Causes and consequences of social interactions in free-living female house mice (Mus musculus domesticus)

Social interactions are central to many organisms and influence behaviour and ultimately fitness. The social environment is considered to be the most complex and fluctuating component of an individual’s environment because of its high degree of flexibility and intrinsic unpredictability. The significance of social interactions therefore should best be studied in the animals’ natural environment, combining proximate and ultimate causation of social interactions. Since 2002 we study a natural population of house mice (Mus musculus domesticus) in a barn near Zurich. This study population is unique since mice can emigrate and immigrate, and can be individually followed from birth until they are found dead or leave the barn. We further use RFID technology to monitor social and reproductive behaviour, and combine this data set with pedigree information based on microsatellite analyses.Here, we intend to analyse social interactions in an evolutionary framework, using an integrative approach. Our goal is to understand how differences in the social environment lead to differences in social interactions that translate into fitness differences between individuals, analysed in the animals’ natural social environment. To do so, we will focus on female-female interactions and use three approaches.First, we use social network analysis to identify the complexity of social interactions. For interacting females we use social network statistics and map genetic and morphological traits to identify what best explains the social position in the network. This dataset further allows testing the importance of social networks for cooperation. We will empirically test the function of social interactions with a “knockout” experiment where we remove individuals with differing behavioural phenotypes according to network characteristics.Second, we will link behavioural variation at the individual level with underlying mechanisms of endocrine and neural nature. We will analyse steroid hormones incorporated in the animals’ hair and will analyse whether instability in the social environment results in a chronically stressful situation that influences a female’s and her daughters’ reproductive success. We will use the brains of the mice from our “knockout” experiment to analyse whether distribution differences in ligands and receptors of the nonapeptide system (oxytocin, vasopressin) are associated with differences in social phenotypes of wild female mice.Third, we intend to use genome-wide genotyping for quantitative genetic analyses to study the genetic architecture of naturally varying social traits observed in our study population. The very large number of mice, the systematic way in which we monitor the population, and the resulting richness of the data offers an extraordinary opportunity to quantify the sources of variation in life-history traits and, particularly, of social behaviours.