ADSA®-EAAP speaker exchange presentation: Genetic analysis of multivariate indices of detailed fatty acid profile determined by gas chromatography in bovine milk

Abstract Text:

The genetic improvement of fatty acid (FA) composition is a crucial point for enhancing milk dietary and nutritive properties. However the development of an appropriate breeding goal for this trait is hampered by the large number of FA and the complex correlation pattern among them. Multivariate factor analysis (MFA) is able to derive synthetic variables that can describe efficiently a multivariate system with a complex covariance structure. This study was aimed at:i) elucidating the structure of relationships between milk yield, composition and detailed FA composition by using the MFA; ii) estimating genetic parameters for the new-derived synthetic variables. Individual milk samples were collected from 1,158 Brown Swisscowsand gas chromatography wasused to obtain detailed milk FA compositions. MFA was carried out on 53 variables(i.e., 47FA and 6 milk production and composition traits). A total of twelve factors were extracted, able to explain about the 75% of the total variance. Factor scores were then used as new phenotypes for estimating (co)variance components using a Bayesian linear animal model via Gibbs Sampling. The model accounted for the effect of days in milk, parity, herd and animal additive genetic effect. Factor scores exhibited a clear structure in term of relationship with the original variables and they were classified, from a biological point of view, as: ‘de novo FA’, ‘milk yield-branched FA’, ‘biohydrogenation’, ‘long chain FA’, ‘short chain FA’, ‘milk-fat-protein’, ‘odd FA’, ‘conjugated linoleic acid’, ‘linoleic’, ‘udder health’ and ‘vaccelenic’, respectively. Marginal posterior means (SD) of heritabilitiesfor the aforementioned factor scores ranged from 0.048(0.02) for ‘vaccelenic’to 0.310(0.09) for ‘desaturation’. Moderate heritability estimates were observed for ‘milk yield-branched FA’(0.214±0.07),‘linoleic’(0.201±0.08),‘biohydrogenation’ (0.193±0.08), and ‘short chain FA’ (0.157±0.07), respectively.Results highlight the existence of important and exploitable genetic variation in these derived phenotypes. In particular factors strongly associated with variables related to mammary neo-synthesis and desaturation, may offer interesting perspectives for improving milkfatnutritional properties by selective breeding.

Keywords: Fatty Acid profile, Multivariate Factor Analysis, heritability