Groupe Keller

Evolutionary genetics and ecology of social life

The goal of our group is understand the principles governing the evolution of animal societies and the ecological and evolutionary consequences of social life. To study these questions we combine disciplines of animal behaviour, ecology, evolutionary genetics and genomics. Our current interests include:

There is tremendous variation in life span between castes in ants, with queens living over 500 times longer than males and 50 times longer than sterile workers. The fact that individuals of these different castes share the same genome and exhibit such enormous variation in lifespan allows us to ask fundamental questions about the evolution of lifespan with respect to the expression of several groups of genes implicated in the ageing process. We investigate the mechanisms that underlie the deterioration of an individual's body as well as the control of these processes at the molecular level. This will allow us to gain further insight into the physiology of ageing as well as shed light on its evolution.

The development of kin selection theory has been one of the major breakthroughs in evolutionary biology. However, kin selection has rarely been tested experimentally although such tests would not only permit validation of the Hamiltonian hypothesis of indirect gene transmission but also provide evidence for Darwin's concept of evolution by natural selection as a whole. We have worked on sex allocation in ants of the genus Formica, showing kin-selected worker manipulation of the secondary sex ratio. We are also investigating partitioning of reproduction between queens as well as between workers in several ants and an Australian Allodapine bee. Finally, in collaboration with Prof. Dario Floreano and Roland Siegwart at the EPFL (http://asl.epfl.ch/) we are investigating the role of relatedness (genetic similarity), levels of selection, and group size on the evolution of collective actions and division of labour in artificial ant colonies. We are using colonies of artificial ants implemented as small mobile robots with simple vision and communication abilities. Selection experiments are done with the help of genetic algorithms.

The number of queens within a social insect colony strongly influences opportunities for the evolution and maintenance of kin-selected behaviors, and thus is an important trait in social evolution Queen number can be highly variable both between and within species, and we are interested in how and why transitions in social structure occur. In collaboration with Dr. Kenneth Ross (University of Georgia), we have identified a gene (Gp-9) influencing social behaviour in the fire ant Solenopsis invicta. We are currently quantifying the genetic variability of this gene as well as other closely linked genes using a genomic approach. We are also interested in the ecological factors influencing the evolution of queen number in three polygyne species of wood ants. Another project is about the evolution and consequences of unicoloniality in the Argentine ant. This species has invaded the Mediterranean coast about 80 years ago and there is no aggression between workers from Spain, France and Italy

Kin selection theory predicts that whether a female ant develops into a sterile worker or a reproductive queen should be determined by environmental factors rather than by a genetic predisposition. However, some extraordinary cases of genetic determination of reproductive caste have been found in harvester ants and fire ants. In these systems, two separate genetic groups must interbreed to produce workers, but produce new reproductive queens only from within-group matings. We have been investigating the evolutionary origins of this phenomenon, which appears to be derived from past interspecific hybridisation events, using various molecular genetic markers and the consequences on the biology of the species.