The livestock sector is highly concerned by the global food system, both by an expected increase in the demand for animal products such as milk, and by the ecological footprint of animal production, which must be minimized. Increasing feed efficiency (FE) in dairy cows would reduce some of the direct emissions (methane and ammonia) from livestock production but would also have a substantial positive impact on the induced emissions associated with crop production, due to the better feed conversion. Genetic improvement of FE is a particularly attractive strategy because it would impact most of the dairy farms for a limited cost. The decreased use of feed inputs implied by such an efficiency gain would give a competitive advantage to dairy production, but will also contribute to reducing environmental impacts. Thus, this project is expected to provide the essential elements needed for genetic selection strategies to improve FE in dairy cows. It fits with the fifth of the major societal challenges of ANR and the first research theme of the Apis-Gene company on FE and limitation of N pollution and methane emission by ruminants to improve the overall efficiency of ruminants. Selecting for FE is not as straightforward as it might first seem, there is evidence to suggest that robustness and adaptive capacity, especially for reproductive females, can be adversely affected by short-sighted strategies to improve efficiency. Thus, the choice of indicators used to assess FE is of great important, and it is essential to verify and validate the anticipated benefits of any such strategies to improve efficiency for their long-term consequences. Another key issue is to be able to better exploit new possibilities to target specific characteristics that contribute in part to FE. Such characters have rarely been studied because they have been very difficult to phenotype. The project will use new phenotyping technologies and the newly available information from them to develop selection for efficient use of body reserves whilst limiting the risks of undesirable trade-offs with other life functions that have been associated with high levels of production in dairy cows. Deffilait aims to elucidate ways by which to improve the FE of dairy cows without decreasing their robustness, to build strategies for doing this, and models to predict the future increases in FE attainableby selection programs, and directly on farm. The project will first involve developing new tools for large-scale phenotyping of the major biological characteristics that are directly involved in FE. Theproject will produce new tools to better estimate body condition, morphology, and digestive efficiency in large scale studies. These phenotypic measures will also impact on our capabilities for on-farm advising, and monitoring in livestock, which are also levers for improving efficiency at farm level. Then, to study the major determinants of FE, the project will also build an original database of dairycow lactations with a large set of phenotypes to describe the main sources of energy transformation, thus explaining the observed between-animal variability in FE. This dataset will then be used to quantify the contribution of the different mechanisms to the variability in FE, and to test different indicators and strategies to improve FE. A specific focus will be made on body reserves mobilization in early lactation to assess its genetic components and correlation with other traits with a larger dataset involving commercial farms. The project will then develop simulation tools to predict the short- and long-term consequences of different selection strategies in different environments. The expected results will contribute to the definition of strategies of selection to combine efficiency and robustness. The project will provide a coherent framework to undertake a balanced genetic selection on these traits, and thereby make a significant - and lasting - contribution to improving FE.

The project was retained by the ANR for a total amount of aid requested of € 703,429 and a duration of 4 years (2016-2019). The project is funded by € 448,000 by ApisGene.