Application of neural networks with back-propagation to genome-enabled prediction of complex traits in Holstein-Friesian and German Fleckvieh cattle

Background

Recently, artificial neural networks (ANN) have been proposed as promising machines for marker-based genomic predictions of complex traits in animal and plant breeding. ANN are universal approximators of complex functions, that can capture cryptic relationships between SNPs (single nucleotide polymorphisms) and phenotypic values without the need of explicitly defining a genetic model. This concept is attractive for high-dimensional and noisy data, especially when the genetic architecture of the trait is unknown. However, the properties of ANN for the prediction of future outcomes of genomic selection using real data are not well characterized and, due to high computational costs, using whole-genome marker sets is difficult. We examined different non-linear network architectures, as well as several genomic covariate structures as network inputs in order to assess their ability to predict milk traits in three dairy cattle data sets using large-scale SNP data. For training, a regularized back propagation algorithm was used. The average correlation between the observed and predicted phenotypes in a 20 times 5-fold cross-validation was used to assess predictive ability. A linear network model served as benchmark.

Results

Predictive abilities of different ANN models varied markedly, whereas differences between data sets were small. Dimension reduction methods enhanced prediction performance in all data sets, while at the same time computational cost decreased. For the Holstein-Friesian bull data set, an ANN with 10 neurons in the hidden layer achieved a predictive correlation of r=0.47 for milk yield when the entire marker matrix was used. Predictive ability increased when the genomic relationship matrix (r=0.64) was used as input and was best (r=0.67) when principal component scores of the marker genotypes were used. Similar results were found for the other traits in all data sets.

Conclusion

Artificial neural networks are powerful machines for non-linear genome-enabled predictions in animal breeding. However, to produce stable and high-quality outputs, variable selection methods are highly recommended, when the number of markers vastly exceeds sample size.     

Right-hand-side updating for fast computing of genomic breeding values

Background

Since both the number of SNPs (single nucleotide polymorphisms) used in genomic prediction and the number of individuals used in training datasets are rapidly increasing, there is an increasing need to improve the efficiency of genomic prediction models in terms of computing time and memory (RAM) required.

Methods

In this paper, two alternative algorithms for genomic prediction are presented that replace the originally suggested residual updating algorithm, without affecting the estimates. The first alternative algorithm continues to use residual updating, but takes advantage of the characteristic that the predictor variables in the model (i.e. the SNP genotypes) take only three different values, and is therefore termed “improved residual updating”. The second alternative algorithm, here termed “right-hand-side updating” (RHS-updating), extends the idea of improved residual updating across multiple SNPs. The alternative algorithms can be implemented for a range of different genomic predictions models, including random regression BLUP (best linear unbiased prediction) and most Bayesian genomic prediction models. To test the required computing time and RAM, both alternative algorithms were implemented in a Bayesian stochastic search variable selection model.

Results

Compared to the original algorithm, the improved residual updating algorithm reduced CPU time by 35.3 to 43.3%, without changing memory requirements. The RHS-updating algorithm reduced CPU time by 74.5 to 93.0% and memory requirements by 13.1 to 66.4% compared to the original algorithm.

Conclusions

The presented RHS-updating algorithm provides an interesting alternative to reduce both computing time and memory requirements for a range of genomic prediction models.     

Genomic prediction based on data from three layer lines: a comparison between linear methods

Background

The prediction accuracy of several linear genomic prediction models, which have previously been used for within-line genomic prediction, was evaluated for multi-line genomic prediction.

Methods

Compared to a conventional BLUP (best linear unbiased prediction) model using pedigree data, we evaluated the following genomic prediction models: genome-enabled BLUP (GBLUP), ridge regression BLUP (RRBLUP), principal component analysis followed by ridge regression (RRPCA), BayesC and Bayesian stochastic search variable selection. Prediction accuracy was measured as the correlation between predicted breeding values and observed phenotypes divided by the square root of the heritability. The data used concerned laying hens with phenotypes for number of eggs in the first production period and known genotypes. The hens were from two closely-related brown layer lines (B1 and B2), and a third distantly-related white layer line (W1). Lines had 1004 to 1023 training animals and 238 to 240 validation animals. Training datasets consisted of animals of either single lines, or a combination of two or all three lines, and had 30 508 to 45 974 segregating single nucleotide polymorphisms.

Results

Genomic prediction models yielded 0.13 to 0.16 higher accuracies than pedigree-based BLUP. When excluding the line itself from the training dataset, genomic predictions were generally inaccurate. Use of multiple lines marginally improved prediction accuracy for B2 but did not affect or slightly decreased prediction accuracy for B1 and W1. Differences between models were generally small except for RRPCA which gave considerably higher accuracies for B2. Correlations between genomic predictions from different methods were higher than 0.96 for W1 and higher than 0.88 for B1 and B2. The greater differences between methods for B1 and B2 were probably due to the lower accuracy of predictions for B1 (~0.45) and B2 (~0.40) compared to W1 (~0.76).

Conclusions

Multi-line genomic prediction did not affect or slightly improved prediction accuracy for closely-related lines. For distantly-related lines, multi-line genomic prediction yielded similar or slightly lower accuracies than single-line genomic prediction. Bayesian variable selection and GBLUP generally gave similar accuracies. Overall, RRPCA yielded the greatest accuracies for two lines, suggesting that using PCA helps to alleviate the “n ≪ p” problem in genomic prediction.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-014-0057-5) contains supplementary material, which is available to authorized users.     

Improved precision of QTL mapping using a nonlinear Bayesian method in a multi-breed population leads to greater accuracy of across-breed genomic predictions

Background

Genomic selection is increasingly widely practised, particularly in dairy cattle. However, the accuracy of current predictions using GBLUP (genomic best linear unbiased prediction) decays rapidly across generations, and also as selection candidates become less related to the reference population. This is likely caused by the effects of causative mutations being dispersed across many SNPs (single nucleotide polymorphisms) that span large genomic intervals. In this paper, we hypothesise that the use of a nonlinear method (BayesR), combined with a multi-breed (Holstein/Jersey) reference population will map causative mutations with more precision than GBLUP and this, in turn, will increase the accuracy of genomic predictions for selection candidates that are less related to the reference animals.

Results

BayesR improved the across-breed prediction accuracy for Australian Red dairy cattle for five milk yield and composition traits by an average of 7% over the GBLUP approach (Australian Red animals were not included in the reference population). Using the multi-breed reference population with BayesR improved accuracy of prediction in Australian Red cattle by 2 – 5% compared to using BayesR with a single breed reference population. Inclusion of 8478 Holstein and 3917 Jersey cows in the reference population improved accuracy of predictions for these breeds by 4 and 5%. However, predictions for Holstein and Jersey cattle were similar using within-breed and multi-breed reference populations. We propose that the improvement in across-breed prediction achieved by BayesR with the multi-breed reference population is due to more precise mapping of quantitative trait loci (QTL), which was demonstrated for several regions. New candidate genes with functional links to milk synthesis were identified using differential gene expression in the mammary gland.

Conclusions

QTL detection and genomic prediction are usually considered independently but persistence of genomic prediction accuracies across breeds requires accurate estimation of QTL effects. We show that accuracy of across-breed genomic predictions was higher with BayesR than with GBLUP and that BayesR mapped QTL more precisely. Further improvements of across-breed accuracy of genomic predictions and QTL mapping could be achieved by increasing the size of the reference population, including more breeds, and possibly by exploiting pleiotropic effects to improve mapping efficiency for QTL with small effects.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-014-0074-4) contains supplementary material, which is available to authorized users.     

Accuracy of genomic predictions in Bos indicus (Nellore) cattle

Background

Nellore cattle play an important role in beef production in tropical systems and there is great interest in determining if genomic selection can contribute to accelerate genetic improvement of production and fertility in this breed. We present the first results of the implementation of genomic prediction in a Bos indicus (Nellore) population.

Methods

Influential bulls were genotyped with the Illumina Bovine HD chip in order to assess genomic predictive ability for weight and carcass traits, gestation length, scrotal circumference and two selection indices. 685 samples and 320 238 single nucleotide polymorphisms (SNPs) were used in the analyses. A forward-prediction scheme was adopted to predict the genomic breeding values (DGV). In the training step, the estimated breeding values (EBV) of bulls were deregressed (dEBV) and used as pseudo-phenotypes to estimate marker effects using four methods: genomic BLUP with or without a residual polygenic effect (GBLUP20 and GBLUP0, respectively), a mixture model (Bayes C) and Bayesian LASSO (BLASSO). Empirical accuracies of the resulting genomic predictions were assessed based on the correlation between DGV and dEBV for the testing group.

Results

Accuracies of genomic predictions ranged from 0.17 (navel at weaning) to 0.74 (finishing precocity). Across traits, Bayesian regression models (Bayes C and BLASSO) were more accurate than GBLUP. The average empirical accuracies were 0.39 (GBLUP0), 0.40 (GBLUP20) and 0.44 (Bayes C and BLASSO). Bayes C and BLASSO tended to produce deflated predictions (i.e. slope of the regression of dEBV on DGV greater than 1). Further analyses suggested that higher-than-expected accuracies were observed for traits for which EBV means differed significantly between two breeding subgroups that were identified in a principal component analysis based on genomic relationships.

Conclusions

Bayesian regression models are of interest for future applications of genomic selection in this population, but further improvements are needed to reduce deflation of their predictions. Recurrent updates of the training population would be required to enable accurate prediction of the genetic merit of young animals. The technical feasibility of applying genomic prediction in a Bos indicus (Nellore) population was demonstrated. Further research is needed to permit cost-effective selection decisions using genomic information.     

Genomic prediction using imputed whole-genome sequence data in Holstein Friesian cattle

Background

In contrast to currently used single nucleotide polymorphism (SNP) panels, the use of whole-genome sequence data is expected to enable the direct estimation of the effects of causal mutations on a given trait. This could lead to higher reliabilities of genomic predictions compared to those based on SNP genotypes. Also, at each generation of selection, recombination events between a SNP and a mutation can cause decay in reliability of genomic predictions based on markers rather than on the causal variants. Our objective was to investigate the use of imputed whole-genome sequence genotypes versus high-density SNP genotypes on (the persistency of) the reliability of genomic predictions using real cattle data.

Methods

Highly accurate phenotypes based on daughter performance and Illumina BovineHD Beadchip genotypes were available for 5503 Holstein Friesian bulls. The BovineHD genotypes (631,428 SNPs) of each bull were used to impute whole-genome sequence genotypes (12,590,056 SNPs) using the Beagle software. Imputation was done using a multi-breed reference panel of 429 sequenced individuals. Genomic estimated breeding values for three traits were predicted using a Bayesian stochastic search variable selection (BSSVS) model and a genome-enabled best linear unbiased prediction model (GBLUP). Reliabilities of predictions were based on 2087 validation bulls, while the other 3416 bulls were used for training.

Results

Prediction reliabilities ranged from 0.37 to 0.52. BSSVS performed better than GBLUP in all cases. Reliabilities of genomic predictions were slightly lower with imputed sequence data than with BovineHD chip data. Also, the reliabilities tended to be lower for both sequence data and BovineHD chip data when relationships between training animals were low. No increase in persistency of prediction reliability using imputed sequence data was observed.

Conclusions

Compared to BovineHD genotype data, using imputed sequence data for genomic prediction produced no advantage. To investigate the putative advantage of genomic prediction using (imputed) sequence data, a training set with a larger number of individuals that are distantly related to each other and genomic prediction models that incorporate biological information on the SNPs or that apply stricter SNP pre-selection should be considered.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-015-0149-x) contains supplementary material, which is available to authorized users.     

Selection of haplotype variables from a high-density marker map for genomic prediction

Background

Using haplotype blocks as predictors rather than individual single nucleotide polymorphisms (SNPs) may improve genomic predictions, since haplotypes are in stronger linkage disequilibrium with the quantitative trait loci than are individual SNPs. It has also been hypothesized that an appropriate selection of a subset of haplotype blocks can result in similar or better predictive ability than when using the whole set of haplotype blocks. This study investigated genomic prediction using a set of haplotype blocks that contained the SNPs with large effects estimated from an individual SNP prediction model. We analyzed protein yield, fertility and mastitis of Nordic Holstein cattle, and used high-density markers (about 770k SNPs). To reach an optimum number of haplotype variables for genomic prediction, predictions were performed using subsets of haplotype blocks that contained a range of 1000 to 50 000 main SNPs.

Results

The use of haplotype blocks improved the prediction reliabilities, even when selection focused on only a group of haplotype blocks. In this case, the use of haplotype blocks that contained the 20 000 to 50 000 SNPs with the highest effect was sufficient to outperform the model that used all individual SNPs as predictors (up to 1.3 % improvement in prediction reliability for mastitis, compared to individual SNP approach), and the achieved reliabilities were similar to those using all haplotype blocks available in the genome data (from 0.6 % lower to 0.8 % higher reliability).

Conclusions

Haplotype blocks used as predictors can improve the reliability of genomic prediction compared to the individual SNP model. Furthermore, the use of a subset of haplotype blocks that contains the main SNP effects from genomic data could be a feasible approach to genomic prediction in dairy cattle, given an increase in density of genotype data available. The predictive ability of the models that use a subset of haplotype blocks was similar to that obtained using either all haplotype blocks or all individual SNPs, with the benefit of having a much lower computational demand.     

Genomic relationships based on X chromosome markers and accuracy of genomic predictions with and without X chromosome markers

Background

Although the X chromosome is the second largest bovine chromosome, markers on the X chromosome are not used for genomic prediction in some countries and populations. In this study, we presented a method for computing genomic relationships using X chromosome markers, investigated the accuracy of imputation from a low density (7K) to the 54K SNP (single nucleotide polymorphism) panel, and compared the accuracy of genomic prediction with and without using X chromosome markers.

Methods

The impact of considering X chromosome markers on prediction accuracy was assessed using data from Nordic Holstein bulls and different sets of SNPs: (a) the 54K SNPs for reference and test animals, (b) SNPs imputed from the 7K to the 54K SNP panel for test animals, (c) SNPs imputed from the 7K to the 54K panel for half of the reference animals, and (d) the 7K SNP panel for all animals. Beagle and Findhap were used for imputation. GBLUP (genomic best linear unbiased prediction) models with or without X chromosome markers and with or without a residual polygenic effect were used to predict genomic breeding values for 15 traits.

Results

Averaged over the two imputation datasets, correlation coefficients between imputed and true genotypes for autosomal markers, pseudo-autosomal markers, and X-specific markers were 0.971, 0.831 and 0.935 when using Findhap, and 0.983, 0.856 and 0.937 when using Beagle. Estimated reliabilities of genomic predictions based on the imputed datasets using Findhap or Beagle were very close to those using the real 54K data. Genomic prediction using all markers gave slightly higher reliabilities than predictions without X chromosome markers. Based on our data which included only bulls, using a G matrix that accounted for sex-linked relationships did not improve prediction, compared with a G matrix that did not account for sex-linked relationships. A model that included a polygenic effect did not recover the loss of prediction accuracy from exclusion of X chromosome markers.

Conclusions

The results from this study suggest that markers on the X chromosome contribute to accuracy of genomic predictions and should be used for routine genomic evaluation.     

Empirical and deterministic accuracies of across-population genomic prediction

Background

Differences in linkage disequilibrium and in allele substitution effects of QTL (quantitative trait loci) may hinder genomic prediction across populations. Our objective was to develop a deterministic formula to estimate the accuracy of across-population genomic prediction, for which reference individuals and selection candidates are from different populations, and to investigate the impact of differences in allele substitution effects across populations and of the number of QTL underlying a trait on the accuracy.

Methods

A deterministic formula to estimate the accuracy of across-population genomic prediction was derived based on selection index theory. Moreover, accuracies were deterministically predicted using a formula based on population parameters and empirically calculated using simulated phenotypes and a GBLUP (genomic best linear unbiased prediction) model. Phenotypes of 1033 Holstein-Friesian, 105 Groninger White Headed and 147 Meuse-Rhine-Yssel cows were simulated by sampling 3000, 300, 30 or 3 QTL from the available high-density SNP (single nucleotide polymorphism) information of three chromosomes, assuming a correlation of 1.0, 0.8, 0.6, 0.4, or 0.2 between allele substitution effects across breeds. The simulated heritability was set to 0.95 to resemble the heritability of deregressed proofs of bulls.

Results

Accuracies estimated with the deterministic formula based on selection index theory were similar to empirical accuracies for all scenarios, while accuracies predicted with the formula based on population parameters overestimated empirical accuracies by ~25 to 30%. When the between-breed genetic correlation differed from 1, i.e. allele substitution effects differed across breeds, empirical and deterministic accuracies decreased in proportion to the genetic correlation. Using a multi-trait model, it was possible to accurately estimate the genetic correlation between the breeds based on phenotypes and high-density genotypes. The number of QTL underlying the simulated trait did not affect the accuracy.

Conclusions

The deterministic formula based on selection index theory estimated the accuracy of across-population genomic predictions well. The deterministic formula using population parameters overestimated the across-population genomic accuracy, but may still be useful because of its simplicity. Both formulas could accommodate for genetic correlations between populations lower than 1. The number of QTL underlying a trait did not affect the accuracy of across-population genomic prediction using a GBLUP method.     

Genomic Prediction for Tuberculosis Resistance in Dairy Cattle

Background

The increasing prevalence of bovine tuberculosis (bTB) in the UK and the limitations of the currently available diagnostic and control methods require the development of complementary approaches to assist in the sustainable control of the disease. One potential approach is the identification of animals that are genetically more resistant to bTB, to enable breeding of animals with enhanced resistance. This paper focuses on prediction of resistance to bTB. We explore estimation of direct genomic estimated breeding values (DGVs) for bTB resistance in UK dairy cattle, using dense SNP chip data, and test these genomic predictions for situations when disease phenotypes are not available on selection candidates.

Methodology/Principal Findings

We estimated DGVs using genomic best linear unbiased prediction methodology, and assessed their predictive accuracies with a cross validation procedure and receiver operator characteristic (ROC) curves. Furthermore, these results were compared with theoretical expectations for prediction accuracy and area-under-the-ROC-curve (AUC). The dataset comprised 1151 Holstein-Friesian cows (bTB cases or controls). All individuals (592 cases and 559 controls) were genotyped for 727,252 loci (Illumina Bead Chip). The estimated observed heritability of bTB resistance was 0.23±0.06 (0.34 on the liability scale) and five-fold cross validation, replicated six times, provided a prediction accuracy of 0.33 (95% C.I.: 0.26, 0.40). ROC curves, and the resulting AUC, gave a probability of 0.58, averaged across six replicates, of correctly classifying cows as diseased or as healthy based on SNP chip genotype alone using these data.

Conclusions/Significance

These results provide a first step in the investigation of the potential feasibility of genomic selection for bTB resistance using SNP data. Specifically, they demonstrate that genomic selection is possible, even in populations with no pedigree data and on animals lacking bTB phenotypes. However, a larger training population will be required to improve prediction accuracies.     

Accuracy of genomic selection for a sib-evaluated trait using identity-by-state and identity-by-descent relationships

Background

GBLUP (genomic best linear unbiased prediction) uses high-density single nucleotide polymorphism (SNP) markers to construct genomic identity-by-state (IBS) relationship matrices. However, identity-by-descent (IBD) relationships can be accurately calculated for extremely sparse markers. Here, we compare the accuracy of prediction of genome-wide breeding values (GW-BV) for a sib-evaluated trait in a typical aquaculture population, assuming either IBS or IBD genomic relationship matrices, and by varying marker density and size of the training dataset.

Methods

A simulation study was performed, assuming a population with strong family structure over three subsequent generations. Traditional and genomic BLUP were used to estimate breeding values, the latter using either IBS or IBD genomic relationship matrices, with marker densities ranging from 10 to ~1200 SNPs/Morgan (M). Heritability ranged from 0.1 to 0.8, and phenotypes were recorded on 25 to 45 sibs per full-sib family (50 full-sib families). Models were compared based on their predictive ability (accuracy) with respect to true breeding values of unphenotyped (albeit genotyped) sibs in the last generation.

Results

As expected, genomic prediction had greater accuracy compared to pedigree-based prediction. At the highest marker density, genomic prediction based on IBS information (IBS-GS) was slightly superior to that based on IBD information (IBD-GS), while at lower densities (≤100 SNPs/M), IBD-GS was more accurate. At the lowest densities (10 to 20 SNPs/M), IBS-GS was even outperformed by the pedigree-based model. Accuracy of IBD-GS was stable across marker densities performing well even down to 10 SNPs/M (2.5 to 6.1% reduction in accuracy compared to ~1200 SNPs/M). Loss of accuracy due to reduction in the size of training datasets was moderate and similar for both genomic prediction models. The relative superiority of (high-density) IBS-GS over IBD-GS was more pronounced for traits with a low heritability.

Conclusions

Using dense markers, GBLUP based on either IBD or IBS relationship matrices proved to perform better than a pedigree-based model. However, accuracy of IBS-GS declined rapidly with decreasing marker densities, and was even outperformed by a traditional pedigree-based model at the lowest densities. In contrast, the accuracy of IBD-GS was very stable across marker densities.     

A comparison of principal component regression and genomic REML for genomic prediction across populations

Background

Genomic prediction faces two main statistical problems: multicollinearity and n ≪ p (many fewer observations than predictor variables). Principal component (PC) analysis is a multivariate statistical method that is often used to address these problems. The objective of this study was to compare the performance of PC regression (PCR) for genomic prediction with that of a commonly used REML model with a genomic relationship matrix (GREML) and to investigate the full potential of PCR for genomic prediction.

Methods

The PCR model used either a common or a semi-supervised approach, where PC were selected based either on their eigenvalues (i.e. proportion of variance explained by SNP (single nucleotide polymorphism) genotypes) or on their association with phenotypic variance in the reference population (i.e. the regression sum of squares contribution). Cross-validation within the reference population was used to select the optimum PCR model that minimizes mean squared error. Pre-corrected average daily milk, fat and protein yields of 1609 first lactation Holstein heifers, from Ireland, UK, the Netherlands and Sweden, which were genotyped with 50 k SNPs, were analysed. Each testing subset included animals from only one country, or from only one selection line for the UK.

Results

In general, accuracies of GREML and PCR were similar but GREML slightly outperformed PCR. Inclusion of genotyping information of validation animals into model training (semi-supervised PCR), did not result in more accurate genomic predictions. The highest achievable PCR accuracies were obtained across a wide range of numbers of PC fitted in the regression (from one to more than 1000), across test populations and traits. Using cross-validation within the reference population to derive the number of PC, yielded substantially lower accuracies than the highest achievable accuracies obtained across all possible numbers of PC.

Conclusions

On average, PCR performed only slightly less well than GREML. When the optimal number of PC was determined based on realized accuracy in the testing population, PCR showed a higher potential in terms of achievable accuracy that was not capitalized when PC selection was based on cross-validation. A standard approach for selecting the optimal set of PC in PCR remains a challenge.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-014-0060-x) contains supplementary material, which is available to authorized users.     

Accounting for trait architecture in genomic predictions of US Holstein cattle using a weighted realized relationship matrix

Background

Genomic BLUP (GBLUP) can predict breeding values for non-phenotyped individuals based on the identity-by-state genomic relationship matrix (G). The G matrix can be constructed from thousands of markers spread across the genome. The strongest assumption of G and consequently of GBLUP is that all markers contribute equally to the genetic variance of a trait. This assumption is violated for traits that are controlled by a small number of quantitative trait loci (QTL) or individual QTL with large effects. In this paper, we investigate the performance of using a weighted genomic relationship matrix (wG) that takes into consideration the genetic architecture of the trait in order to improve predictive ability for a wide range of traits. Multiple methods were used to calculate weights for several economically relevant traits in US Holstein dairy cattle. Predictive performance was tested by k-means cross-validation.

Results

Relaxing the GBLUP assumption of equal marker contribution by increasing the weight that is given to a specific marker in the construction of the trait-specific G resulted in increased predictive performance. The increase was strongest for traits that are controlled by a small number of QTL (e.g. fat and protein percentage). Furthermore, bias in prediction estimates was reduced compared to that resulting from the use of regular G. Even for traits with low heritability and lower general predictive performance (e.g. calving ease traits), weighted G still yielded a gain in accuracy.

Conclusions

Genomic relationship matrices weighted by marker realized variance yielded more accurate and less biased predictions for traits regulated by few QTL. Genome-wide association analyses were used to derive marker weights for creating weighted genomic relationship matrices. However, this can be cumbersome and prone to low stability over generations because of erosion of linkage disequilibrium between markers and QTL. Future studies may include other sources of information, such as functional annotation and gene networks, to better exploit the genetic architecture of traits and produce more stable predictions.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-015-0100-1) contains supplementary material, which is available to authorized users.     

Genomic prediction of genetic merit using LD-based haplotypes in the Nordic Holstein population

Background

A haplotype approach to genomic prediction using high density data in dairy cattle as an alternative to single-marker methods is presented. With the assumption that haplotypes are in stronger linkage disequilibrium (LD) with quantitative trait loci (QTL) than single markers, this study focuses on the use of haplotype blocks (haploblocks) as explanatory variables for genomic prediction. Haploblocks were built based on the LD between markers, which allowed variable reduction. The haploblocks were then used to predict three economically important traits (milk protein, fertility and mastitis) in the Nordic Holstein population.

Results

The haploblock approach improved prediction accuracy compared with the commonly used individual single nucleotide polymorphism (SNP) approach. Furthermore, using an average LD threshold to define the haploblocks (LD≥0.45 between any two markers) increased the prediction accuracies for all three traits, although the improvement was most significant for milk protein (up to 3.1 % improvement in prediction accuracy, compared with the individual SNP approach). Hotelling’s t-tests were performed, confirming the improvement in prediction accuracy for milk protein. Because the phenotypic values were in the form of de-regressed proofs, the improved accuracy for milk protein may be due to higher reliability of the data for this trait compared with the reliability of the mastitis and fertility data. Comparisons between best linear unbiased prediction (BLUP) and Bayesian mixture models also indicated that the Bayesian model produced the most accurate predictions in every scenario for the milk protein trait, and in some scenarios for fertility.

Conclusions

The haploblock approach to genomic prediction is a promising method for genomic selection in animal breeding. Building haploblocks based on LD reduced the number of variables without the loss of information. This method may play an important role in the future genomic prediction involving while genome sequences.     

Generalized Seasonal Autoregressive Integrated Moving Average Models for Count Data with Application to Malaria Time Series with Low Case Numbers

Introduction

With the renewed drive towards malaria elimination, there is a need for improved surveillance tools. While time series analysis is an important tool for surveillance, prediction and for measuring interventions’ impact, approximations by commonly used Gaussian methods are prone to inaccuracies when case counts are low. Therefore, statistical methods appropriate for count data are required, especially during “consolidation” and “pre-elimination” phases.

Methods

Generalized autoregressive moving average (GARMA) models were extended to generalized seasonal autoregressive integrated moving average (GSARIMA) models for parsimonious observation-driven modelling of non Gaussian, non stationary and/or seasonal time series of count data. The models were applied to monthly malaria case time series in a district in Sri Lanka, where malaria has decreased dramatically in recent years.

Results

The malaria series showed long-term changes in the mean, unstable variance and seasonality. After fitting negative-binomial Bayesian models, both a GSARIMA and a GARIMA deterministic seasonality model were selected based on different criteria. Posterior predictive distributions indicated that negative-binomial models provided better predictions than Gaussian models, especially when counts were low. The G(S)ARIMA models were able to capture the autocorrelation in the series.

Conclusions

G(S)ARIMA models may be particularly useful in the drive towards malaria elimination, since episode count series are often seasonal and non-stationary, especially when control is increased. Although building and fitting GSARIMA models is laborious, they may provide more realistic prediction distributions than do Gaussian methods and may be more suitable when counts are low.     

Consistency of linkage disequilibrium between Chinese and Nordic Holsteins and genomic prediction for Chinese Holsteins using a joint reference population

Background

In China, the reference population of genotyped Holstein cattle is relatively small with to date, 80 bulls and 2091 cows genotyped with the Illumina 54 K chip. Including genotyped Holstein cattle from other countries in the reference population could improve the accuracy of genomic prediction of the Chinese Holstein population. This study investigated the consistency of linkage disequilibrium between adjacent markers between the Chinese and Nordic Holstein populations, and compared the reliability of genomic predictions based on the Chinese reference population only or the combined Chinese and Nordic reference populations.

Methods

Genomic estimated breeding values of Chinese Holstein cattle were predicted using a single-trait GBLUP model based on the Chinese reference dataset, and using a two-trait GBLUP model based on a joint reference dataset that included both the Chinese and Nordic Holstein data.

Results

The extent of linkage disequilibrium was similar in the Chinese and Nordic Holstein populations and the consistency of linkage disequilibrium between the two populations was very high, with a correlation of 0.97. Genomic prediction using the joint versus the Chinese reference dataset increased reliabilities of genomic predictions of Chinese Holstein bulls in the test data from 0.22, 0.15 and 0.11 to 0.51, 0.47 and 0.36 for milk yield, fat yield and protein yield, respectively. Using five-fold cross-validation, reliabilities of genomic predictions of Chinese cows increased from 0.15, 0.12 and 0.15 to 0.26, 0.17 and 0.20 for milk yield, fat yield and protein yield, respectively.

Conclusions

The linkage disequilibrium between the two populations was very consistent and using the combined Nordic and Chinese reference dataset substantially increased reliabilities of genomic predictions for Chinese Holstein cattle.     

Predicting complex traits using a diffusion kernel on genetic markers with an application to dairy cattle and wheat data

Background

Arguably, genotypes and phenotypes may be linked in functional forms that are not well addressed by the linear additive models that are standard in quantitative genetics. Therefore, developing statistical learning models for predicting phenotypic values from all available molecular information that are capable of capturing complex genetic network architectures is of great importance. Bayesian kernel ridge regression is a non-parametric prediction model proposed for this purpose. Its essence is to create a spatial distance-based relationship matrix called a kernel. Although the set of all single nucleotide polymorphism genotype configurations on which a model is built is finite, past research has mainly used a Gaussian kernel.

Results

We sought to investigate the performance of a diffusion kernel, which was specifically developed to model discrete marker inputs, using Holstein cattle and wheat data. This kernel can be viewed as a discretization of the Gaussian kernel. The predictive ability of the diffusion kernel was similar to that of non-spatial distance-based additive genomic relationship kernels in the Holstein data, but outperformed the latter in the wheat data. However, the difference in performance between the diffusion and Gaussian kernels was negligible.

Conclusions

It is concluded that the ability of a diffusion kernel to capture the total genetic variance is not better than that of a Gaussian kernel, at least for these data. Although the diffusion kernel as a choice of basis function may have potential for use in whole-genome prediction, our results imply that embedding genetic markers into a non-Euclidean metric space has very small impact on prediction. Our results suggest that use of the black box Gaussian kernel is justified, given its connection to the diffusion kernel and its similar predictive performance.     

International genomic evaluation methods for dairy cattle

Background

Genomic evaluations are rapidly replacing traditional evaluation systems used for dairy cattle selection. Higher reliabilities from larger genotype files promote cooperation across country borders. Genomic information can be exchanged across countries using simple conversion equations, by modifying multi-trait across-country evaluation (MACE) to account for correlated residuals originating from the use of foreign evaluations, or by multi-trait analysis of genotypes for countries that use the same reference animals.

Methods

Traditional MACE assumes independent residuals because each daughter is measured in only one country. Genomic MACE could account for residual correlations using daughter equivalents from genomic data as a fraction of the total in each country and proportions of bulls shared. MACE methods developed to combine separate within-country genomic evaluations were compared to direct, multi-country analysis of combined genotypes using simulated genomic and phenotypic data for 8,193 bulls in nine countries.

Results

Reliabilities for young bulls were much higher for across-country than within-country genomic evaluations as measured by squared correlations of estimated with true breeding values. Gains in reliability from genomic MACE were similar to those of multi-trait evaluation of genotypes but required less computation. Sharing of reference genotypes among countries created large residual correlations, especially for young bulls, that are accounted for in genomic MACE.

Conclusions

International genomic evaluations can be computed either by modifying MACE to account for residual correlations across countries or by multi-trait evaluation of combined genotype files. The gains in reliability justify the increased computation but require more cooperation than in previous breeding programs.     

Comparison of Bayesian models to estimate direct genomic values in multi-breed commercial beef cattle

Background

While several studies have examined the accuracy of direct genomic breeding values (DGV) within and across purebred cattle populations, the accuracy of DGV in crossbred or multi-breed cattle populations has been less well examined. Interest in the use of genomic tools for both selection and management has increased within the hybrid seedstock and commercial cattle sectors and research is needed to determine their efficacy. We predicted DGV for six traits using training populations of various sizes and alternative Bayesian models for a population of 3240 crossbred animals. Our objective was to compare alternate models with different assumptions regarding the distributions of single nucleotide polymorphism (SNP) effects to determine the optimal model for enhancing feasibility of multi-breed DGV prediction for the commercial beef industry.

Results

Realized accuracies ranged from 0.40 to 0.78. Randomly assigning 60 to 70% of animals to training (n ≈ 2000 records) yielded DGV accuracies with the smallest coefficients of variation. Mixture models (BayesB95, BayesCπ) and models that allow SNP effects to be sampled from distributions with unequal variances (BayesA, BayesB95) were advantageous for traits that appear or are known to be influenced by large-effect genes. For other traits, models differed little in prediction accuracy (~0.3 to 0.6%), suggesting that they are mainly controlled by small-effect loci.

Conclusions

The proportion (60 to 70%) of data allocated to training that optimized DGV accuracy and minimized the coefficient of variation of accuracy was similar to large dairy populations. Larger effects were estimated for some SNPs using BayesA and BayesB95 models because they allow unequal SNP variances. This substantially increased DGV accuracy for Warner-Bratzler Shear Force, for which large-effect quantitative trait loci (QTL) are known, while no loss in accuracy was observed for traits that appear to follow the infinitesimal model. Large decreases in accuracy (up to 0.07) occurred when SNPs that presumably tag large-effect QTL were over-regressed towards the mean in BayesC0 analyses. The DGV accuracies achieved here indicate that genomic selection has predictive utility in the commercial beef industry and that using models that reflect the genomic architecture of the trait can have predictive advantages in multi-breed populations.

Electronic supplementary material

The online version of this article (doi:10.1186/s12711-015-0106-8) contains supplementary material, which is available to authorized users.     

Comparison of Prediction Model for Cardiovascular Autonomic Dysfunction Using Artificial Neural Network and Logistic Regression Analysis

Background

This study aimed to develop the artificial neural network (ANN) and multivariable logistic regression (LR) analyses for prediction modeling of cardiovascular autonomic (CA) dysfunction in the general population, and compare the prediction models using the two approaches.

Methods and Materials

We analyzed a previous dataset based on a Chinese population sample consisting of 2,092 individuals aged 30–80 years. The prediction models were derived from an exploratory set using ANN and LR analysis, and were tested in the validation set. Performances of these prediction models were then compared.

Results

Univariate analysis indicated that 14 risk factors showed statistically significant association with the prevalence of CA dysfunction (P<0.05). The mean area under the receiver-operating curve was 0.758 (95% CI 0.724–0.793) for LR and 0.762 (95% CI 0.732–0.793) for ANN analysis, but noninferiority result was found (P<0.001). The similar results were found in comparisons of sensitivity, specificity, and predictive values in the prediction models between the LR and ANN analyses.

Conclusion

The prediction models for CA dysfunction were developed using ANN and LR. ANN and LR are two effective tools for developing prediction models based on our dataset.