Adsa Logo White Adsa Title White

Estrus prediction model for dairy Gyr heifers.

L. El Faro Zadra

Events

06-23-2020

Join L. El Faro Zadra on this page for a live text chat!
6:00 PM - 8:00 PM GMT

Abstract:

T15
Estrus prediction model for dairy Gyr heifers.
V. Vilela Andrade1, P. Arrigucci Bernardes2, R. Ribeiro Vicentini3, A. Penido Oliveira4, L. El Faro Zadra*1. 1Instituto de Zootecnia Sert�ozinho, SP, Brazil, 2Universidade Federal de Santa Catarina Florian�polis, SC, Brazil, 3Universidade Federal de Juiz de Fora Juiz de Fora, MG, Brazil, 4Empresa de Pesquisa Agropecu�ria de Minas Gerais Uberaba, MG, Brazil.

In view of the particularities of the dairy Gyr breed in terms of estrous behavior and of the importance of efficient estrus detection, precision technologies can assist in the monitoring of estrus-related parameters. This study aimed to evaluate variations in core body temperature using continuous data of reticulorumen temperature (RRT) and animal activity (ACT) during estrus of dairy Gyr heifers and to use these measures in prediction models. The animals were submitted to an estrus synchronization protocol. The data were obtained with bolus instruments administered in the reticuloruminal cavity of 45 heifers, which recorded information at intervals of 10 min. The mean RRT and ACT were compared at different time intervals in relation to estrus (established by visual observation of mounting acceptance in the field). Mixed models were used for ANOVA of the 2 variables. Logistic regression, random forest and linear discriminant analysis were tested as prediction models using RRT, ACT, time of day and temperature humidity index (THI) as predictors, and the presence and absence of estrus as the variable to be predicted. The mean RRT during the period corresponding to estrus were higher (P < 0.05) than the mean temperatures on the day before (+0.22�C) and after estrus (+0.36�C) for all times studied. These results can be attributed to the hormonal changes that occur during this period. The mean ACT was higher (P < 0.05) for all times after removal of the progestogen implant compared with the times of the previous day, which can be explained by the behavioral changes characteristic of this phase. Among the prediction models, the random forest model provided the best mean performance values using RRT, ACT, time of day and THI as predictors. The ability to correctly predict the period of estrus (sensitivity) was 51.69% and the ability to predict the period of non-estrus (specificity) was 93.1%. The accuracy (ratio between the number of events correctly classified by the total number of events) was 0.87. Changes in RRT and ACT occur during estrus, which are important variables to be included in prediction models.

Keywords: sensors, body temperature, Zebu.

Biography: AcknowledgementsLenira El Faro is Animal Scientist, with MSc in Genetics and Animal Breeding by Paulista State University Julio de Mesquita Filho and DSc in Animal Science by the same institution. Currently, she is a researcher at Institute of Animal Science of S�o Paulo State and professor at the Graduate Program in Sustainable Animal Production of the same institute. Lenira has interest on new phenotypes, dairy cattle, quantitative genetics, genetic evaluation and genetic parameters.