Grazing personality genetics of beef cattle in New Zealand rangelands : A thesis submitted in partial fulfilment of the requirements for the Degree of Doctor of Philosophy at Lincoln University

Abstract

The uneven distribution of grazing cattle on pastures and rangelands has been of concern to livestock managers since the very early days of grassland science, not least because of the effects of grazing patterns on ecosystem functions and the sustainability of farming systems. Mountainous terrain imposes additional limitations for free-range grazing animals; which may avoid using vegetation at higher altitudes, on steeper slopes, or at greater distance from drinking water. Consequently, areas that are more easily accessible might be overgrazed, potentially leading to diminished ecological functions and reduced productivity. In recent years, animal personality theory has suggested that individual animals do not behave in the same way, and instead display consistent and distinctive sets of behaviours or ‘personality’. Animal personality could explain the distinct grazing patterns reported for free-range cattle, where individual animals have preference for certain habitats over others as a result of their behaviour. Preliminary studies have also reported associations between cattle gene regions (i.e., quantitative trait loci) and indexes that describe terrain use, suggesting the potential for genes that might explain variation in the grazing personality of beef cattle.

This thesis contains the following chapters:

A literature review about grazing behaviour and personality (Chapter 1). While grazing lands can offer a diverse range of forages, individuals within herds appear to prefer to graze some habitats and not others. They can have consistent differences in grazing patterns and occupy specific spatial domains, whilst developing tactics and strategies for foraging that are individual specific. Accordingly, in this chapter, a new understanding of grazing personality was developed. This entailed the development of a ‘grazing personality model’ (GP-model) that accounts for the personality of individual animals and for the collective behaviour of herds. The GP-model postulates that the grazing personalities of ruminants and other large herbivores are determined genetically and tempered epigenetically in interaction with the social and biophysical environments of the cattle. They may also reflect the emotional state of animals.

While the selection of one grazing personality may be adequate for homogeneous pastoral systems, the design of herds with a range of grazing personalities that are matched to the habitat diversity may be a better approach to improving the distribution of grazing animals, thus potentially enhancing ecosystem services and maximizing productivity.

An investigation of whether the movement of cattle and potential measures of their grazing personalities might be determined genetically, was undertaken in chapter 2. Genetic variation within the glutamate metabotropic receptor 5 gene (GRM5 ), a ‘grazing gene’ candidate was investigated. Associations between variation in that gene and variation in grazing personality behaviours (GP-behaviours) were tested with mature cows (n = 303) under free-range management during winter grazing in the steep and rugged rangelands of New Zealand. Grazing behaviours were calculated using data from global positioning system (GPS) tracking collars and, satellite-derived data. Eight GP-behaviours were fitted into mixed models to ascertain their associations with variant sequences and genotypes of GRM5 . Three new GRM5 variants (A, B and C ) were discovered and six possible genotypes were identified in the cattle studied. The mixed models revealed that A was associated (P < 0.05) with elevation range, home range and movement tortuosity. Similarly, GRM5 genotypes were significantly associated (P < 0.05) with home range and movement tortuosity, while trends towards association (P < 0.1) were revealed for elevation range and horizontal distance travelled. Most of the GP-behaviour models were improved when corrected with the ‘cow age-class’ factor and the results suggested that grazing personality might be stable when cows reached 4 years of age. Home range and movement tortuosity were not only associated with GRM5 variation, but also negatively correlated with each other (r = -0.27, P < 0.001). Thus, there seems to be a genetically determined trade-off between home range and movement tortuosity that may be useful in beef cattle breeding programmes that aim to improve the grazing distribution and utilisation of steep and rugged rangelands.

The results of Chapter 2 suggested that differences in grazing patterns are associated with nucleotide sequence variation in GRM5 . Association analyses require large datasets to detect genotype-phenotype associations, hence, most large-scale studies aiming to identify behavioural linkages with grazing genes typically apply random sampling from existing setups without a priori control over the genotypic composition of the sample. This can lead to unbalanced experiments with over or under representation of any given group analysed. An alternative approach (Chapter 3) was used to perform a discriminant analysis of a balanced dataset that was generated by under-sampling the larger dataset (n = 303) described in Chapter 2. In this analysis, a training dataset of mature cows (n = 80) that equally represented five of the six GRM5 genotypes and four farms were selected. The GP-behaviours were derived from 5-min GPS relocations measured over 15 d and the analysis aimed to select a combination of GP-behaviours that assist the identification of specific GRM5 genotypes, and to investigate behavioural differences between cows of various genotypes of this ‘grazing gene’. Two sets of grazing behaviours were selected to build quadratic discriminant models (QDMs) that achieved 87% of accuracy in ascertain GRM5 genotype with a training balanced dataset. An ‘exploration discriminant model’ built with the GP-behaviours related to elevation, slope and exploration correctly predicted the genotypes of 85% of the individuals of a testing dataset that were not included in the model’s training. MANOVA and ANOVA analyses highlighted the relative importance of GP-behaviours to discriminate between GRM5 genotypes and showed behavioural differences between cows of various GRM5 genotypes. The results extend the list of key behaviours linked to GRM5 in agreement with genotypephenotype associations between GRM5 and GP-behaviours reported in chapter 2 and in the literature. In conclusion, sets of key GP-behaviours might be useful for predicting the variation in putative ‘grazing genes’ and QDMs applied to small-scale experiments with balanced designs seems to be a promising approach for behavioural genetics.

Overall, this research proposed a model for individual and collective grazing personalities for cattle. The analysis suggested consistent differences between individuals associated with GRM5 variation. Furthermore, linkages between bovine GRM5 and key grazing behaviours may characterise specific genotypes and assist with their identification. The research provides a conceptual model of grazing personality and experimental evidence suggesting possible applications of behavioural genetics to potentially optimise the distribution of beef cattle in steep and rugged terrain. More research is however needed to validate these findings.