Hot topic: performance of bovine high-density genotyping platforms in Holsteins and Jerseys.

Two high-density single nucleotide polymorphism (SNP) genotyping arrays have recently become available for bovine genomic analyses, the Illumina High-Density Bovine BeadChip Array (777,962 SNP) and the Affymetrix Axiom Genome-Wide BOS 1 Array (648,874 SNP). These products each have unique design and chemistry attributes, and the extent of marker overlap and their potential utility for quantitative trait loci fine mapping, detection of copy number variation, and multibreed genomic selection are of significant interest to the cattle community. This is the first study to compare the performance of these 2 arrays. Deoxyribonucleic acid samples from 16 dairy cattle (10 Holstein, 6 Jersey) were used for the comparison. An independent set of DNA samples taken from 46 Jersey cattle and 18 Holstein cattle were used to ascertain the amount of SNP variation accounted by the 16 experimental samples. Data were analyzed with SVS7 software (Golden Helix Inc., Bozeman, MT) to remove SNP having a call rate less than 90%, and linkage disequilibrium pruning was used to remove linked SNP (r² ≥ 0.9). Maximum, average, and median gaps were calculated for each analysis based on genomic position of SNP on the bovine UMD3.1 genome assembly. All samples were successfully genotyped (≥ 98% SNP genotyped) with both platforms. The average number of genotyped SNP in the Illumina platform was 775,681 and 637,249 for the Affymetrix platform. Based on genomic position, a total of 107,896 SNP were shared between the 2 platforms; however, based on genotype concordance, only 96,031 SNP had complete concordance at these loci. Both Affymetrix BOS 1 and Illumina BovineHD genotyping platforms are well designed and provide high-quality genotypes and similar coverage of informative SNP. Despite fewer total SNP on BOS 1, 19% more SNP remained after linkage disequilibrium pruning, resulting in a smaller gap size (5.2 vs. 6.9 kb) in Holstein and Jersey samples relative to BovineHD. However, only 224,115 Illumina and 241,038 Affymetrix SNP remained following removal of SNP with a minor allele frequency of zero in Holstein and Jersey samples, resulting in an average gap size of 11,887 bp and 11,018 bp, respectively. Combining the 354,348 informative (r² ≥ 0.9), polymorphic (minor allele frequency ≥ 0), unique SNP data from both platforms decreased the average gap size to 7,560 bp. Genome-wide copy number variant analyses were performed using intensity files from both platforms. The BovineHD platform provided an advantage to the copy number variant data compared with the BOS 1 because of the larger number of SNP, higher intensity signals, and lower background effects. The combined use of both platforms significantly improved coverage over either platform alone and decreased the gap size between SNP, providing a valuable tool for fine mapping quantitative trait loci and multibreed animal evaluation.

Assessment of inbreeding depression in a Guzerat dairy herd: effects of individual increase in inbreeding coefficients on production and reproduction.

Influences of inbreeding on daily milk yield (DMY), age at first calving (AFC), and calving intervals (CI) were determined on a highly inbred zebu dairy subpopulation of the Guzerat breed. Variance components were estimated using animal models in single-trait analyses. Two approaches were employed to estimate inbreeding depression: using individual increase in inbreeding coefficients or using inbreeding coefficients as possible covariates included in the statistical models. The pedigree file included 9,915 animals, of which 9,055 were inbred, with an average inbreeding coefficient of 15.2%. The maximum inbreeding coefficient observed was 49.45%, and the average inbreeding for the females still in the herd during the analysis was 26.42%. Heritability estimates were 0.27 for DMY and 0.38 for AFC. The genetic variance ratio estimated with the random regression model for CI ranged around 0.10. Increased inbreeding caused poorer performance in DMY, AFC, and CI. However, some of the cows with the highest milk yield were among the highly inbred animals in this subpopulation. Individual increase in inbreeding used as a covariate in the statistical models accounted for inbreeding depression while avoiding overestimation that may result when fitting inbreeding coefficients.

A preorganized active site in the crystal structure of the Tetrahymena ribozyme.

Group I introns possess a single active site that catalyzes the two sequential reactions of self-splicing. An RNA comprising the two domains of the Tetrahymena thermophila group I intron catalytic core retains activity, and the 5.0 angstrom crystal structure of this 247-nucleotide ribozyme is now described. Close packing of the two domains forms a shallow cleft capable of binding the short helix that contains the 5' splice site. The helix that provides the binding site for the guanosine substrate deviates significantly from A-form geometry, providing a tight binding pocket. The binding pockets for both the 5' splice site helix and guanosine are formed and oriented in the absence of these substrates. Thus, this large ribozyme is largely preorganized for catalysis, much like a globular protein enzyme.

Crystal structure of a group I ribozyme domain: principles of RNA packing.

Group I self-splicing introns catalyze their own excision from precursor RNAs by way of a two-step transesterification reaction. The catalytic core of these ribozymes is formed by two structural domains. The 2.8-angstrom crystal structure of one of these, the P4-P6 domain of the Tetrahymena thermophila intron, is described. In the 160-nucleotide domain, a sharp bend allows stacked helices of the conserved core to pack alongside helices of an adjacent region. Two specific long-range interactions clamp the two halves of the domain together: a two-Mg2+-coordinated adenosine-rich corkscrew plugs into the minor groove of a helix, and a GAAA hairpin loop binds to a conserved 11-nucleotide internal loop. Metal- and ribose-mediated backbone contacts further stabilize the close side-by-side helical packing. The structure indicates the extent of RNA packing required for the function of large ribozymes, the spliceosome, and the ribosome.

RNA tertiary structure mediation by adenosine platforms.

The crystal structure of a group I intron domain reveals an unexpected motif that mediates both intra- and intermolecular interactions. At three separate locations in the 160-nucleotide domain, adjacent adenosines in the sequence lie side-by-side and form a pseudo-base pair within a helix. This adenosine platform opens the minor groove for base stacking or base pairing with nucleotides from a noncontiguous RNA strand. The platform motif has a distinctive chemical modification signature that may enable its detection in other structured RNAs. The ability of this motif to facilitate higher order folding provides one explanation for the abundance of adenosine residues in internal loops of many RNAs.

Stepwise conversion of a mesophilic to a thermophilic ribozyme.

A fundamental question in RNA folding is the mechanism of thermodynamic stability. We investigated the equilibrium folding of a series of sequence variants in which one to three motifs of a 255-nucleotide mesophilic ribozyme were substituted with the corresponding motifs from its thermophilic homologue. Substitution of three crucial motifs individually or in groups results in a continual increase in the stability and folding cooperativity in a stepwise fashion. We find an unexpected relationship between stability and folding cooperativity. Without changing the folding cooperativity, RNAs having a similar native structure can only achieve moderate change in stability and likewise, without changing stability, RNAs having a similar native structure can only achieve moderate change in folding cooperativity. This intricate relationship must be included in the predictions of tertiary RNA stability.

Ribosomal proteins S5 and L6: high-resolution crystal structures and roles in protein synthesis and antibiotic resistance.

Antibiotic resistance is rapidly becoming a major medical problem. Many antibiotics are directed against bacterial ribosomes, and mutations within both the RNA and protein components can render them ineffective. It is well known that the majority of these antibiotics act by binding to the ribosomal RNA, and it is of interest to understand how mutations in the ribosomal proteins can produce resistance. Translational accuracy is one important target of antibiotics, and a number of ribosomal protein mutations in Escherichia coli are known to modulate the proofreading mechanism of the ribosome. Here we describe the high-resolution structures of two such ribosomal proteins and characterize these mutations. The S5 protein, from the small ribosomal unit, is associated with two types of mutations: those that reduce translational fidelity and others that produce resistance to the antibiotic spectinomycin. The L6 protein, from the large subunit, has mutations that cause resistance to several aminoglycoside antibiotics, notably gentamicin. In both proteins, the mutations occur within their putative RNA-binding sites. The L6 mutations are particularly drastic because they result in large deletions of an RNA-binding region. These results support the hypothesis that the mutations create local distortions of the catalytic RNA component.When combined with a variety of structural and biochemical data, these mutations also become important probes of the architecture and function of the translational machinery. We propose that the C-terminal half of S5, which contains the accuracy mutations, organizes RNA structures associated with the decoding region, and the N-terminal half, which contains the spectinomycin-resistance mutations, directly interacts with an RNA helix that binds this antibiotic. As regards L6, we suggest that the mutations indirectly affect proofreading by locally distorting the EF-Tu.GTP.aminoacyl tRNA binding site on the large subunit.

Crystals by design: a strategy for crystallization of a ribozyme derived from the Tetrahymena group I intron.

Recently, the 2.8 A crystal structure of one domain of the self-splicing Tetrahymena group I intron was reported. Although it revealed much about RNA tertiary interactions, it contained only half of the active site. We have now designed a series of larger molecules that contain about 70% of the intron and all of the catalytic core. These RNAs were efficient in cleavage of a substrate RNA, consisting of the approximately 100 nucleotides from the 5' end of the intron, at a site corresponding to the 5' splice site. A sparse matrix was designed specifically for large RNAs and used to screen for preliminary crystallization conditions. Of the six RNAs initially tested, five were crystallized in this initial trial. Two of these crystals were further examined. The first diffracted X-rays to only approximately 16 A resolution, even when the crystal were very large. The second diffracted as high as 3.5 A, but the crystals were twinned and therefore unusable for structural studies. Site-specific mutagenesis was performed on the latter RNA to disrupt interactions that might have been responsible for the twinning. One of these mutant RNAs produced large, single, diffraction-quality crystals. The crystals belong to the tetragonal space group P42212 and have large unit cell dimensions, a=b=178 A and c=199 A. Thus, by variation of both sequence elements and crystallization conditions, crystals of a 247 nucleotide catalytic RNA were obtained.

Technical note: changes in genetic predictions between subsequent evaluations.

A procedure was developed to compute the proportion (P) of future genetic predictions that would be within 1 SE of previous predictions. The procedure is based on the Central Limit Theorem. Whatever the distribution function, provided only that it has a finite variance, the sample mean will have approximately the normal distribution for large samples. The proportion of new individual genetic predictions being within 1 SE of their previous evaluation is expressed as a function of the change in accuracy (ACC) between the previous and subsequent evaluations. If little additional information is made available since the previous evaluation, the increase in ACC will be almost negligible. As anticipated the vast majority of genetic predictions will be within 1 SE of their previous evaluation. The proportion determined from the results of the analysis can be compared to P. An additional appealing feature of the procedure presented is the ease of implementation with most computer softwares. Finally, application to both simulated and field data is presented.

Approximate confidence intervals for heritability from method R estimates.

Method R estimates of heritability (h2) and associated confidence intervals (CI) were obtained from simulated data using a single trait, direct effects, full animal model, with 50% subsampling. Five hundred data sets were simulated for each of five levels of h2 (.10, .20, .30, .40, and .50) and two types of pedigree structure (random pedigree structure [N = 2,000] that varied over simulations, or the pedigree structure from a real data set [N = 2,644] that was constant for all simulations). The first 10, 20, and all 50 h2 estimates were used to obtain 80, 90, 95, and 99% CI for each data set. The variance of h2 estimates within data sets approximated the sampling variance of the h2 estimates. The Box-Cox transformation was used to normalize the distribution of estimates from each data set. Confidence intervals were computed on the transformed scale as CI = mu +/- (T x sigma), where mu and sigma = the mean and SD of the N transformed h2 estimates, respectively, and T = the critical value from the T distribution for a 1-alpha CI, with df = N-1. Upper and lower CI bounds were converted back to the original scale by reversing the transformation. The percentages of CI containing the true h2 value, pooled across all levels of h2, types of pedigree, and number of estimates used to obtain CI, for 80, 90, 95, and 99% CI were 81.14, 90.96, 95.27, and 98.76%, respectively. These results suggested that Method R h2 estimates can be used to obtain reliable CI.

Technical note: detection of bias in genetic predictions.

The theoretical development of a procedure to detect bias in genetic predictions is presented. The procedure is based on the expectation of three statistics. These statistics detect bias by identifying systematic, unexpected change in subsequent analyses. Expectations of the following statistics were obtained: linear correlation coefficient between subsequent predictions, linear regression of recent (more accurate) on previous (less accurate) genetic prediction, and variance of the genetic prediction difference (recent minus previous genetic prediction). Deviations from these expectations can be used to indicate bias. The covariance between subsequent BLUP of genetic value is shown to equal the variance of the early estimate, implying that the expected value of the regression of recent on previous genetic prediction equals 1 regardless of the distribution of the observations and predictions. Also, the expected value of the linear correlation coefficient between subsequent genetic predictions equals the square root of the ratio of the means of the square of accuracy values. The expected value of the variance of the genetic prediction difference was shown to be equal to the difference between prediction error variances.

A performance programmed method for computing inbreeding coefficients from large data sets for use in mixed-model analyses.

Coefficients of inbreeding are commonly used in mixed-model methods for forming inverses of Wright's numerator relationship matrix and transformation matrices used in variance component estimation and national cattle evaluation. Computation of exact coefficients of inbreeding from very large data sets has been believed to be too expensive or too difficult a task to perform. Approximate methods have been used instead. The effects of using approximation methods for inbred data that appear in national cattle data sets are demonstrated. An algorithm is given for the computation of inbreeding coefficients for large data sets. The algorithm feasibly computes inbreeding coefficients for large data sets even on small computing architectures.

Method R variance components procedure: application on the simple breeding value model.

An algorithm for estimating variance components (Method R) based on the linear regression coefficient (R) of recent (more accurate) on previous (less accurate) individual genetic predictions is presented. The previous prediction is obtained by analyzing a subsample of the whole data set. First raw moment of R equals 1 regardless of the distribution of observations and predictions. A condition such as the use of inappropriate variance components ratio (VC) can cause this regression to deviate from its expectation. If the computed R (Rc) is greater than 1, then VC ratio has been underestimated, and if Rc is less than 1, then VC ratio has been overestimated. Several iterations are performed, changing the VC ratio at each iteration, until Rc approximately equal to 1. When an Rc is obtained that is acceptably close to 1 (precision is reached), then the appropriate VC has been used. Method R does not require computation of the inverse of the coefficient matrix and has desirable properties of convergence, precision, and computing feasibility. Additional sampling variance in the estimate of VC is expected due to the requirement of taking a subsample of the entire data set to obtain the lower accuracy predictions. This sampling variance is shown to be small for simulated datasets of size n = 10,000 with no selection.

Within-herd genetic analyses of stayability of beef females.

Nonlinear mixed-model procedures for analysis of binary data were used to estimate heritability (h2), predict individual genetic merit, and determine genetic and environmental trends for four measures of stayability of beef females. Traits considered were probabilities of a female having 2 [S(2/1)], 5 [S(5/1)], 8 [S(8/1)] and 11 [S(11/1)] calves, given that she calved once. Colorado State University Beef Improvement Center (BIC) and Beckton Stock Farm (BSF) provided data for the analyses. Heritability was estimated using animal model marginal maximum likelihood (AM MML), sire model marginal maximum likelihood (SM MML), and animal model Method R (AM MR). Individual genetic merit was predicted using single-trait animal models with each h2 estimate. Birth year was treated as fixed in all analyses. Only AM MML yielded h2 estimates for all traits in both herds. The AM MML h2 estimates for S(2/1), S(5/1), S(8/1), and S(11/1) were .09, .11, .07, and .20, respectively, for BSF data and .02, .14, .09, and .07, respectively, for BIC data. Differing h2 estimates did not substantially influence rank of individual predictions. Genetic trends in stayability were positive in both herds, although birth year solutions indicated variable or negative environmental trends. Genetic improvement of stayability may be accelerated by incorporating predictions of genetic merit for stayability in selection criteria. S(5/1) may be the most useful trait for consideration in national cattle evaluations.

The effect of selective reporting on estimates of weaning weight parameters in beef cattle.

The effect of selective reporting on estimates of weaning weight parameters in beef cattle was evaluated by comparing REML estimates from unaltered and altered simulated data. Selective reporting reduced estimates of weaning weight direct (WWD), maternal milk (MAT), and error variances. However, heritability estimates were not greatly affected because the reductions in variance estimates were relatively proportionate. When the true value for the direct-maternal (DM) correlation was zero or negative, selective reporting caused estimates of DM to be less positive or more negative in 50 of 62 comparisons, with an average change of -.136. When the true value for DM was positive, selective reporting increased the positive magnitude of DM estimates in 12 of 20 comparisons, with an average change of +.040. In BLUP of unaltered data with a true DM value of -.09, using a -.28 and a zero DM correlation reduced the correlation of MAT EPD with true values .065 and .041, respectively. These results suggest that the reliability of parameter estimates (and BLUPs) would be improved by estimating parameters from representative subsets of data free of reporting bias.

Additive genetic groups for animals evaluated in more than one breed association national cattle evaluation.

Additive genetic groups were included in the 1993 Red Angus Association of America national cattle evaluation for phantom parents of individuals who were registered with the American Angus Association (AAA). Genetic groups were formed for each component in two multiple-trait evaluations in which all animal effects were fit. Additive direct effects were included for birth weight, weaning weight (WW), and milk (MILK). In a second analysis the additive direct effect of 160-d postweaning gain was analyzed with WW and MILK. Of the 387,665 animals, 50,838 had at least one phantom parent assigned to one of five genetic groups fit as fixed effects for each additive component. Of these 50,838 animals, 1,324 were identified as registered with the AAA. An average of 906 animals per component had an AAA EPD available. Animals with a known AAA EPD were designated into one of three groups of equal numbers based on AAA EPD for each component (1 = low, 2 = medium, 3 = high). Animals in the fourth genetic group were those registered with the AAA but with no EPD available for the component. The fifth genetic group included all other animals with phantom parents. Grouping on AAA EPD allowed for EPD on animals out of parent(s) registered with the AAA to be more closely aligned to the AAA EPD because they were regressed from the group solution instead of zero. Grouping based on EPD from another NCE should be considered in the production of multibreed EPD.

Genetic parameter estimates of ultrasound-measured longissimus muscle area and 12th rib fat thickness in Brangus cattle.

Data from 2,101 Brangus calves born from 1986 to 1990 were analyzed with a REML procedure using a derivative-free algorithm in a mixed linear animal model to obtain variance component estimates of ultrasound-measured longissimus muscle area and fat thickness. Direct additive heritabilities (ha2) of .39 and .40 were obtained for age-constant weaning and yearling longissimus muscle area (WLMA and YLMA, respectively), with a genetic correlation (rg) of .66 between them. The rg of YLMA with birth weight (BWT), weaning weight (WWT), postweaning gain (PWG), yearling weight (YWT), frame score (FS), and scrotal circumference (SC) were .17, .29, .43, .38, .01, and .19, respectively. The ha2 of age-constant yearling 12th rib fat thickness (FAT) was .14, and cattle averaged .44 cm (SD = .19). Positive rg were obtained between FAT and WLMA (.19) and YLMA (.12). Negative rg of FAT with WWT, YWT, and SC were -.17, -.53, and -.33, respectively. Positive rg were obtained between FAT and BWT (.52), PWG (.44), and FS (.14). Maternal heritabilities (hm2) of WLMA, YLMA, and FAT were .01, .01, and .10, respectively. Weight-constant WLMA, YLMA, and FAT ha2 were .36, .39, and .11, respectively. Selection based on either age-constant YLMA or FAT could potentially result in 1.06 cm2 or .005 cm change per year, respectively, which would be slightly greater than change from selection based on weight-constant YLMA or FAT. Selection based on WLMA or YLMA should be effective, and changes in these traits, growth, and SC should be possible in tandem.(ABSTRACT TRUNCATED AT 250 WORDS)

Genetic analysis of absolute growth measurements, relative growth rate and restricted selection indices in red Angus cattle.

Performance records on 41,184 Red Angus cattle were analyzed and estimates of parameters calculated for absolute growth rate, relative growth rate and restricted selection indices. Heritability estimates for birth weight, 205-d weight, 365-d weight and postweaning gain were .46 +/- .02, .39 +/- .02, .40 +/- .02 and .36 +/- .02, respectively. Heritability estimates for preweaning, postweaning and postnatal relative growth rates were identical (.33 +/- .02). Heritability estimates for restricted selection indices were .31 +/- .02, .33 +/- .02 and .31 +/- .02 for weaning index, yearling index and postweaning index, respectively. The genetic correlation between preweaning and postweaning absolute growth rate was .15. The genetic correlation between consecutive measurements of relative growth rate (RGR) was -.33. Genetic correlations of birth weight with preweaning RGR and postnatal RGR were -.68 and -.71, respectively. Correlations among measures of relative growth rate using simulated data were similar to correlations of actual data, indicating that these relationships are the result of numerator/denominator relationships and not biological causes. The genetic correlation between weaning and postweaning indices was near zero. Small genetic coefficients of variation for preweaning and postnatal relative growth rates indicate further problems with the expression of growth in this manner. Restricted selection indices exhibited much larger genetic coefficients of variation than measurements of RGR. Genetic standard deviations were 7.8%, 7.2% and 13.7% of the means for weaning, yearling and postweaning indices, respectively.

Estimation of variance components and prediction of breeding values for scrotal circumference and weaning weight in Limousin cattle.

Variance and covariance components were estimated for yearling scrotal circumference and weaning weight from Limousin field data. Records of 8.226 bulls were used to evaluate 584 sires and 653 maternal grandsires. Data included all herdbook records of bulls with a recorded scrotal circumference and their weaning contemporaries. Analyses were performed by restricted maximum likelihood techniques employing the expected maximization algorithm and fitting both single- and two-trait models. Scrotal circumference was first fitted in a single-trait, sire model to obtain starting values for variances for a later analysis. Likewise, weaning weight was fitted in a single-trait, sire-maternal grandsire model to obtain priors for (co)variances for a later analysis. Scrotal circumference and weaning weight were then fitted together in a two-trait model to estimate variance components. Estimates of variance components were calculated by equating (co)variances obtained from the models to their expectations. Estimates of heritability of scrotal circumference, direct weaning weight, and maternal weaning weight were .46, .25, and .19, respectively. Estimates of genetic correlations between yearling scrotal circumference and direct weaning weight, scrotal circumference and maternal weaning weight, and direct weaning weight and maternal weaning weight were .14, -.22, and -.44, respectively. The estimate of the environmental correlation between scrotal circumference and weaning weight was .61. Genetic parameters obtained were then used in two-trait, reduced animal mixed-model equations for a maternally influenced trait to predict breeding values for animals.

Additive genetic parameter estimates for heifer pregnancy and subsequent reproduction in Angus females.

A primary objective of this study was to determine whether the binary traits heifer pregnancy (HP) and subsequent rebreeding (SR) were heritable in an experimental population of Angus cattle. A second objective was to determine the nature of the additive genetic relationships among HP, SR, and stayability (S(5/1)) in the same population. Heifer pregnancy was defined as the observation of a heifer conceiving and remaining pregnant to palpation at 120 d, given exposure during the breeding season. Subsequent rebreeding was defined as the observation of a 2-yr-old conceiving and remaining pregnant to palpation at 105 d, given pregnancy as a yearling and exposure during the breeding season. Stayability was defined as the probability of a female having at least five calves, given she becomes a dam as a 2 yr old. Data were analyzed using a maximum a posteriori probit threshold model to predict breeding values on the liability scale and Method R procedures to estimate variance components in the determination of heritability (h2). Additive genetic groups were used in determining the additive genetic relationships among these fertility traits. Additive genetic groups were formed on one trait's breeding values and used in the prediction of another trait's breeding values. Analyses yielded h2 estimates that were out of the parameter space 8.5 and 46.3% for HP and SR, respectively, and 5.9% for the reestimation of S(5/1). The majority of point estimates outside the parameter space for SR converged toward 0, whereas those for HP and S(5/1) primarily converged toward 1. From the subsamples producing h2 estimates within the parameter space, average h2 for HP, SR, and S(5/1) were .21, .19, and .15, with standard deviations of .12, .14, and .08, respectively. The estimates of h2 indicate that HP and S(5/1) were heritable and should respond favorably to selection; however, SR did not appear heritable due to the large number of subsamples producing h2 estimates out of the parameter space. Fixed effect estimates for age of dam were significant for HP. From the analyses using additive genetic groups, the relationship among HP and S(5/1) appeared to be nonlinear. This potential nonlinear relationship seen between HP and S(5/1) indicates that selection for improved female fertility would be most effective by having predictions on both traits.