Search for ultrasound and metabolic biomarkers to optimize pig insemination Like for many breeding species, determining the time of ovulation in sows correctly influences the success of insemination and therefore achievement of an optimal fertility rate on economically competitive farms. Today, detecting sow estrus after weaning remains a problem in farm management. It requires a large amount of work for an uncertain result due to its relative inefficiency, involving the libido of the boar used for detection, the sexual behavior of the sow, the level of expertise and the time spent handling animals. These factors contribute to the resulting percentage of infertility on farms and explain why 2 or 3 artificial inseminations (AI) are needed to reach a fertility rate of 90%. The aim of this study was therefore to validate predictive parameters of sow estrus before ovulation. In this study, eight sows were followed after weaning. Ultrasound scanning was used to analyze uterus development, while the metabolome of salivary samples was processed using 1H Nuclear Magnetic Resonance. Ultrasound scanning detected significant changes in uterine development. Based on these analyses, 24 hours after uterine development, two inseminations were performed 12 hours apart with semen from boars of different breeds, showing that the fertile AI was the first AI for 73% of piglets. Finally, analysis of metabolomic profiles showed a significant change in the composition of saliva and differentiated ca. ten compounds of interest that evolved from the weaning of piglets to insemination. Further studies are needed to confirm these preliminary results and to find a biomarker that correlated exactly with uterine and ovarian development during the weaning/estrus period for all sows.