Variance parameter estimation for age at puberty phenotypes under 2 levels of phenotype censorship

Abstract

Age at puberty (AGEP) is a moderately heritable trait in cattle that may be predictive of an animal's genetic merit for reproductive success later in life. In addition, under some mating strategies (for example, where mating begins before all animals have attained puberty) animals that attain puberty at a relatively young age will also likely conceive earlier than their herd mates, and thus begin their productive life earlier. Unfortunately, AGEP is challenging to measure because animals must be observed over a period of several months. Our objectives for this study were twofold. The first objective was to produce variance components for AGEP. The second objective was to investigate the implications of a simplified phenotyping strategy for AGEP, when the interval between repeated blood plasma progesterone measures was extended from weekly to monthly, increasing the extent of left, interval, and right censoring. We measured AGEP in a closely monitored population of around 500 Holstein-Friesian heifers, born in 2015 and managed under a seasonal, pasture-based dairy system. Animals were blood tested weekly from approximately 240 to 440 d of age and were deemed to have reached puberty when blood plasma progesterone elevation (>1 ng/mL) was detected in 2 of 3 consecutive blood tests (AGEP_Weekly). To simulate a simplified phenotyping strategy based on monthly herd visits (AGEP_Monthly), we selectively disregarded data from all but 3 blood test events, when animals were around 300, 330, and 360 d of age (standard deviation = 14.5 d). The posterior mean of estimated heritabilities for AGEP_Weekly was 0.54, with a 90% credibility interval (90% CRI) of 0.41 to 0.66, whereas it was 0.44 (90% CRI 0.32 to 0.57) for AGEP_Monthly. The correlation between EBVs for AGEP_Weekly and AGEP_Monthly was 0.87 (90% CRI, 0.84 to 0.89). We conclude that in this population, AGEP is a moderately heritable trait. Further, increasing phenotype censorship from weekly to monthly observations would not have altered the main conclusions of this analysis. Our results support the strategic use of censoring to reduce costs and animal ethics considerations associated with collection of puberty phenotypes.

Summary: We measured age at puberty (AGEP) in a closely monitored population of approximately 500 Holstein-Friesian cows, born in 2015 and managed under a seasonal, pasture-based dairy system. Animals were blood tested weekly from approximately 240 to 440 days old and were deemed to have reached puberty once blood plasma progesterone (BP4) elevation (>1 ng/mL) was detected in 2 of 3 consecutive blood tests (AGEP_Weekly). To simulate a simplified phenotyping strategy based upon monthly herd visits (AGEP_Monthly), we selectively disregarded data from all but 3 blood test events, when animals were around 300, 330, and 360 days old (SD = 14.5 days). The correlation between estimated breeding values for AGEP_Weekly and AGEP_Monthly was 0.87 with a 90% credibility interval (CRI) of 0.84 to 0.89. The posterior mean of estimated heritabilities for AGEP_Weekly was 0.54 (90% CRI 0.41 to 0.66). Our results support the strategic use of censoring to reduce costs and animal ethics considerations associated with collection of puberty phenotypes.