Page 137 - Linear Models for the Prediction of Animal Breeding Values 3rd Edition

8 Social Interaction Models

8.1 Introduction

Social interaction among animals, such as competition and cooperation, can have a profound
effect on the expressions of performance and welfare traits in domestic livestock popula-
tions (Muir, 2005; Bijma et al., 2007a). When a group of animals rely on a limiting resource
(e.g. feed) to achieve an outcome (e.g. growth) the observed phenotype of an individual
(e.g. growth rate) can be influenced by both the phenotype (e.g. ability to fight for food)
and the genotype (which confers this ability) of the competitors in the group. So the growth
rate of piglets, for instance, can be reduced due to competition for food. In laying hen pro-
duction systems, social interactions can result in mortality due to cannibalism when hens
are housed in groups, and this poses both economic and welfare problems.
Although a major component of the social interaction among group members may
appear to be environmental, there is a genetic component (Wolf et al., 1998) attributable
to the genes carried by others in the group which affects how they compete; generally
referred to as indirect genetic effects (IGE) (Cheverud and Moore, 1994; Moore et al.,
1997). A selection experiment to reduce mortality due to cannibalism in domestic chick-
ens (Muir, 1996) has shown that heritable interactions (or IGE) can contribute substan-
tially to response to selection. Selection schemes that ignore this social effect of an
individual on the phenotypes of its group members could result in less optimum response
or even response in the opposite direction (Griffing, 1967). This social effect or indirect
genetic effect (Cheverud and Moore, 1994) is often referred to as an associative effect
(Griffing, 1967). In addition, Bijma et al. (2007b) indicated that the existence of social
interaction among individuals may increase the total heritable variance in a trait. They
found that heritable variance in survival days expressed as a proportion of phenotypic
variance increased from 7 to 20% due to social interactions, indicating that about two-
thirds of heritable variation is due to interactions among individuals. One possible solu-
tion for improving traits affected by social interaction is to undertake group selection
(Griffing, 1967). However, an optimum individual selection scheme to improve traits
affected by interactions among individuals will involve the use of models that account for:
1. direct effects due to the direct effects of the genes of the individual; and
2. indirect effects due to the associative effect of the individual on its group members.
The phenotype (P ) of an animal i for a trait influenced by social interaction
i
belonging to a group with n members where interaction occurs may be modelled as:
n−1 n−1
P = A + Q + E D i ∑ A S j ∑ Q + E
+
+
i D i , D i , , , S j , S j ,
ji ≠ ji ≠
where j is one of the n − 1 group mates, A and A are the additive direct effect and
D,i S,j
sum of the additive indirect effects of each of the n − 1 group mates, with corresponding
© R.A. Mrode 2014. Linear Models for the Prediction of Animal Breeding Values, 121
3rd Edition (R.A. Mrode)