Genetic determinism of milk composition in fatty acids and proteins in ruminants, and selection potential

Abstract

This study presents the main genetic results obtained from the PhénoFinlait project with regards to genetic parameters and QTL detection for milk composition in fatty acids (FA) and proteins in three dairy cattle breeds (Montbeliarde, Normande, and Holstein), two goat breeds (Alpine and Saanen) and two sheep breeds (Lacaune and Manech Tete Rousse). Milk composition was estimated from midinfra red spectrometry. Genetic parameters were estimated from about 102,000 test-day records from 22,000 cows in first lactation, 67,000 records from 20,000 ewes and 45,000 records from 13,700 goats. Genetic parameter results were very homogeneous across species and breeds. They were found to be sensitive to the mode of expression of the traits, in % of milk or in % of fat or protein. Expressed in % of milk, test-day saturated FA (SAT) had higher heritability estimates than unsaturated FA (UNSAT) but this difference was smaller when traits were in % of fat. In goats, the results were markedly different with higher heritability estimates found for traits expressed in % of fat. FA measurements were highly genetically correlated across different stages of lactation except in the beginning of the lactation. Genetic correlation estimates were found to be positive across saturated FA, and also across unsaturated FA. Between saturated and unsaturated FA, correlation estimates were positive when FA were expressed in % milk but negative when FA were expressed in % fat. Saturated FA were strongly correlated with fat content. With regards to proteins, heritability estimates were very high for beta-lactoglobulin, moderate to high for caseins, moderate for alpha-lactalbumin. Correlation pattern showed a strong similarity between FA and proteins. Indeed, caseins were strongly correlated with each other and with protein content. Their correlation with whey proteins was positive or very negative, whether proteins were expressed in % milk or in % proteins.

QTL detection analyses were based on pangenomic genotyping data of 7800 cows, 1800 ewes, and 2300 goats. On average, 9 QTL were detected per FA trait and cattle breed. The most important QTL were found on chromosome 14 (DGAT1 gene), 5, 19, 27, 17, 11, and 13. A strong co-location of QTL was observed for FA sharing a common metabolic origin. A large proportion of the QTL seems to be shared across breeds. Twenty-two to 29 QTL were detected for each protein. The most significant QTL were found on chromosome 6 (2 regions close to ABCG2 gene and to casein cluster), 11 (beta-lactoglobulin gene), and 20 (2 regions, around GHR gene and around 58 Mb). The DGAT1 gene (BTA14, around 1.8 Mb) was also found to affect many proteins when expressed in % of milk.

These results show that milk composition in FA or proteins can be significantly modified by selection, even if the major characteristics cannot be changed. For instance, it is possible to increase the casein percentage in total proteins. It is also possible to increase the unsaturated FA fraction in fat, but at the expense of a decrease in fat content in milk.