DENSE DNA MARKER INFORMATION IS WIDELY USED TO MAKE PREDICTIONS OF PHENOTYPE, WITH MAJOR BENEFITS FOR HUMAN MEDICINE AND ANIMAL AND PLANT PRODUCTION SYSTEMS. THANKS TO THE RAPID ADVANCES IN MOLECULAR TECHNIQUES, GENOMICS IS BECOMING THE LINCHPIN OF GENETIC PROGRESS IN LIVESTOCK AND IS KEY TO FEEDING AN EVER-GROWING HUMAN POPULATION. LARGE INVESTMENTS HAVE ALREADY BEEN MADE TO SEQUENCE LIVESTOCK AND TO COLLECT GENOTYPES AND PHENOTYPES FOR GENOMIC PREDICTION. THIS PROJECT INTENDS TO CAPITALIZE ON THESE DATA BY DEVELOPING NEW ARTIFICIAL INTELLIGENCE (AI) METHODS THAT AIM TO INCREASE ACCURACY OF GENOMIC PREDICTION AND AT THE SAME TIME REDUCE COSTS TO ENABLE A BROADER INDUSTRY ADOPTION. AS WE MOVE TOWARDS FULL SEQUENCE DATA FOR PREDICTION, IT IS CLEAR THAT WE ARE ON A DIMINISHING RETURNS CURVE WITH CURRENT METHODS. THE INCREASE IN PREDICTION ACCURACY IS MOSTLY MARGINAL COMPARED TO LOWER DENSITY MARKER PANELS. AI EXCELS IN IDENTIFYING ROBUST PREDICTORS FROM LARGE NOISY DATA - THE OBJECTIVEOF THIS PROJECT IS TO TREAT GENOMIC PREDICTION AS A FEATURE SELECTION PROBLEM AND DEVELOP AI METHOD TO SUBSET SEQUENCE DATA INTO SMALLER TRAIT SPECIFIC MARKER PANELS THAT ARE MORE ACCURATE THAN USING ALL THE GENOMIC INFORMATION. IMPROVEMENTS IN ACCURACY OF PREDICTION LEAD TO HIGHER RATES OF GENETIC IMPROVEMENT WITH MULTIPLICATIVE EFFECTS ON THE PRODUCTIVITY OF LIVESTOCK SYSTEMS. THE ANTICIPATED RESULTS OF THIS PROJECT WILL SUPPORT INTERNATIONAL COMPETITIVENESS AS WELL AS ENABLE HIGHER YIELD AND MORE EFFICIENT FOOD PRODUCTION. PROJECT OUTCOMES ARE ALSO APPLICABLE TO PREDICTION OF HUMAN DISEASE RISK AND GENETIC SELECTION IN PLANTS.