Genomic investigation of milk production in Italian buffalo

Abstract

The aim of this study was to test the feasibility of genomic selection in the Italian Mediterranean water buffalo, which is farmed mainly in the south Italy for milk, and mozzarella, production. A total of 498 animals were genotyped at 49,164 loci. Test day records (80,417) of milk (MY), fat (FY) and protein (PY) yields from 4127 cows, born between 1975 and 2009, were analysed in a three-trait model. Cows born in 2008 and 2009 with phenotypes and genotypes were selected as validation animals (n ¼ 50). Variance components (VC) were estimated with BLUP and ssGBLUP. Heritabilities for BLUP were 0.25 ± 0.02 (MY), 0.16 ± 0.01 (FY) and 0.25 ± 0.01 (PY). Breeding values were computed using BLUP and ssGBLUP, using VC estimated from BLUP. ssGBLUP was applied in five scenarios, each with a different number of genotypes available: (A) bulls ( 35); (B) validation cows (50); (C) bulls and validation cows ( 85); (D) all genotyped cows (463); (E) all genotypes (498). Model validation was performed using the LR method: correlation, accuracy, dispersion, and bias statistics were calculated. Average correlations were 0.71 ± 0.02 and 0.82 ± 0.01 for BLUP and ssGBLUP-E, respectively. Accuracies were also higher in ssGBLUP-E (0.75 ± 0.03) compared to BLUP (0.57 ± 0.03). The best dispersions (i.e. closer to 1) were found for ssGBLUP-C. The use of genotypes only for the 35 bulls did not change the validation values compared to BLUP. Results of the present study, even if based on small number of animals, showed that the inclusion of genotypes of females can improve breeding values accuracy in the Italian Buffalo. HIGHLIGHTSThe genotypes of males did not improve the predictions. Genotypes of females improve breeding values accuracy. Slight increase in prediction accuracy with weighted ssGBLUP.

Genomic Prediction in Local Breeds: The Rendena Cattle as a Case Study

“…However, those investigations were conducted with breeds with much larger population sizes, and results were expressed in terms of reliability [ 46 , 47 , 48 ]. Interestingly, in Cesarani et al [ 49 ], an analogous number of animals was used, and results in terms of bias and dispersion agree with results obtained in this manuscript, with a similar influence of weighting strategies, although the number of animals with genotype in their study is much lower than ours. While generally, different weighting strategies have led to different increases in the accuracy of the breeding value predictions [ 21 ], extreme shrinkage strategies (i.e., quadratic weight) can lead rapidly to a decline in accuracy as the interactions increase and generally present greater biased prediction [ 23 ].…”

Polymorphism detection of <i>DGAT1</i> and <i>Lep</i> genes in Anatolian water buffalo (<i>Bubalus bubalis</i>) populations in Turkey

“…In Murrah buffalo breed, Cardoso et al (2015) reported that DGAT1-VNTR (variable number of tandem repeats) in the promoter region was significant for the fat percentage of milk because of a located binding site of Sp1 (specificity protein 1), which is a transcription factor for the mRNA synthesis. Cesarani et al (2021) revealed that WssGBLUP is a weighted version of ssGBLUP, which is some major genes or SNPs, as DGAT1 for fat content has more weights that could increase the accuracy of breeding values for growth, milk traits.…”

Multibreed genomic evaluation for production traits of dairy cattle in the United States using single-step genomic best linear unbiased predictor

“…The ssGBLUP involves the inverse of a realized relationship matrix (H −1 ; Aguilar et al, 2010), which combines the pedigree relationship matrix (A) and the genomic relationship matrix (G). Despite past limits due to problems with unknown parent groups (UPG) and computational cost, recent studies demonstrated the validity of this method for evaluating breeding values in several different livestock species including dairy (Himmelbauer et al, 2021;Liu and Alkhoder, 2021;Pimentel et al, 2021) and beef cattle (Lourenco et al, 2015), buffalo (Aspilcueta-Borquis et al, 2015;Cesarani et al, 2021a), goats (Teissier et al, 2018), and sheep (Cesarani et al, 2019;Macedo et al, 2020). Recently, better prediction features of ssGBLUP compared with BLUP in US Holstein were reported (Cesarani et al, 2021b).…”