Global breeding programs

A three-tier breeding structure for genomic selection in Sarda dairy sheep: concept and initial evidence

---

Abstract

Genetic improvement in Sarda dairy sheep has been limited by the low use of artificial insemination, incomplete pedigrees, and small progeny groups, which reduce the accuracy of sire evaluations. Therefore, reference populations for genomic selection must rely substantially on females. In collaboration with the Sheep Breeders' Association, responsible for the breeding program, we designed a three-tier structure to implement genomic selection: Nucleus (N) flocks, with intensive recording and genotyping of females and males; Traditional Breeders (TB), providing pedigree data, routine phenotypes, and male genotyping; and User Flocks (UFs), providing male genotyping. A case study was conducted on milk protein content with the aim of evaluating the efficiency of this approach by the GEBV reliability of breeding males. Data, provided by the national recording system, included 17,662 lactations from 5,363 ewes recorded from 2000 to 2024 in N and 8,897 lactations from 5,509 ewes with simplified recording schemes in TB. In total, 7,907 animals were genotyped with the Illumina OvineSNP50k BeadChip: 5,284 N ewes and 2,623 males, including 313 sires and 2,310 rams from N, TB, and UFs. Two genomic evaluation scenarios were compared: GBLUP, based on phenotypes and genotypes from N, and SSGBLUP, which combined N and TB data through the H matrix. Reliability, predictive ability, and the linear regression method were assessed on 664 ewes from N and 349 from TB. Reliability of rams' GEBV was computed from the prediction error variance, and pedigree-based reliabilities (0.79 for progeny-tested sires and 0.42 for pedigree indexes) were used as benchmarks. Genotyped males were classified as sires in both N and TB (108), in N only (144), in TB only (61), and as young rams from N (191), TB (430), and UFs (1,689). For both GBLUP and SSGBLUP, predictive abilities were markedly greater in N ewes (0.59-0.59) compared with TB ewes (0.27-0.29). Rams' reliability varied across categories and scenarios. Sires with daughters in both N and TB achieved the highest values (0.77 in GBLUP; 0.86 in SSGBLUP), with the proportion above 0.79 increasing from 57 to 84%. For sires only in N, reliability averaged 0.76-0.80, and 42-57% surpassed 0.79. For sires only in TB, reliabilities rose from 0.48 to 0.75, and 0-44% exceeded 0.79. Among young rams, reliability was 0.53-0.58 for N, 0.41-0.59 for TB, and 0.34-0.39 for UFs, with 98-100%, 41-59%, and 34-39% exceeding 0.42, respectively. The expected reliabilities for the breeding male categories confirm the feasibility and potential of the three-tier structure for genomic selection. The balance between N and TB flocks, along with animal exchange across tiers, is crucial for improving GEBV reliability and disseminating genetic progress. Genotyping females in TB flocks will enhance accuracy and the sustainability of the program.

Keywords: 2026

How to Cite:

Carta, A., Grande, S., Negro, A., Salaris, S., Sechi, T. & Usai, M., (2026) “A three-tier breeding structure for genomic selection in Sarda dairy sheep: concept and initial evidence”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2286362. doi: https://doi.org/10.31274/wcgalp.23957

Rights: 1