Genetic Analysis of Milk and Beef Traits in the Autochthonous Rendena Dual Purpose Breed

The Rendena breed is an Italian small cattle population accounting for about 4,000 registered cows in the herd book, native of the homonymous Valley in Trento province and currently reared in the North-east of Italy. The selection objective in Rendena breed is the improvement of both quality and quantity of milk and meat with higher emphasis for milk, in tune with the maintenance of the fertility, longevity and rusticity in the breed, and with a special attention to the control of inbreeding, considering the reduced population size. This thesis has aimed to deepen some aspects of genetic selection for the main productive aptitudes in the breed (milk and meat) investigating the possibility of implementing new breeding goals on the basis of new phenotypes such as udder health and milk technological properties considering carefully their genetic correlations with the quantity and quality of milk. Beside these objectives, the investigation on possible variance heterogeneity for milk yield traits and on the relationship between beef traits obtained in primiparous cows and young candidate bulls have been carried out. Therefore, the ultimate aim was a further improve of the knowledge on the genetic structure of the whole panel of traits accounting for the dual purpose attitude in the Rendena breed looking at possible additional improvement of the selection scheme. The first three steps of the study have been concerned different aspects of the milk production aptitude in the breed, using the currently repeatability test-day model applied in routinely genetic evaluation for milk. The first chapter of this thesis evaluated possible heterogeneity of variance for productive traits using average milk yield recorded in Rendena farms under functional control in years 2004-2013 (114 farms). Medium and high productive groups (average milk yield of 15.63 ± 4.96 and 19.95 ± 5.95, respectively) were identified via cluster analysis. Subsequently, the estimated heritability and genetic correlations between groups were obtained via bi-trait animal model using 73,916 test-day records of milk, fat and protein yield belonging to a sample of 49 herds with the same numerical proportion as in medium and highly productive levels and considering yields in the two productive groups as different traits. The genetic components and heritability of milk, fat and protein yield were heterogeneous between the two productive groups, with higher values for the highly productive group (0.16 and 0.30, respectively). However, the genetic correlations for milk yield traits between productive groups resulted sufficiently great (≥ 0.88) to support the hypotheses of no heterogeneity between productive groups. Even the rank correlation between EBVs of bulls that had daughters in both groups confirmed this finding (0.98). The results suggested that genetic evaluation not accounting for the heterogeneous variance may be applied without producing biased estimates for the bulls EBVs. However, a more careful selection should be considered for bull dams. In the second study genetic parameters for the content of milk somatic cells (SCC) were estimated using different and alternative methods. In particular, the SCC traits defined for genetic evaluation were: somatic cells score (SCS), log total somatic cell counts (LTSCC) and three different thresholds (80,000, 150,000 and 400,000 cells/ml) for somatic cells. A total of 69,796 TD records of milk, fat, protein yields and SCC belonging to 4,325 cows was used to analyse heritability and genetic correlations between SCC and yield traits using bi-trait animal model under Bayesian approach. The threshold traits showed heritability values expressed under a liability scale ranging from 0.04 to 0.07, lower than those obtained for SCS and LTSCC traits that were equal to 0.09 and 0.10, respectively. The genetic correlations between yield traits and the SCC traits were almost null (from -0.09 between fat yield and SCS, to 0.13 between protein yield and SCC400), with the only exception of LTSCC (0.29 for milk, 0.16 for fat and 0.28 for protein). These results indicated that the genetic selection for milk, fat and protein production does not negatively influence the content of somatic cells in milk. Nevertheless, the complete framework of genetic relationships of SCC with all traits under selection should be considered to decide for possible inclusion of SCC in the breeding program. In the subsequent study, milk technological traits for cheese production were analysed using their Fourier Transform Infrared (FTIR) predictions (milk coagulation properties, MCP, curd firmness model parameters, cheese yields, the nutrient recoveries in the curd and the daily cheese yields), obtained from routinely functional controls carried out in the Trento province. A total of 9,947 FTIR predictions representing the test-day records of 997 Rendena cows, collected over a 5-yr period, were available for the study. The (co)variance components were estimated for the milk technological traits, milk production, and milk composition traits (fat, protein and lactose percentages, and SCS), via a bi-trait analysis by means of a Gibbs sampling algorithm. Milk of Rendena cows exhibited good aptitude for coagulation. In general the heritability obtained for all the predicted traits were comparable and in some case greater than those reported for milk production and composition. Heritability estimates for quality traits ranged from 0.11 for fat percentage to 0.39 for time of maximum curd firmness, whereas SCS showed the lowest heritability (0.01). The improvement of milk production favoured some of the traditional MCP traits, but did not promote cheese yield traits (average genetic correlation of -0.34). However, the relationships between milk production and the nutrients recovery traits were almost null. Conversely, favourable genetic correlations were obtained between milk composition and cheese yields and nutrient recoveries traits (average values of 0.57 and 0.34, respectively). These results suggested that it is possible to efficiently improve the technological properties of milk for cheese production and therefore the economic value of milk in the Rendena breed. The improvement of beef aptitude together with the milk ability has played an important role for maintaining the dual attitude in the selection of the Rendena breed. Therefore, the last part of the study was aimed at estimating genetic relationships between muscularity type traits obtained in primiparous cows (front muscularity; back, loins, and rump; thigh, buttocks side and rear views) with the same types scored on candidate young bulls at the end of the performance test for beef aptitude, and the performance test traits (average daily gain, ADG; EUROP fleshiness evaluation and dressing percentage). Muscularity linear type traits recorded on 11,992 first parity cows and both performance test traits and the muscularity types scored on 957 candidate young bulls at the end of the performance test were used. Through a series of single-trait REML analyses, genetic parameters for both muscularity traits of cows and young bulls, and beef performance test traits of young bulls were obtained. The models for the two datasets of cows and bulls accounted for different fixed effects, i.e., herd-year-classifier, age at calving and days in milk for muscularity evaluated on primpiparous cows, and group of contemporary, parity order of mother for muscularity and performance test traits relative to young bulls. The only communality between data sets was the random additive genetic variance. All the previous effects were jointly retained for the AIREML bivariate analysis to assess the genetic correlations between beef traits. Heritability estimates were moderate for both groups of traits and ranged from 0.25 for front muscularity in primiparous cows, to 0.36 for both back, loins and rump and dressing percentage recorded in young bulls. Medium to high genetic correlations were found between performance test and muscularity type traits collected in young bulls (from 0.47 between thigh, buttocks side view and ADG to 0.97 between thigh, buttocks side view and in vivo EUROP score). Low genetic relationships were obtained between muscularity linear type traits of primiparous cows and performance test traits or the same types traits scored on young bulls , which showed genetic correlations ranging from 0.07 to 0.82. These results indicate a common way for the meat selection route in Rendena, suggesting the possibility to account for young bull muscularity traits in genetic improvement of the breed. In conclusion, all the four studies suggested possible positive aspects, either methodological or related to productive or functional traits, likely to be used in the selection practice of Rendena breed