Service Communautaire d'Information sur la Recherche et le Développement - CORDIS

QTL for milk production traits in the Lacaune and Manech populations

For QTL detection, the purebred French families corresponded to a grand-daughter design (GDD), taking advantage of the extensive use of AI in the breeding programmes of the Lacaune and Manech breeds (table 2). At the end of the second year of the project, 3,836 blood samples (for DNA storage) of Lacaune, Basco-Bearnaise or Manech AI rams have been collected.

Twenty two GDD families were chosen, 18 in Lacaune, 1 in Basco-Bearnaise, 3 in Manech breeds, for QTL detection: these 22 GDD families corresponded to 783 AI rams (22 sires and 761 sons) progeny tested in the nucleus flocks, with on average 34.5 sons per family. The sons were born between 1992 and 2002 and had 89 daughters on average.

Nine traits were considered in the analysis: milk yield (MY), fat yield (FY), protein yield (PY), fat content (FC), protein content (PC), lactation SCS (LSCS), teat angle (TA), udder cleft (UC), and udder depth (UD). The phenotypic unit of measurement is daughter yield deviations (DYD), i.e. the average of the phenotypes of the daughters of each son adjusted for the environmental effects and additive genetic merit of the daughter’s dams, obtained from the French official dairy sheep breeding value estimates (BVE).

For milk traits considered on a lactation basis, a BLUP animal model with repeated records is used including the fixed effects of flock x year x parity, age and month at lambing within parity and year, for milk yield the lambing-first test-day interval within year and parity, for fat and protein yield, fat and protein contents the effect of qualitative recorded category defined with the number of test-days x average lactation stage at recording; and genetic groups of unknown parents.

Since 1999, the milk traits BVE account for heterogeneity of variance: genetic, residual, and permanent environment variances are fitted according to flocks, year and parity. The model for genetic evaluation of SCS was also defined as a repeatability model for the lactation average LSCS including fixed effects comparable to those defined for milk traits. Finally for udder-type traits, a multiple-trait animal model was applied including the fixed effects of flock x year x classifier, year x lactation stage x date of lambing, year x number of lambs.

When a QTL for milk traits is detected for the 22 families and also only for the 18 Lacaune families, it is tabulated 2 times in the results, but of course only one QTL is counted.

A total of 11 QTL were found at the genome wise significance threshold with a suggestive linkage (p=0.038): 1 for MY on OAR 14, 1 for FY on OAR 16, 2 for PY on OAR 16 and OAR 17, 4 for F% on OAR1, OAR 2, OAR 9 and OAR 10, and 3 for P % on OAR 2, OAR 5, OAR9. Each detected QTL segregated in 2 to 5 families. The estimated substitution effects ranged from 0.7 to 3.0 standard genetic deviation, illustrating as a trend the limited power of the design, and sometimes also a weak informativeness of the closest markers for the informative families. Furthermore 5 results were chromosome-wise (p < 0.05) or nearly chromosome-wise significant : 1 for FY on OAR 12, 2 for PY on OAR 3 and OAR 14, 1 for F % on OAR 4, and 1 for P % on OAR 6.

Informations connexes

Reported by

INRA - SAGA- BP 52627
31326 Castanet-Tolosan