Abstract :
The fatty acid (FA) profile of milk fat may affect the nutritional and market value of sheep
dairy products. In this study we present repeatability estimates and results of quantitative
trait loci (QTL) detection analyses for FA profile of sheep milk obtained by a genome
scan approach applied to an experimental Sardinian×Lacaune back-cross population. The
experimental flockwas constituted by 967 daughters of 10 F1 sires corresponding to a classical
daughter design for QTL detection. Fatty acid contents were determined twice in the
middle of 2nd (847 records) and 3rd lactation (795 records) by gas chromatography. Means,
standard deviations as well as raw correlations were calculated for the 17 most abundant
fatty acids, their sums and ratios, based on 1454 records from 727 ewes with two samplings.
A repeated measurement model including, as a fixed effect, the year of production
by management group interaction and the random effects of sire and animal within sire
was applied to all the analysed variables. For each trait R2 statistics associated with the
fixed part of the model, the sire and the animal within sire components of the total variation
were estimated. The genotypes at 146 microsatellite markers spanning the 26 ovine
autosomeswere available. The individual solutions for the 727 ewes with two recordswere
used as phenotypes for QTL analyses. An across-family single trait QTL analysis was carried
out by within-family linear regression. Chromosome locations significant at least at the 5%
chromosome-wise level were retained. The overall pattern of raw correlations showed that
FA profile is a very complex system resulting from numerous inter-relationships between
basic FA. The portion of variation absorbed by the fixed part of the model differed markedly
across variables, as showed by R2 statistics which ranged from 11.3% to 78.3%. FA showed
large variability in terms of repeatability estimates, which ranged between 14.1% for C18:3
and 57.3% for C4. The most repeatable FA were C4, C14:1 and C17. For almost all the analysed
variables, significant genome locationswere found. Only for C4was no genome region
affecting the trait variability identified. The highest number of significant locations was
found for C14 1 and its ratio with C14. The most significant QTL were detected for C14 and
C16 on OAR 11 and for monounsaturated FA (MUFA) on OAR 6. A total of 82 locations significant
at least at the 5% chromosome-wise threshold were detected. The present study
suggests that for fatty acid profile of sheep milk both classical quantitative genetic approach
and genomic information based selection approaches will be realistic options in the future.
© 2008 Elsevier B.V. All rights reserved
