Beef cattle

Modeling age at first calving in Nelore cattle using Survival Analysis

---

Abstract

Age at first calving (AFC) is one of the main indicators of sexual precocity and reproductive efficiency in beef cattle and is traditionally evaluated using the categorical phenotype P30, which classifies cows according to whether they have calved by 30 months of age. Although operationally simple, this approach disregards the actual variation in time to the event and treats biologically distinct performances within the early category as equivalent, leading to information loss and a potential reduction in the detectable genetic signal. In this study, we evaluated the use of Survival Analysis as an alternative framework for modeling AFC in Nelore cattle, directly comparing it with the traditional P30 classification in terms of phenotypic discrimination and genetic parameter estimation. Data from approximately 100,000 cows born between 2010 and 2017 from 18 farms participating in the Nelore Brazil Genetic Improvement Program (ANCP) were analyzed. Cox proportional hazards models with a random sire effect (frailty) were fitted using R, including fixed effects of year of birth, calving season, and farm. Significant variation was observed among birth years, with a progressive increase in the hazard of first calving by 30 months in more recent cohorts (p < 0.01), suggesting genetic and/or management-related improvements over time. Calving season also had a significant effect, with cows exposed to the rainy season presenting a higher instantaneous probability of calving compared to those in the dry season (p < 0.001). The variance of the sire effect was estimated at 0.206 on the log-hazard scale, corresponding to an approximate heritability of 0.33 for sexual precocity under this time-to-event definition. In contrast, the binary P30-based analysis yielded a lower heritability estimate (h² ≈ 0.18), indicating a substantial loss of genetic information when timing is ignored. Moreover, the correlation between predicted genetic values from the two approaches was moderate, with noticeable re-ranking of sires. This re-ranking was especially pronounced among sires in the extreme quartiles of genetic merit, potentially leading to different selection outcomes depending on the methodology adopted. Overall, these results demonstrate that Survival Analysis captures additional genetic variation by explicitly incorporating both time to event and censoring. Therefore, although both methodologies are operationally feasible, the time-dependent approach provides a more informative and biologically meaningful representation of reproductive performance, with important implications for the genetic evaluation of sexual precocity in breeding programs. Further studies using this framework are currently underway to expand its applicability and robustness.

Keywords: 2026

How to Cite:

Caetano, S., Munari, D., Alvares, D., Ramos, S., Rey, F., da Silva, R. & Barbosa, L., (2026) “Modeling age at first calving in Nelore cattle using Survival Analysis”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2286881. doi: https://doi.org/10.31274/wcgalp.24154

Rights: 1