Modelling growth curves and GxE in beef cattle and sheep using random regression models (PhD)

The Challenge

Beef and sheep production in the United Kingdom contributes significantly to wider agricultural production for the nation, and ultimately for the world. As human population sizes have increased, urbanization has progressed, and available resources have diminished the demand on the industry to produce beef and sheep products faster and more efficiently has grown. Of course, efficiency is not limited, simply, to yield or weight. Rather, producers must consider components as various as: environmental impact, longevity, fertility, welfare, long-term health, and feed efficiency. Simultaneously the producer must balance their own costs to make long term economic gain. This is no small task. Historically, research has aided the producer by identifying the best breeding stock to optimize the genetic material of a producer’s flock or herd. In order to do this, recorded weights are plotted as a function of age to identify the factors impacting an animal’s growth. Traditionally, the repeated measurements for weight were analyzed with the assumption that the mean and covariance structure remains the same over time. Then evaluations were analyzed with the assumption that each weight was a genetically different, but correlated trait. These models result in inaccurate selection criteria because they neglect a great deal of random variability. Additionally, they require the producer to take weight measurements within fixed, possibly inconvenient, timeframes.

The Project

My project seeks to solve these issues by using a Random Regression Model. This model will represent the genetic and environmental covariance structure of growth traits in beef and sheep by fitting various growth curves. By capturing a more accurate representation of growth and including a wider range of data (i.e. economic data, maternal environment, carcase measurements) this project will result in better selection criteria. Thus, producers will be able to correctly identify superior breeding stock and optimize growth and economic returns.

Student

Emma Mutch, SRUC