I. Aguilar

Hot topic: A unified approach to utilize phenotypic, full pedigree, and genomic information for genetic evaluation of Holstein final score1

The first national single-step, full-information (phenotype, pedigree, and marker genotype) genetic evaluation was developed for final score of US Holsteins. Data included final scores recorded from 1955 to 2009 for 6,232,548 Holsteins cows. BovineSNP50 (Illumina, San Diego, CA) genotypes from the Cooperative Dairy DNA Repository (Beltsville, MD) were available for 6,508 bulls. Three analyses used a repeatability animal model as currently used for the national US evaluation. The first 2 analyses used final scores recorded up to 2004. The first analysis used only a pedigree-based relationship matrix. The second analysis used a relationship matrix based on both pedigree and genomic information (single-step approach). The third analysis used the complete data set and only the pedigree-based relationship matrix. The fourth analysis used predictions from the first analysis (final scores up to 2004 and only a pedigree-based relationship matrix) and prediction using a genomic based matrix to obtain genetic evaluation (multiple-step approach). Different allele frequencies were tested in construction of the genomic relationship matrix. Coefficients of determination between predictions of young bulls from parent average, single-step, and multiple-step approaches and their 2009 daughter deviations were 0.24, 0.37 to 0.41, and 0.40, respectively. The highest coefficient of determination for a single-step approach was observed when using a genomic relationship matrix with assumed allele frequencies of 0.5. Coefficients for regression of 2009 daughter deviations on parent-average, single-step, and multiple-step predictions were 0.76, 0.68 to 0.79, and 0.86, respectively, which indicated some inflation of predictions. The single-step regression coefficient could be increased up to 0.92 by scaling differences between the genomic and pedigree-based relationship matrices with little loss in accuracy of prediction. One complete evaluation took about 2h of computing time and 2.7 gigabytes of memory. Computing times for single-step analyses were slightly longer (2%) than for pedigree-based analysis. A national single-step genetic evaluation with the pedigree relationship matrix augmented with genomic information provided genomic predictions with accuracy and bias comparable to multiple-step procedures and could account for any population or data structure. Advantages of single-step evaluations should increase in the future when animals are pre-selected on genotypes.

A relationship matrix including full pedigree and genomic information.

Dense molecular markers are being used in genetic evaluation for parts of the population. This requires a two-step procedure where pseudo-data (for instance, daughter yield deviations) are computed from full records and pedigree data and later used for genomic evaluation. This results in bias and loss of information. One way to incorporate the genomic information into a full genetic evaluation is by modifying the numerator relationship matrix. A naive proposal is to substitute the relationships of genotyped animals with the genomic relationship matrix. However, this results in incoherencies because the genomic relationship matrix includes information on relationships among ancestors and descendants. In other words, using the pedigree-derived covariance between genotyped and ungenotyped individuals, with the pretense that genomic information does not exist, leads to inconsistencies. It is proposed to condition the genetic value of ungenotyped animals on the genetic value of genotyped animals via the selection index (e.g., pedigree information), and then use the genomic relationship matrix for the latter. This results in a joint distribution of genotyped and ungenotyped genetic values, with a pedigree-genomic relationship matrix H. In this matrix, genomic information is transmitted to the covariances among all ungenotyped individuals. The matrix is (semi)positive definite by construction, which is not the case for the naive approach. Numerical examples and alternative expressions are discussed. Matrix H is suitable for iteration on data algorithms that multiply a vector times a matrix, such as preconditioned conjugated gradients.

Computing procedures for genetic evaluation including phenotypic, full pedigree, and genomic information.

Currently, genomic evaluations use multiple-step procedures, which are prone to biases and errors. A single-step procedure may be applicable when genomic predictions can be obtained by modifying the numerator relationship matrix A to H = A + A(Delta), where A(Delta) includes deviations from expected relationships. However, the traditional mixed model equations require H(-1), which is usually difficult to obtain for large pedigrees. The computations with H are feasible when the mixed model equations are expressed in an alternate form that also applies for singular H and when those equations are solved by the conjugate gradient techniques. Then the only computations involving H are in the form of Aq or A(Delta)q, where q is a vector. The alternative equations have a nonsymmetric left-hand side. Computing A(Delta)q is inexpensive when the number of nonzeros in A(Delta) is small, and the product Aq can be calculated efficiently in linear time using an indirect algorithm. Generalizations to more complicated models are proposed. The data included 10.2 million final scores on 6.2 million Holsteins and were analyzed by a repeatability model. Comparisons involved the regular and the alternative equations. The model for the second case included simulated A(Delta). Solutions were obtained by the preconditioned conjugate gradient algorithm, which works only with symmetric matrices, and by the bi-conjugate gradient stabilized algorithm, which also works with nonsymmetric matrices. The convergence rate associated with the nonsymmetric solvers was slightly better than that with the symmetric solver for the original equations, although the time per round was twice as much for the nonsymmetric solvers. The convergence rate associated with the alternative equations ranged from 2 times lower without A(Delta) to 3 times lower for the largest simulated A(Delta). When the information attributable to genomics can be expressed as modifications to the numerator relationship matrix, the proposed methodology may allow the upgrading of an existing evaluation to incorporate the genomic information.

Different genomic relationship matrices for single-step analysis using phenotypic, pedigree and genomic information

BackgroundThe incorporation of genomic coefficients into the numerator relationship matrix allows estimation of breeding values using all phenotypic, pedigree and genomic information simultaneously. In such a single-step procedure, genomic and pedigree-based relationships have to be compatible. As there are many options to create genomic relationships, there is a question of which is optimal and what the effects of deviations from optimality are.MethodsData of litter size (total number born per litter) for 338,346 sows were analyzed. Illumina PorcineSNP60 BeadChip genotypes were available for 1,989. Analyses were carried out with the complete data set and with a subset of genotyped animals and three generations pedigree (5,090 animals). A single-trait animal model was used to estimate variance components and breeding values. Genomic relationship matrices were constructed using allele frequencies equal to 0.5 (G05), equal to the average minor allele frequency (GMF), or equal to observed frequencies (GOF). A genomic matrix considering random ascertainment of allele frequencies was also used (GOF*). A normalized matrix (GN) was obtained to have average diagonal coefficients equal to 1. The genomic matrices were combined with the numerator relationship matrix creating H matrices.ResultsIn G05 and GMF, both diagonal and off-diagonal elements were on average greater than the pedigree-based coefficients. In GOF and GOF*, the average diagonal elements were smaller than pedigree-based coefficients. The mean of off-diagonal coefficients was zero in GOF and GOF*. Choices of G with average diagonal coefficients different from 1 led to greater estimates of additive variance in the smaller data set. The correlation between EBV and genomic EBV (n = 1,989) were: 0.79 using G05, 0.79 using GMF, 0.78 using GOF, 0.79 using GOF*, and 0.78 using GN. Accuracies calculated by inversion increased with all genomic matrices. The accuracies of genomic-assisted EBV were inflated in all cases except when GN was used.ConclusionsParameter estimates may be biased if the genomic relationship coefficients are in a different scale than pedigree-based coefficients. A reasonable scaling may be obtained by using observed allele frequencies and re-scaling the genomic relationship matrix to obtain average diagonal elements of 1.

Bias in genomic predictions for populations under selection

Prediction of genetic merit or disease risk using genetic marker information is becoming a common practice for selection of livestock and plant species. For the successful application of genome-wide marker-assisted selection (GWMAS), genomic predictions should be accurate and unbiased. The effect of selection on bias and accuracy of genomic predictions was studied in two simulated animal populations under weak or strong selection and with several heritabilities. Prediction of genetic values was by best-linear unbiased prediction (BLUP) using data either from relatives summarized in pseudodata for genotyped individuals (multiple-step method) or using all available data jointly (single-step method). The single-step method combined genomic- and pedigree-based relationship matrices. Predictions by the multiple-step method were biased. Predictions by a single-step method were less biased and more accurate but under strong selection were less accurate. When genomic relationships were shifted by a constant, the single-step method was unbiased and the most accurate. The value of that constant, which adjusts for non-random selection of genotyped individuals, can be derived analytically.

Genome-wide association mapping including phenotypes from relatives without genotypes.

A common problem for genome-wide association analysis (GWAS) is lack of power for detection of quantitative trait loci (QTLs) and precision for fine mapping. Here, we present a statistical method, termed single-step GBLUP (ssGBLUP), which increases both power and precision without increasing genotyping costs by taking advantage of phenotypes from other related and unrelated subjects. The procedure achieves these goals by blending traditional pedigree relationships with those derived from genetic markers, and by conversion of estimated breeding values (EBVs) to marker effects and weights. Additionally, the application of mixed model approaches allow for both simple and complex analyses that involve multiple traits and confounding factors, such as environmental, epigenetic or maternal environmental effects. Efficiency of the method was examined using simulations with 15,800 subjects, of which 1500 were genotyped. Thirty QTLs were simulated across genome and assumed heritability was 0·5. Comparisons included ssGBLUP applied directly to phenotypes, BayesB and classical GWAS (CGWAS) with deregressed proofs. An average accuracy of prediction 0·89 was obtained by ssGBLUP after one iteration, which was 0·01 higher than by BayesB. Power and precision for GWAS applications were evaluated by the correlation between true QTL effects and the sum of m adjacent single nucleotide polymorphism (SNP) effects. The highest correlations were 0·82 and 0·74 for ssGBLUP and CGWAS with m=8, and 0·83 for BayesB with m=16. Standard deviations of the correlations across replicates were several times higher in BayesB than in ssGBLUP. The ssGBLUP method with marker weights is faster, more accurate and easier to implement for GWAS applications without computing pseudo-data.

Genome-wide marker-assisted selection combining all pedigree phenotypic information with genotypic data in one step: An example using broiler chickens

Data of broiler chickens for 2 pure lines across 3 generations were used for genomic evaluation. A complete population (full data set; FDS) consisted of 183,784 and 164,246 broilers for the 2 lines. The genotyped subsets (SUB) consisted of 3,284 and 3,098 broilers with 57,636 SNP. Genotyped animals were preselected based on more than 20 traits with different index applied to each line. Three traits were analyzed: BW at 6 wk (BW6), ultrasound measurement of breast meat (BM), and leg score (LS) coded 1 = no and 2 = yes for leg defect. Some phenotypes were missing for BM. The training population consisted of the first 2 generations including all animals in FDS or only genotyped animals in SUB. The validation data set contained only genotyped animals in the third generation. Genetic evaluations were performed using 3 approaches: 1) phenotypic BLUP, 2) extending BLUP methodologies to utilize pedigree and genomic information in a single step (ssGBLUP), and 3) Bayes A. Whereas BLUP and ssGBLUP utilized all phenotypic data, Bayes A could use only those of the genotyped subset. Heritabilities were 0.17 to 0.20 for BW6, 0.30 to 0.35 for BM, and 0.09 to 0.11 for LS. The average accuracies of the validation population with BLUP for BW6, BM, and LS were 0.46, 0.30, and <0 with SUB and 0.51, 0.34, and 0.28 with FDS. With ssGBLUP, those accuracies were 0.60, 0.34, and 0.06 with SUB and 0.61, 0.40, and 0.37 with FDS, respectively. With Bayes A, the accuracies were 0.60, 0.36, and 0.09 with SUB. With SUB, Bayes A and ssGBLUP had similar accuracies. For traits of high heritability, the accuracy of Bayes A/SUB and ssGBLUP/FDS were similar, and up to 50% better than BLUP/FDS. However, with low heritability, ssGBLUP/FDS was 4 to 6 times more accurate than Bayes A/SUB and 50% better than BLUP/FDS. An optimal genomic evaluation would be multi-trait and involve all traits and records on which selection is based.

Genetic components of heat stress for dairy cattle with multiple lactations

Data included 585,119 test-day records for milk, fat, and protein yields from the first, second, and third parities of 38,608 Holsteins in Georgia. Daily temperature-humidity indexes (THI) were available from public weather stations. Models included a repeatability test-day model with a random regression on a function of THI and a test-day random regression model using linear splines with knots at 5, 50, 200, and 305 d in milk and a function of THI. Random effects were additive genetic and permanent environmental in the repeatability model and additive genetic, permanent environmental, and herd year in the random regression model. Additionally, models included fixed effects for herd test day, calving age, milking frequency, and lactation stage. Phenotypic variance increased by 50 to 60% from the first to second parity for all yield traits with the repeatability model and by 12 to 15% from the second to third parity. General additive genetic variance increased by 25 to 35% from the first to second parity for all yield traits but decreased slightly from the second to third parity for milk and protein yields. Genetic variance for heat tolerance doubled from the first to second parity and increased by 20 to 100% from the second to third parity. Genetic correlations among general additive effects were lowest between the first and second parities (0.84 to 0.88) and were highest between the second and third parities (0.96 to 0.98). Genetic correlations among parities for the effect of heat tolerance ranged from 0.56 to 0.79. Genetic correlations between general and heat-tolerance effects across parities and yield traits ranged from -0.30 to -0.50. With the random regression model, genetic variance for heat tolerance for milk yield was approximately one-half that of the repeatability model. For milk yield, the most negative genetic correlation (approximately -0.45) between general and heat-tolerance effects was between 50 and 200 d in milk for the first parity and between 200 and 305 d in milk for the second and third parities. The genetic variance of heat tolerance increased substantially from the first to third parity. Genetic estimates of heat tolerance may be inflated with the repeatability model because of timing of lactations to avoid peak yield during hot seasons.

Effect of different genomic relationship matrices on accuracy and scale

Phenotypic data on BW and breast meat area were available on up to 287,614 broilers. A total of 4,113 birds were genotyped for 57,636 SNP. Data were analyzed by a single-step genomic BLUP (ssGBLUP), which accounts for all phenotypic, pedigree, and genomic information. The genomic relationship matrix (G) in ssGBLUP was constructed using either equal (0.5; GEq) or current (GC) allele frequencies, and with all SNP or with SNP with minor allele frequencies (MAF) below multiple thresholds (0.1, 0.2, 0.3, and 0.4) ignored. Additionally, a pedigree-based relationship matrix for genotyped birds (A(22)) was available. The matrices and their inverses were compared with regard to average diagonal (AvgD) and off-diagonal (AvgOff) elements. In A(22), AvgD was 1.004 and AvgOff was 0.014. In GEq, both averages decreased with the increasing thresholds for MAF, with AvgD decreasing from 1.373 to 1.020 and AvgOff decreasing from 0.722 to 0.025. In GC, AvgD was approximately 1.01 and AvgOff was 0 for all MAF. For inverses of the relationship matrices, all AvgOff were close to 0; AvgD was 2.375 in A(22), varied from 11.563 to 12.943 for GEq, and increased from 8.675 to 12.859 for GC as the threshold for MAF increased. Predictive ability with all GEq and GC was similar except that at MAF = 0.4, they declined by 0.01 for BW and improved by 0.01 for breast meat area. Compared with BLUP, EBV in the ssGBLUP were, on average, increased by up to 1 additive SD greater with GEq and decreased by 2 additive SD less with GC. Genotyped animals were biased upward with GEq and downward with GC. The biases and differences in EBV could be controlled by adding a constant to GC; they were eliminated with a constant of 0.014, which corresponds to AvgOff in A(22). Unbiased evaluation in the ssGBLUP may be obtained with GC scaled to be compatible with A(22). The reduction of SNP with small MAF has a small effect on the real accuracy, but it may falsely increase the estimated accuracies by inversion.

Using recursion to compute the inverse of the genomic relationship matrix

Computing the inverse of the genomic relationship matrix using recursion was investigated. A traditional algorithm to invert the numerator relationship matrix is based on the observation that the conditional expectation for an additive effect of 1 animal given the effects of all other animals depends on the effects of its sire and dam only, each with a coefficient of 0.5. With genomic relationships, such an expectation depends on all other genotyped animals, and the coefficients do not have any set value. For each animal, the coefficients plus the conditional variance can be called a genomic recursion. If such recursions are known, the mixed model equations can be solved without explicitly creating the inverse of the genomic relationship matrix. Several algorithms were developed to create genomic recursions. In an algorithm with sequential updates, genomic recursions are created animal by animal. That algorithm can also be used to update a known inverse of a genomic relationship matrix for additional genotypes. In an algorithm with forward updates, a newly computed recursion is immediately applied to update recursions for remaining animals. The computing costs for both algorithms depend on the sparsity pattern of the genomic recursions, but are lower or equal than for regular inversion. An algorithm for proven and young animals assumes that the genomic recursions for young animals contain coefficients only for proven animals. Such an algorithm generates exact genomic EBV in genomic BLUP and is an approximation in single-step genomic BLUP. That algorithm has a cubic cost for the number of proven animals and a linear cost for the number of young animals. The genomic recursions can provide new insight into genomic evaluation and possibly reduce costs of genetic predictions with extremely large numbers of genotypes.