Debugging collective digging in Drosophila melanogaster using an agent-based model

Abstract

Being a member of a group entails a range of possible costs and benefits; observing when and with whom groups form can help identify how these are weighted. Clustering behaviour in fruit fly (Drosophila melanogaster) third-instar larvae has been shown to provide individuals that cluster with improved foraging opportunities on otherwise inaccessible food. To isolate how environmental and genetic factors affect who clusters and when, we created an on-lattice agent-based model and systematically explored the effects of several variables, such as population density, resource distribution, and the phenotypic composition of the group. Our simple model reproduces most of the key features of larval clustering and demonstrates drastic differences in collective behaviour and fitness in response to changes in environmental variables, in ways that align with the limited empirical data available. This model can improve our understanding of this complex social behaviour and make testable predictions about the behavioural and ecological mechanisms of larval clustering.