Naslov
Genomic selection in populations with low connectedness between herds

Autori
Kasap, Ante ; Mioč, Boro ; Hickey, John ; Gorjanc, Gregor

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
Book of Abstracts of the 69th Annual Meeting of the European Federation of Animal Science / EAAP Scientific Committee - Wageningen : Wageningen Academic Publishers, 2018, 301-301

ISBN
978-90-8686-323-5

Skup
69th Annual Meeting of the European Federation of Animal Science (EAAP)

Mjesto i datum
Dubrovnik, Hrvatska, 27.08.2018. - 31.08.2018

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
genomic selection, connectedness

Sažetak
Low connectedness between herds reduces accuracy of estimated breeding values and ranking of animals across herds. This is a common issue in breeding programs that do not use artificial insemination or do not actively exchange sires between herds. Availability of genomic data may present an opportunity to mitigate this issue by providing information on realised genome sharing between animals in different herds. In this study we used simulation to compare the accuracy of pedigree and genomic based estimates of breeding values in populations with varying level of connectedness. The simulation mimicked the structure of a typical sheep breeding program in Croatia. For genomic analyses, we used the single-step method and a training population that comprised either 1000 sires each with 20 phenotyped progeny or 1000 phenotyped progeny. Heritability was 0.3. Accuracy was evaluated for sires and for non-phenotyped progeny. Both pedigree and genomic based models included herd-year as a fixed effect. When the training population comprised sires with phenotyped progeny and each sire had progeny only in one herd both pedigree and genomic estimates had practically null accuracy. This was due to complete confounding between sire breeding values and herd-year effects. When sires had progeny in two herds accuracy increased. The increase was larger for genomic than for pedigree based estimates. Testing sires in more than two herds increased accuracy marginally. When the training population was comprised of phenotyped progeny and each sire had progeny only in one herd the accuracy of genomic estimates was higher than when training population was comprised of sires. When sires had progeny in two herds the accuracy was the same for both training population designs. In conclusion, when training population is comprised of sires low connectedness affects accuracy of pedigree and genomic based estimates in a similar way. In such cases training population for genomic selection should comprise phenotyped individuals rather than sires with phenotyped progeny in disconnected herds.

Izvorni jezik
Engleski

Znanstvena područja
Poljoprivreda (agronomija), Interdisciplinarne biotehničke znanosti

POVEZANOST RADA