Genetic analysis of new progesterone-based fertility traits in dairy cows measured on-farm

Diana Sorg,Monika Wensch-Dorendorf,Kati Schöpke,Gunter Martin,Renate Schafberg,Nicole Reinhold,Steffen Pache,Hermann Swalve
DOI: https://doi.org/10.3168/jds.2016-11865
Abstract:The objectives of this study were (1) to analyze the agreement of a standard laboratory ELISA for progesterone (P4) with an automated on-farm ELISA kit operated under commercial conditions in 1,297 milk samples from 50 dairy cows; (2) to study the influence of the method of detection of luteal activity on genetic parameters of fertility traits based on P4 measured with an automated on-farm ELISA once weekly from wk 3 to 9 postpartum in the milk of 1,304 cows; and (3) to study the influence of sampling frequency (once or twice weekly from wk 3 to 9) on the same traits from 296 cows. Luteal activity can be detected when there is an active corpus luteum in the ovary producing P4 and indicating the onset of reproductive cyclicity after calving. The on-farm ELISA overestimated P4 contents by a mean square error of prediction of 2.76 ng/mL and had an intermediate Spearman correlation with the laboratory kit (0.54). For the second objective, the postpartum interval to the commencement of luteal activity (C-LA), proportion of luteal activity between d 15 and 63 postpartum (P-LA), calculated as the number of samples above the threshold for high P4 values divided by the number of all samples, and delay of first ovulation (DOV1), defined as C-LA occurring later than d 45 postpartum, were derived from the P4 profiles. Both C-LA and DOV1 were determined by (a) thorough qualitative visual inspection of the profile, (b) the profile's mean as threshold for the first increase in P4 postpartum, indicating commencement of luteal activity, and (c) 3 ng/mL as threshold for the first increase in P4, a value that has been used by many other studies. Similarly, P-LA was determined by using methods (b) and (c). Estimates of heritability were 0.04 to 0.13 for C-LA, 0.12 to 0.23 for P-LA, and 0.03 to 0.07 for DOV1. Genetic correlation of P-LA with C-LA and with the profile's mean P4 was -1.00. The profile's mean had a higher estimate of heritability (0.11-0.12) than C-LA or DOV1. It can be calculated as the arithmetic mean of all P4 values of a profile, whereas C-LA, P-LA, and DOV1 need a definition of a threshold for high P4 values. We therefore suggest the profile's mean as a promising candidate for further research. For the third objective, once-weekly sampling was mimicked by neglecting every second sample, and C-LA and DOV1 shifted toward a later onset of cyclicity. Thus, a common standard for sampling regimen and detection algorithm is essential to avoid incompatibility between studies.
What problem does this paper attempt to address?