R.F. Veerkamp and Y. de Haas (eds)
Proceedings of 12
th
World Congress on Genetics Applied to Livestock Production (WCGALP) 2031
DOI: 10.3920/978-90-8686-940-4_489, © V.B. Ambike et al. 2022
489. Random regressions for modelling semen production traits in
HF purebred and crossbred bulls using Bayesian framework
V.B. Ambike
1*
, R. Venkataramanan
2
, S.M.K. Karthickeyan
1
, K.G. Tirumurugaan
3
, K.G. Bhave
4
and
M. Swaminathan
4
1
Madras Veterinary College, Chennai 600 007, India;
2
The office of Vice-Chancellor, TANUVAS, Chennai
600 051, India;
3
Zoonosis Research Laboratory, Chennai 600 051, India;
4
BAIF, Pune 412 202, India;
vrindaambike@gmail.com
Abstract
Semen traits including ejaculate volume (EV, ml), sperm concentration (SC, 10
9
/ml), initial sperm
motility (ISM, %) and post-thaw motility (PTM, %) were analysed using random regression models. Data
pertaining to 105,321 ejaculates from 297 Holstein-Friesian (HF) purebred and crossbred bulls collected
from Bharatiya Agro Industries Foundation were utilized. 200,000 Gibbs samples were generated with a
burn-in of 20,000 and thinning interval of 50 in a Bayesian framework. Legendre Polynomials with orders
of fit up to five for additive and permanent environmental effects were used. Heritability estimates ranged
from 0.2 to 0.5 for EV and 0.35 to 0.5 for SC in purebreds while it ranged from 0.05 to 0.45 for EV and 0.2
to 0.7 for SC in crossbreds. Estimates were very low for motility traits in both purebred and crossbred bulls.
Random regression models in a Bayesian framework were useful for obtaining solutions for mixed models
even with large number of parameters.
Introduction
e HF breed has been extensively used for crossbreeding to increase milk production. Since most of the
selection is for milk yield, it is important to assess the semen parameters periodically to maintain optimum
fertility. e Bayesian framework provides the advantage of robustness with less stringent assumptions
about the distribution of parameters.
Random regression models have the ability to model the covariance of repeated data (longitudinal data)
measured on the same animal through covariance structures. is ability to model a trait with age as a
control variable helps in understanding the trajectory of various genetic parameters over the age of the
bulls. Worldwide there are very few studies conducted using random regression models for semen traits in
cattle (Al-Kanaan et al. 2015; Carabaño et al. 2007). e present study was carried out to understand the
genetic variability of various semen traits over age the age of the bulls in HF purebred and crossbred cattle.
Materials & methods
Data pertaining to 105,321 ejaculates from 120 purebred and 177 crossbred HF bulls collected from
Bharatiya Agro Industries Foundation (BAIF) were utilized for this study. e BAIF frozen semen stations
are located at Uruli Kanchan (Maharashtra, India; 18.5°N and 73.8°E) and Dharouli (Haryana, Jind; 29.2°N
and 76.2°E). Semen production data were available for 11 years (2010 to 2020) along with the pedigree
information for 1,317 bulls. e founder population was imported between 1967-1974 from Canada, the
USA, Denmark, and the Netherlands, where these bulls were proven through progeny testing.
Traits and factors included. Semen production traits included in the study were ejaculate volume (EV,
ml), sperm concentration (SC, 10
9
/ml), initial sperm motility (ISM, %) and post-thaw motility (PTM, %).
e fixed effects included in the model were location (Pune or Jind), breed of bull (purebred or crossbred),
order of ejaculate (first or second), the season (summer, monsoon, or winter), the year (2010 to 2020), age
https://www.wageningenacademic.com/doi/pdf/10.3920/978-90-8686-940-4_489 - Saturday, February 11, 2023 1:16:00 PM - IP Address:54.243.9.176