Prediction of genetic merit in Beetal goats using machine learning algorithms
Poonam Sevda, Yogesh C. Bangar, Sunil Kumar, Man Singh, Rakesh Nehra, Ankit MagotraContext
Accurate early prediction of growth traits is essential for guiding selection decisions and accelerating genetic improvement in livestock. In Beetal goats, 6-month body weight is an important indicator of future performance, yet estimating its genetic merit traditionally requires time-consuming and less flexible analytical methods.
Aims
This study aimed to evaluate the effectiveness of advanced machine learning (ML) algorithms in predicting the genetic merit (breeding value) of 6-month body weight in Beetal goats.
Methods
Data from 889 goats recorded over two decades (2004–2023) at the Goat Breeding Farm, Lala Lajpat Rai University of Veterinary and Animal Science, Hisar, Haryana, India, were used. The dataset included pedigree information, birth-related factors, dam characteristics and early growth records. Nine supervised ML algorithms, namely, gradient boosting machine, random forest, support vector machine, artificial neural network, Bayesian regression, Gaussian process, multivariate adaptive regression splines, sequential minimal optimization regression, and multiple linear regression were trained and tested. Model performance was compared using R2, root mean square error, mean absolute error, mean absolute percentage error and bias metrics.
Key results
Gradient boosting machine showed the highest predictive accuracy, yielding the greatest R2 (85.29%), and the lowest root mean square error (0.37), mean absolute error (0.28) and bias (−0.02). Random forest performed comparably well, whereas Bayesian regression and multiple linear regression produced the least accurate predictions. Six-month body weight was identified as the most influential predictor for breeding value.
Conclusions
Advanced ML algorithms, particularly gradient boosting machine and random forest, demonstrated strong potential for accurately predicting breeding values for 6-month body weight in Beetal goats.
Implications
The adoption of these ML tools can support earlier and more precise selection decisions, thereby enhancing genetic progress and improving productivity in goat breeding programs.