Comparison of equations to estimate lactation curves using different sampling strategies in Angus and Brown Swiss cattle and their crosses

Abstract

THE OBJECTIVES OF THIS STUDY WERE TO COMPARE FOUR EQUATIONS TO ESTIMATE LACTATION CURVES IN ANGUS AND EUROPEAN BROWN SWISS CATTLE, AS WELL AS IN THEIR RECIPROCAL-CROSSES, BASED ON FOUR SAMPLING FREQUENCIES OF MILK PRODUCTION, AND TO DETERMINE THE MINIMUM NUMBER OF SAMPLES TO ADEQUATELY PREDICT LACTATION CURVES. MILK YIELD (MY) DATA (297 LACTATIONS) WERE COLLECTED WEEKLY BY THE WEIGH-SUCKLE-WEIGH TECHNIQUE UP TO 180 DAYS AFTER CALVING. EQUATIONS EVALUATED FOR EACH LACTATION WERE: 1) YT = A EXP(BT+CT 2); 2) YT = ATBEXP(-CT); 3) YT = T/(A+BT+CT 2); AND 4) YT = A+BT+CT 2. THE SAMPLING STRATEGIES WERE 26, 13, 7, AND 4 MEASUREMENTS OF DAILY MY DURING LACTATION. FRIEDMAN´S TEST WAS USED AS THE MAIN CRITERION TO COMPARE THE RANK OF THE EQUATIONS IN EACH SAMPLING FREQUENCY. THE CRITERION TO DETERMINE THE MINIMUM NUMBER OF SAMPLES TO PROPERLY FIT LACTATION CURVES WAS THE STANDARD ERROR OF THE ESTIMATED PARAMETERS ON WEEKLY SAMPLING AS COMPARED TO THE OTHER SAMPLING STRATEGIES. THE EQUATION OF BEST FIT DEPENDED ON COW´S GENOTYPE AND SAMPLING STRATEGY. HOWEVER, EQUATIONS 1 AND 2 CONSISTENTLY SHOWED THE BEST PERFORMANCE REGARDLESS OF SAMPLING FREQUENCY AND GENOTYPE STUDIED. THE SIMILAR VALUES OF STANDARD ERRORS FOR THE PARAMETERS ESTIMATED WHEN MY WAS REGISTERED WEEKLY OR EVERY OTHER WEEK, SUGGEST THAT EITHER EQUATION 1 OR 2 IS A GOOD CHOICE TO FIT LACTATION CURVES WHEN DAILY MY IS RECORDED EVERY OTHER WEEK IN THE STUDIED GENOTYPES.