Genomic correlations between milk mid-infrared spectra and, milk production and feed efficiency related traits in French Lacaune dairy sheep

Archive ouverte

Robert-Granié, Christèle | Machefert, Coralie | Larroque, Hélène

Edité par CCSD -

International audience. In French dairy sheep, Fourier-Transform infrared (FTIR) milk spectra are routinely stored in national recording database for predicting milk components concentrations, used as traits in national genetic evaluations. Spectral data revealing milk's chemical composition could be also considered as predictors reflecting the physiological status of animal. The genetic variability of milk FTIR spectra was explored by estimating heritability at each individual spectral point between 0 and 0.42 from French Lacaune dairy sheep’s milk. Genomic correlations of spectra with currently selected traits (especially if low heritability) or new phenotypes (costly measures) could help to improve genomic selection of these traits. This study aimed to estimate genomic correlations between milk FTIR spectra and milk yield, milk components, feed efficiency and body reserves variations in Lacaune sheep breed.From 2020-2021, 36,873 milk samples within six-monthly test-days were collected from 4,712 ewes in 8 farms, as part of the SMARTER European project. Daily milk yield (DMY), fat content (FC) and protein content (PC) predicted through mid-infrared spectrometry, somatic cell count (SCC) and spectral data were recorded. The FTIR spectrum provided by spectrometers containing 1,060 wavenumbers (5,012 to 926 cm-1) was narrowed to 446 by removing water absorption-related noise. Each test-day, lactation feed conversion ratio (LFCR) and residual energy intake (REI) were computed and considered traits for lactation net energy feed efficiency based on largely collective feed data. Body reserves variations (BCSΔ) were evaluated by comparing five body condition scores throughout the production period for each ewe. 1,794 ewes were genotyped with low-density chip then imputed to SNP50 Bead-Chip. Genomic correlations between each wavenumber and each trait were estimated by bivariate animal models using AI-REML including pedigree and genomic relationships. Parity and an interaction of flock-year-lactation stage were included as fixed effects and an additive genetic effect on the animal, permanent environmental effect and residual as random effects. The average heritability estimations for milk-related traits were 0.15 (0.02) for DMY, 0.32 (0.02) for FC, 0.39 (0.02) for PC, 0.09 (0.01) for SCS, and was 0.29 (0.02) for all the wavenumbers of FTIR spectra. Average heritabilities of feed efficiency-related traits were lower with 0.09 (0.01) for LFCR, 0.11 (0.01) for REI and 0.04 (0.01) for BCSΔ. FC and PC showed strong and similar genomic correlations patterns with milk FTIR spectra, in contrast to that observed for DMY with lower amplitude. Genomic correlations with SCS were similar to those for DMY, with predominantly positive values. Regions of spectra had strong genomic correlations with REI (-0.75 to 0.67), modest with LFCR (-0.35 to 0.28) and weaker with BCSΔ (-0.31 to 0.22). The wavenumbers most closely linked genetically to feed efficiency-related traits were in the spectral regions associated with fatty acids and proteins.Project financially supported by INRAE, Occitanie region and European Union’s Horizon 2020 research and innovation program under grant agreement No 772787 (SMARTER).

• Description
• Sujet/Public
• Outil
• Outil d'anticipation
• Mémoire et thèse
• Gestion