResultsIf 23 QTL were simulated to cause additive and dominance effects, both fBayesB and a conventional MCMC sampler BayesB yielded similar results in terms of accuracy of genetic value prediction and bias of variance component estimation based on a model including additive and dominance effects. Applying fBayesB to data with epistasis, accuracy could be improved by 5% when all pairwise interactions were modelled as well. The accuracy decreased more than 20% if genetic variation was spread over 230 QTL. In this scenario, accuracy based on modelling only additive and dominance effects was generally superior to that of the complex model including epistatic effects.ConclusionsThis simulation study showed that the fBayesB approach is convenient for genetic value prediction. Jointly estimating additive and non-additive effects (especially dominance) has reasonable impact on the accuracy of prediction and the proportion of genetic variation assigned to the additive genetic source.

Investigating associations between milk metabolite profiles and milk traits of Holstein cows

In the field of dairy cattle research, it is of great interest to improve the detection and prevention of diseases (e.g., mastitis and ketosis) and monitor specific traits related to the state of health and management. During the standard milk performance test, traditional milk traits are monitored, and quality and quantity are screened. In addition to the standard test, it is also now possible to analyze milk metabolites in a high-throughput manner and to consider them in connection with milk traits to identify functionally important metabolites that can also serve as biomarker candidates. We present a study in which 190 milk metabolites and 14 milk traits of 1,305 Holstein cows on 18 commercial farms were investigated to characterize interrelations of milk metabolites between each other, to milk traits from the milk standard performance test, and to influencing factors such as farm and sire effect (half-sib structure). The effect of influencing factors (e.g., farm) varied among metabolites and traditional milk traits. The investigations of associations between metabolites and milk traits revealed groups of metabolites that show, for example, positive correlations to protein and casein, and negative correlations to lactose and pH. On the other hand, groups of metabolites jointly associated with the investigated milk traits can be identified and functionally discussed. To enable a multivariate investigation, 2 machine learning methods were applied to detect important metabolites that are highly correlated with the investigated traditional milk traits. For somatic cell score, uracil, lactic acid, and 9 other important metabolites were detected. Lactic acid has already been proposed as a biomarker candidate for mastitis in the recent literature. In conclusion, we found sets of metabolites eligible to predict milk traits, enabling the analysis of milk traits from a metabolic perspective and discussion of the possible functional background for some of the detected associations.

Milk metabolites and their genetic variability

The composition of milk is crucial to evaluate milk performance and quality measures. Milk components partly contribute to breeding scores, and they can be assessed to judge metabolic and energy status of the cow as well as to serve as predictive markers for diseases. In addition to the milk composition measures (e.g., fat, protein, lactose) traditionally recorded during milk performance test via infrared spectroscopy, novel techniques, such as gas chromatography-mass spectrometry, allow for a further analysis of milk into its metabolic components. Gas chromatography-mass spectrometry is suitable for measuring several hundred metabolites with high throughput, and thus it is applicable to study sources of genetic and nongenetic variation of milk metabolites in dairy cows. Heritability and mode of inheritance of metabolite measurements were studied in a linear mixed model approach including expected (pedigree) and realized (genomic) relationship between animals. The genetic variability of 190 milk metabolite intensities was analyzed from 1,295 cows held on 18 farms in Mecklenburg-Western Pomerania, Germany. Besides extensive pedigree information, genotypic data comprising 37,180 single nucleotide polymorphism markers were available. Goodness of fit and significance of genetic variance components based on likelihood ratio tests were investigated with a full model, including marker- and pedigree-based genetic effects. Broad-sense heritability varied from zero to 0.699, with a median of 0.125. Significant additive genetic variance was observed for highly heritable metabolites, but dominance variance was not significantly present. As some metabolites are particularly favorable for human nutrition, for instance, future research should address the identification of locus-specific genetic effects and investigate metabolites as the molecular basis of traditional milk performance test traits.

Genomic additive and dominance variance of milk performance traits

Milk performance traits are likely influenced by both additive and non-additive (e.g. dominance) genetic effects. Genetic variation can be partitioned using genomic information. The objective of this study was to estimate genetic variance components of production and milk component traits (e.g. acetone, fatty acids), which are particularly important for milk processing or which can provide information on the health status of cows. A genomic relationship approach was applied to phenotypic and genetic information of 1295 Holstein cows for estimating additive genetic and dominance variance components. Most of the 17 investigated traits were mainly affected by additive genetic effects, but protein content and casein content also showed a significant contribution of dominance. The ratio of dominance to additive variance was estimated as 0.64 for protein content and 0.56 for casein content. This ratio was highest for SCS (1.36) although dominance was not significant. Dominance effects were negligible in other moderately heritable milk traits.

Integrating milk metabolite profile information for the prediction of traditional milk traits based on SNP information for Holstein cows.

In this study the benefit of metabolome level analysis for the prediction of genetic value of three traditional milk traits was investigated. Our proposed approach consists of three steps: First, milk metabolite profiles are used to predict three traditional milk traits of 1,305 Holstein cows. Two regression methods, both enabling variable selection, are applied to identify important milk metabolites in this step. Second, the prediction of these important milk metabolite from single nucleotide polymorphisms (SNPs) enables the detection of SNPs with significant genetic effects. Finally, these SNPs are used to predict milk traits. The observed precision of predicted genetic values was compared to the results observed for the classical genotype-phenotype prediction using all SNPs or a reduced SNP subset (reduced classical approach). To enable a comparison between SNP subsets, a special invariable evaluation design was implemented. SNPs close to or within known quantitative trait loci (QTL) were determined. This enabled us to determine if detected important SNP subsets were enriched in these regions. The results show that our approach can lead to genetic value prediction, but requires less than 1% of the total amount of (40,317) SNPs., significantly more important SNPs in known QTL regions were detected using our approach compared to the reduced classical approach. Concluding, our approach allows a deeper insight into the associations between the different levels of the genotype-phenotype map (genotype-metabolome, metabolome-phenotype, genotype-phenotype).

Evaluating the temporal and situational consistency of personality traits in adult dairy cattle

Recent research suggests that personality, defined as consistent individual behavioral variation, in farm animals could be an important factor when considering their health, welfare, and productivity. However, behavioral tests are often performed individually and they might not reflect the behavioral differences manifested in every-day social environments. Furthermore, the contextual and longer-term temporal stability of personality traits have rarely been investigated in adult dairy cattle. In this study, we tested three groups of lactating Holstein cows (40 cows) using an individual arena test and a novel object test in groups to measure the contextual stability of behavior. Among the recorded individual test parameters, we used seven in the final analysis, which were determined by a systematic parameter reduction procedure. We found positive correlations between novel object contact duration in the group test and individual test parameters object contact duration (Rs = 0.361, P = 0.026) and movement duration (Rs = 0.336, P = 0.039). Both tests were repeated 6 months later to investigate their temporal stability whereby four individual test parameters were repeatable. There was no consistency in the group test results for 25 cows tested twice, possibly due to group composition changes. Furthermore, based on the seven individual test parameters, two personality traits (activity/exploration and boldness) were identified by principal component analysis. We found a positive association between the first and second tests for activity/exploration (Rs = 0.334, P = 0.058) and for boldness (Rs = 0.491, P = 0.004). Our results support the multidimensional nature of personality in adult dairy cattle and they indicate a link between behavior in individual and within-group situations. The lack of stability according to the group test results implies that group companions might have a stronger influence on individual behavior than expected. We suggest repeating the within-group behavioral measurements to study the relationship between the social environment and the manifestation of personality traits in every-day situations.

Short communication: Estimating lactation curves for highly inhomogeneous milk yield data of an F2 population (Charolais × German Holstein)

Fitting of lactation curves is a common tool to obtain the entire milk yield as well as to estimate the main curve characteristic (such as day of peak milk yield) for a lactation. These models are primarily designed for dairy cattle, but have been applied to nondairy cattle breeds and also for other species. In this study we considered milk yield data of 197 F2 crossbred cows of Charolais and German Holstein (founder breeds) for the first and the beginning of the second lactation. The F2 cows showed a high variability regarding the length of lactation, which varied between 7 and 406 d in milk for the first lactation. Thus, the data also show high variation regarding the daily and overall milk yield. To obtain complete lactation curves, we evaluated the lactation models of Ali-Schaeffer and Wilmink. To compare the 2 lactation models, we evaluated the goodness of fit using 6 evaluation criteria. The results show that the model of Ali-Schaeffer performs better on these highly inhomogeneous data, in contrast to the model of Wilmink. We discuss our findings from a statistical point of view and present possible biological reasons for the high variability regarding milk yield within the F2 population. Hence our findings may be helpful when milk yield data of crosses between dairy and beef cows (dual purpose) are investigated, whose lactation curves may not show the typical characteristics of dairy cattle.