Strong connectedness within Norwegian Cheviot and Fur Sheep ram circles allows reliable estimation of breeding values

Breeding programs for sheep in Norway are based on cooperatives of ram circles (RC). The key features of RC are selection of rams across member flocks and their rotation among RC flocks during the mating season. Genetic gains are disseminated to flocks outside RC (ORC). In both groups, natural service and AI are practiced. The objectives were to investigate 1) connectedness within and across RC and across RC and ORC, which impacts bias in genetic comparisons across flocks, and 2) opportunities to improve accuracy by including data from ORC flocks in genetic evaluation of RC flocks. Weaning weights in Cheviot and Fur Sheep from 1990 to 2010 were used. In Cheviot, in the last year of data (2010), there were 4 RC with 49 flocks and 1,824 ewes. Seventy-seven ORC flocks, with 1,246 ewes, also were recorded that year. In total, 214,391 pedigree and 131,012 performance records in Cheviot were available. For Fur Sheep, there was 1 RC with 8 flocks and 468 ewes in 2010 and 134 ORC flocks with 1,932 ewes. In total, 198,339 pedigree and 110,955 performance records in Fur Sheep were available. Unbiased comparison of EBV requires that genetic means of flock founders are similar or that flocks are genetically connected. The latter requires that rams sire enough progeny across flocks. In RC in both breeds and in 28.6% of Cheviot and 20% of Fur Sheep ORC flocks, the average prediction error correlation of flock mean EBV (flock rij) exceeded a threshold (0.10) for strong connectedness. These flocks also had similar genetic means: the variance between means of flock founders (genetic groups) was 1.05 (Cheviot) and 0.51 (Fur Sheep) times that of the additive variance for weaning weight. With less connected flocks included (flock rij ≤ 0.10), the between genetic group variance increased to 1.6 times the additive variance. When weaning weights from connected ORC flocks were included in the genetic evaluation of RC flocks, the size of the data increased by 1.07 times in Cheviot and by 3.6 times in Fur Sheep. However, the increase in accuracy of EBV of animals in RC flocks was trivial. Because founder animals of RC flocks had similar genetic means and were strongly connected, flock owners can comfortably focus on other aspects of their breeding program. Opportunities for closer cooperation between RC and connected ORC flocks exist and could be explored to accelerate genetic improvement within the Norwegian Cheviot and Fur Sheep breeds.