A total of 7 experimental populations (9 breeds: 1 goat, 4 dairy sheep and 4 meat sheep breeds) were established and intensively analysed for proxies for feed efficiency (FE). A list of phenotypic traits (>=30) showed a relationship with FE. Among the biomarkers analyzed, blood metabolites and milk fatty acid profile standed out as the most promising. Furthermore, a list of candidate phenotypes were specified as possible proxies including live weight, back fat, muscle depth, milk fat, protein and lactose content, body condition score (BCS), feed intake, feeding behavior, rumen lipidome and rumen microbiota and GHG from Pac Chambers. For these FE traits SMARTER provided heritability, correlation with production and GHG traits, and some genomic results. To characterize resilience (R), new phenotypes collected included footrot, mastitis, gastro-intestinal parasites (GIN), survival and longevity, mothering ability and behavioural reactivity traits. Novel phenotypes such immunity related traits, ruminal temperature, metabolic stress, tolerance to heat stress, BCS changes, GPS-generated phenotypes, have been explored. We provided many genetic parameters for these traits and delivered the first breeding values for GIN (sheep and goat), footrot & mastitis (sheep) and longevity (goat). Genomic studies were carried out for 11 of these traits in seven breeds, leading to a better understanding of the genes and mechanisms that control resilience. The genetic link between R&E was studied using existing data, new experimental and genomic data and models. Firstly, comprehensive meta-analysis was completed which identified some trade-offs between R&E traits. Genomics studies were used to identify and characterize pleiotropic loci associated with both R&E traits. Four sheep and one goat selection experiments were analysed for studying the mechanisms underlying such trade-off between R&E traits. The main findings were: i) little evidence of major trade-offs between selection for R&E and ii) some key resilience mechanisms and phenotypes and iii) new methods and indicators to analyze resilience trajectories upon challenge. The experimental data was further used to adapt and calibrate resource allocation models, in order to predict response to challenges and various breeding strategies. The modeling provided the first ‘real’ empirical evidence that immunity has an energy cost. In order to characterize genetic diversity of small ruminants including hardy and underutilized breeds SMARTER created a common repository for existing and newly generated genetic and environmental data for 285 sheep and 166 goat populations (≥17K animals). Using standard population analyses methods, as well as a new statistical method dedicated to the historical annotation of genes, SMARTER characterized the genetic relationships between small ruminants’ populations and identify genomic regions involved in their bioclimatic adaptation. For the analysis and the selection of R&E traits in sheep and goats, methods were developed or adapted. To detect and analyse environmental stresses and traits associated to resilience SMARTER used different approaches (mixture model on variances, resilience in a Norm Reaction model, statistical indicators of deviation to target trajectories). Methods were also developed to i) refine and assess genomic predictions within and across small breeds, ii) ensure good selection programs across multiple environments or with various genotyping strategies, and iii) assess and handle genetic diversity and recessive anomalies. With the aim of improving the efficiency of selection for R&E traits by strengthening international collaboration, SMARTER first created the conditions and tools needed for international genetic evaluation. The first across-country genetic evaluations in small ruminants was implemented, and demonstrated technical and economic benefits. Meanwhile, SMARTER developed a process flow for routine genetic evaluations and assessed the interest of countries for such evaluations. Recommendations on 31 R&E phenotyping on a large scale were achieved, leading to guidelines for ICAR. An informative genotype panel across sheep populations was developed and made available on the SMARTER website. The research outputs from farm and population level agro-ecological impacts of breeding, demonstrated that i)there are significant opportunities to create value through genetic improvement for R&E traits ii)farmers clearly see the need for animal resilience and efficiency in robust future farming systems and iii)that there is a scope to create breeding goals to achieve the desirable balance of selection across production and R&E traits
