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

QTL for udder morphology and kinetic of milk emission in the Sarda x Lacaune population

One of the aim of the project was identifying quantitative trait loci affecting udder morphology and milk emission during machine milking in dairy ewes. Both these features greatly affect milkability of ewes and are thus of major interest for sheep farmers which wish to simplify milking activities and reduce milking time. To achieve our aim we first developed or improved two methods for appraising udder morphology. The first one was based on classical linear scoring of basic udder traits (teat placement, TP; udder attachment UA, udder depth UD, degree of separation of the halves DS). The second was based on the extraction of objective measurements of the udder from digital pictures. The reliability and feasibility of both techniques was assed by estimating repeatability and correlations of measures realised with the 2 methods and their correlations with measurements directly taken on the animals. Results indicated that both udder scoring and digital picture analysis are useful tools for appraising udder morphology in sheep.

The kinetics of milk emission during machine milking was recorded with an automatic device developed by INRA, which provides some measures of milk emission speed. Parameters provided at each individual milking are: average (AMF, ml/s) and maximum milk flow (MMF, ml/s), the moment of maximum milk flow occurrence (TMMF, s), and the time needed to collect the first 160 ml of milk in the jar (latency time, LT, s). Preliminary analysis showed that, among the traits recorded, MMF and TL are the most pertinent for describing milking individual speed, given that high MMF and low LT determine shorter milking time.

On the SardaXLacaune population udder scoring was performed once a month in 2000 and 3 times a year from 2001 to 2003; digital pictures of the udder were taken once a year in 2000 and 2001, milk emission data were collected twice a month for 4 years. Data were adjusted for the main environmental effects and QTL detection was performed on solutions of random individual effects for repeatability mixed models, using the QTLmap software. An original approach was used to increase the power of detection of non pleiotropic QTL when 2 correlated traits are investigated. It is based on the correction of the trait of interest for the genetic covariance with the correlated trait. In all, the analyses of different variables describing the udder morphology allowed for the detection of 53 locations significant at the 5% chromosome wise significance level.

Most of these locations concerned the same udder trait measured in different ways. Consistent results arise from data issued from different measurement approaches. QTL with a high confidence level were detected on OAR 3, 4, 9, 14, 16, 20, 22, 26. In particular, on OAR 3, 4, 14 and 16, QTL affecting measures of udder width were detected. The number of informative families ranged from 2 to 4. QTL effects varied from 0.42 to 1.90 within family s.d. unit. QTL affecting cistern height was mapped on OAR9, OAR14 and OAR26. These finding are of particular importance for sheep milkability, given that the height of the cistern greatly affects stripping. Finally significant QTL affecting udder height or udder attachment was detected on OAR22. The existence of QTL affecting udder UA and UD was suggested for OAR6, OAR20 and OAR 26. QTL which affected milk emission speed were detected on OAR9, 11, 17, and 20. In Particular, a highly significant QTL, which affected milk emission speed, but had no effect on milk production, as detected on OAR11. This finding lead to envisage direct selection on milking speed whenever genetic gain for this trait, due to indirect response on milk yield selection, was not considered sufficient.


At present, only few studies on QTL affecting udder traits and milkability have been published on dairy cattle and none on dairy ewes. Our results represent thus a 1st step towards the possibility of selecting dairy ewes for these functional traits on the basis molecular information. The 2nd step is to verify whether the QTL detected in the backcross population are also segregating in the parental breeds and to define more precisely their location. As far as the Sardinian breed is concerned, this step has already been undertaken. A resource population of approximately 900 7/8 or 15/16 Sardinian x Lacaune (Sardinian recurrent parent) has been procreated. It is organised in 24 half-sib sire families of around 40 daughters and will be measured for udder morphology and milk emission traits as well as for other traits of interest, for 4 years. Twenty-two sires came from the AI centre of the Sardinian breed and are thus representative of the selected population. Furthermore our Institute has the availability of a DNA bank which will allow us to validate QTL for traits routinely recorded in the herd-book farms, i.e., concerning milkability, for udder morphology.

Informations connexes

Reported by

Istituto Zootecnico e Caseario per la Sardegna
S.S. Sassari-Fertilia Km18.6
07040 OLMEDO
Italy