Roberto Steri

Using eigenvalues as variance priors in the prediction of genomic breeding values by principal component analysis

Genome-wide selection aims to predict genetic merit of individuals by estimating the effect of chromosome segments on phenotypes using dense single nucleotide polymorphism (SNP) marker maps. In the present paper, principal component analysis was used to reduce the number of predictors in the estimation of genomic breeding values for a simulated population. Principal component extraction was carried out either using all markers available or separately for each chromosome. Priors of predictor variance were based on their contribution to the total SNP correlation structure. The principal component approach yielded the same accuracy of predicted genomic breeding values obtained with the regression using SNP genotypes directly, with a reduction in the number of predictors of about 96% and computation time of 99%. Although these accuracies are lower than those currently achieved with Bayesian methods, at least for simulated data, the improved calculation speed together with the possibility of extracting principal components directly on individual chromosomes may represent an interesting option for predicting genomic breeding values in real data with a large number of SNP. The use of phenotypes as dependent variable instead of conventional breeding values resulted in more reliable estimates, thus supporting the current strategies adopted in research programs of genomic selection in livestock.

The mathematical description of lactation curves in dairy cattle

This review gives an overview of the mathematical modelling of lactation curves in dairy cattle. Over the last ninety years, the development of this field of study has followed the main requirements of the dairy cattle industry. Non-linear parametric functions have represented the preferred tools for modelling average curves of homogeneous groups of animals, with the main aim of predicting yields for management purposes. The increased availability of records per individual lactations and the genetic evaluation based on test day records has shifted the interest of modellers towards more flexible and general linear functions, as polynomials or splines. Thus the main interest of modelling is no longer the reconstruction of the general pattern of the phenomenon but the fitting of individual deviations from an average curve. Other specific approaches based on the modelling of the correlation structure of test day records within lactation, such as mixed linear models or principal component analysis, have been used to test the statistical significance of fixed effects in dairy experiments or to create new variables expressing main lactation curve traits. The adequacy of a model is not an absolute requisite, because it has to be assessed according to the specific purpose it is used for. Occurrence of extended lactations and of new productive and functional traits to be described and the increase of records coming from automatic milking systems likely will represent some of the future challenges for the mathematical modelling of the lactation curve in dairy cattle.

Use of the canonical discriminant analysis to select SNP markers for bovine breed assignment and traceability purposes

Several market research studies have shown that consumers are primarily concerned with the provenance of the food they eat. Among the available identification methods, only DNA-based techniques appear able to completely prevent frauds. In this study, a new method to discriminate among different bovine breeds and assign new individuals to groups was developed. Bulls of three cattle breeds farmed in Italy - Holstein, Brown, and Simmental - were genotyped using the 50K SNP Illumina BeadChip. Multivariate canonical discriminant analysis was used to discriminate among breeds, and discriminant analysis (DA) was used to assign new observations. This method was able to completely identify the three groups at chromosome level. Moreover, a genome-wide analysis developed using 340 linearly independent SNPs yielded a significant separation among groups. Using the reduced set of markers, the DA was able to assign 30 independent individuals to the proper breed. Finally, a set of 48 high discriminant SNPs was selected and used to develop a new run of the analysis. Again, the procedure was able to significantly identify the three breeds and to correctly assign new observations. These results suggest that an assay with the selected 48 SNP could be used to routinely track monobreed products.

An association analysis between OXT genotype and milk yield and flow in Italian Mediterranean river buffalo

The aim of this study was to evaluate possible associations between three SNPs at the oxytocin locus (AM234538: g.28C>T; g.204A>G and g.1627G>T) and two productive traits, milk yield and milkability, in Italian Mediterranean river buffaloes. Effects of parity, calving season and month of production were also evaluated. A total of 41 980 test-day records belonging to 219 lactations of 163 buffalo cows were investigated. The allele call rate was 98·8% and the major allele frequency for all the investigated loci was 0·76. The OXT genotype was significantly associated with milk yield (P=0·029). The TT genotype showed an average daily milk yield approximately 1·7 kg higher than GT buffaloes. Such a difference represents about 23% more milk/d. A large dominance effect (-1·17±0·43 kg) was estimated, whereas the contribution of OXT genotype (r(2)(OXT)) to the total phenotypic variance in milk yield was equal to 0·06. The TT genotype showed higher values also for the milk flow, even though the estimated difference did not reach a level of statistical significance (P=0·07). Such an association, among the first reported for the oxytocin locus in ruminants, should be tested on a population scale and possible effects on milk composition traits should be evaluated in order to supply useful indications for the application of marker-assisted selection programmes in river buffaloes.

Pre-selection of most significant SNPS for the estimation of genomic breeding values

The availability of a large amount of SNP markers throughout the genome of different livestock species offers the opportunity to estimate genomic breeding values (GEBVs). However, the estimation of many effects in a data set of limited size represent a severe statistical problem. A pre-selection of SNPS based on single regression may provide a reasonable compromise between accuracy of results, number of independent variables to be considered and computing requirements.A total of 595 and 618 SNPS were pre-selected using a simple linear regression for each SNP, based on phenotypes or polygenic EBVs, respectively, with an average distance of 9–10 cM between them. Chromosome four had the largest frequency of selected SNPS. Average correlations between GEBVs and TBVs were about 0.82 and 0.73 for the TRAINING generations when phenotypes or polygenic EBVs were considered as dependent variable, whereas they tend to decrease to 0.66 and 0.54 for the PREDICTION generations. The pre-selection of SNPs using the phenotypes as dependent variable together with a BLUP estimation of marker genotype effects using a variance contribution of each marker equal to σ2a/nsnps resulted in a remarkable accuracy of GEBV estimation (0.77) in the PREDICTION generations.

A single nucleotide polymorphism in the promoter region of river buffalo stearoyl CoA desaturase gene (SCD) is associated with milk yield.

An association study between the milk yield trait and the stearoyl-CoA desaturase (SCD) polymorphism (g.133A > C) in Italian Mediterranean river buffalo was carried out. A full characterization of the river buffalo SCD promoter region was presented. Genotyping information was provided and a quick method for allelic discrimination was developed. The frequency of the C allele was 0·16. Test-day (TD) records (43 510) of milk production belonging to 226 lactations of 169 buffalo cows were analysed with a mixed linear model in order to estimate the effect of g.133A > C genotype, as well as the effect of parity and calving season. The SCD genotype was significantly associated with milk yield (P = 0·02). The genotype AC showed an over-dominance effect with an average daily milk yield approximately 2 kg/d higher than CC buffaloes. Such a difference represents about 28% more milk/d. The effect of the genotype was constant across lactation stages. The contribution of SCD genotype (r(2)SCD) to the total phenotypic variance in milk yield was equal to 0·12. This report is among the first indications of genetic association between a trait of economic importance in river buffalo. Although such results need to be confirmed with large-scale studies in the same and other buffalo populations, they might offer useful indications for the application of MAS programmes in river buffalo and in the future they might be of great economic interest for the river buffalo dairy industry.

Analysis of lactation shapes in extended lactations.

In order to describe the temporal evolution of milk yield (MY) and composition in extended lactations, 21 658 lactations of Italian Holstein cows were analyzed. Six empirical mathematical models currently used to fit 305 standard lactations (Wood, Wilmink, Legendre, Ali and Schaeffer, quadratic and cubic splines) and one function developed specifically for extended lactations (a modification of the Dijkstra model) were tested to identify a suitable function for describing patterns until 1000 days in milk (DIM). Comparison was performed on individual patterns and on average curves grouped according to parity (primiparous and multiparous) and lactation length (standard ≤305 days, and extended from 600 to 1000 days). For average patterns, polynomial models showed better fitting performances when compared with the three or four parameters models. However, LEG and spline regression, showed poor prediction ability at the extremes of the lactation trajectory. The Ali and Schaeffer polynomial and Dijkstra function were effective in modelling average curves for MY and protein percentage, whereas a reduced fitting ability was observed for fat percentage and somatic cell score. When individual patterns were fitted, polynomial models outperformed nonlinear functions. No detectable differences were observed between standard and extended patterns in the initial phase of lactation, with similar values of peak production and time at peak. A considerable difference in persistency was observed between 200 and 305 DIM. Such a difference resulted in an estimated difference between standard and extended cycle of about 7 and 9 kg/day for daily yield at 305 DIM and of 463 and 677 kg of cumulated milk production at 305 DIM for the first- and second-parity groups, respectively. For first and later lactation animals, peak yield estimates were nearly 31 and 38 kg, respectively, and occurred at around 65 and 40 days. The asymptotic level of production was around 9 kg for multiparous cows, whereas the estimate was negative for first parity.

Modelling extended lactation curves for milk production traits in Italian Holsteins

References Test day records of milk production traits (milk yield, fat and protein percentage, and somatic cell score) of 45,132 Italian Holstein cows were analyzed with seven mathematical models in order to assess the main features of lactations of different length. Lactations curves were grouped according to parity (1, 2, and 3) and lactation length (1<350d; 2=from 351 to 450d; 3=from 451 to 650d; 4=651 to 1000d). Models with a larger number of parameters showed better fitting performances for all classes of length for milk yield, whereas poor fitting was observed for fat and protein percentages and SCS in the 651-1000d class. In lactation with length>650d, peak yield was about 31, 37, and 39 kg for first, second, and third parity respectively; peak was predicted at around 60 and 40 days for younger and older animals respectively. The asymptotic level of production was below 10 kg.

Mediterranean river buffalo CSN1S1 gene: search for polymorphisms and association studies

The aim of the present study was to investigate the variability at CSN1S1 locus of the Italian Mediterranean river buffalo and to study possible allele effects on milk yield and its composition. Effects of parity, calving season and month of production were also evaluated. Three single-nucleotide polymorphisms were detected. The first mutation, located at position 89 of the 17th exon (c.628C>T), is responsible for the amino acid change p.Ser178 (B allele)/Leu178 (A allele). The other two polymorphisms, detected at the positions 144 (c.882G>A) and 239 (c.977A>G) of 19th exon, respectively, are silent (3ʹ UTR, untranslated region). Associations between the CSN1S1 genotypes and milk production traits were investigated using 4122 test day records of 503 lactations from 175 buffalo cows. Milk yield, fat and protein percentages were analysed using a mixed linear model. A significant association between the c.628C>T SNP and the protein percentage was found. In particular, the CC genotype showed an average value ~0.04% higher than the CT and TT genotypes. The allele substitution effect of cytosine into thymine was –0.014, with a quite low (0.3%) protein percentage contribution to total phenotypic variance. A large dominance effect was detected. Characterisation of the CSN1S1 transcripts and a method based on MboI amplification created restriction site PCR for a rapid genotyping of c.628C>T are provided.

Use of partial least squares regression to predict single nucleotide polymorphism marker genotypes when some animals are genotyped with a low-density panel.

High-density single nucleotide polymorphism (SNP) platforms are currently used in genomic selection (GS) programs to enhance the selection response. However, the genotyping of a large number of animals with high-throughput platforms is rather expensive and may represent a constraint for a large-scale implementation of GS. The use of low-density marker (LDM) platforms could overcome this problem, but different SNP chips may be required for each trait and/or breed. In this study, a strategy of imputation independent from trait and breed is proposed. A simulated population of 5865 individuals with a genome of 6000 SNP equally distributed on six chromosomes was considered. First, reference and prediction populations were generated by mimicking high- and low-density SNP platforms, respectively. Then, the partial least squares regression (PLSR) technique was applied to reconstruct the missing SNP in the low-density chip. The proportion of SNP correctly reconstructed by the PLSR method ranged from 0.78 to 0.97 when 90% and 50%, respectively, of genotypes were predicted. Moreover, data sets consisting of a mixture of actual and PLSR-predicted SNP or only actual SNP were used to predict genomic breeding values (GEBVs). Correlations between GEBV and true breeding values varied from 0.74 to 0.76, respectively. The results of the study indicate that the PLSR technique can be considered a reliable computational strategy for predicting SNP genotypes in an LDM platform with reasonable accuracy.