My research focuses on the dynamics of multilevel selection, or systems in which natural selection acts simultaneously at more than one level in a biological hierarchy. Any population evolves if variation in heritable traits affects the reproductive success of its members. This is equally true whether the population consists of organisms or any other reproducing entity. Life is hierarchically organized, and most reproducing entities contain collections of smaller reproducing entities. Examples include genes, chromosomes, organelles, cells, organisms, colonies, and  mixed-species communities. As a result, selection can act simultaneously at more than one level of organization. Although each level of selection can be conceptualized as a separate process, they often interact in complex ways that can dramatically affect evolutionary outcomes.

Multilevel selection theory can be a useful tool for studying the evolution of any trait involving cooperation or conflict. Specific examples include problems in: intragenomic conflict, ontogeny, cancer, senescence, sexual reproduction, pathogen virulence, resource exploitation, interference competition, mutualism, social cooperation, and transitions in individuality. I typically study multilevel selection using agent-based computer models to generate the dynamics of interest, and the mathematical models to analyze and interpret those dynamics. In addition to refining multilevel selection theory, I am also working on applying it to several specific problems, as outlined below.

What are the evolutionary dynamics of cancer?
Cancer biologists now recognize that cancer is the end stage of a long and complex evolutionary process within the body of a single individual. During an individual’s lifetime somatic cells gradually accumulate mutations, with selection among cell lines always favoring high rates of cell division and survival. If unchecked this process eventually generates cells that proliferate out of control. On a longer time scale however, selection among individuals shapes developmental patterns that limit the potential for harmful somatic evolution. In collaboration with Carlo Maley at the Fred Hutchinson Cancer Research Center in Seattle, I am using agent-based computer models to investigate the interplay between these two levels of selection. In particular, we are studying how normal cell differentiation patterns block somatic evolution, and how certain critical mutations can change these patterns, triggering an evolutionary process that ultimately results in cancer.

How do genetic systems evolve to become more evolvable?
Biologists have long assumed that natural selection inevitably results in adaptation. More recently it has become clear that the relationship between selection and adaptation is less simple. Research in evolutionary computation has shown that some systems readily evolve solutions to the problems posed by their environments, while others generate little adaptation even after very long periods of mutation, recombination, and selection. These observations raise the issue of “evolvability”, or why some systems readily evolve adaptation while others do not. They also suggest that living organisms’ proficiency at adaptation may itself have evolved. If so, the process that has generated evolvability is not selection among individuals, but rather a slower process of selection among lineages. I am studying a novel example of this process, in which chromosomal linkage patterns evolve over long time scales to increase the adaptive response to recombination. Results from computational models show that linkage patterns affects how well a population can adapt. More importantly, they also show that over long time scales evolution can reorganize chromosomes into a more evolvable architecture.

How can cooperation evolve among unrelated organisms?
Most existing theory for the evolution of cooperation involves either interactions among relatives (kin selection), or flexible behavioral strategies based on memory of past encounters (reciprocal altruism). In nature though, cooperation is common among unrelated organisms with little or no cognition. Together with a number of collaborators, I am studying a novel mechanism for the evolution of cooperation under such conditions, in which feedback between the cooperative behavior and local environmental quality leads to spatial clustering of cooperators. This mimics the clustering of genetic relatives underlying kin selection, and drives the evolution of cooperation by a parallel process. Results from computer and mathematical models suggest that this could be a fairly general mechanism for the evolution of cooperation in nature.

What drives transitions in individuality?
Repeatedly during the history of life, existing organisms have been assembled and integrated to create a new “compound” organism with novel functional organization. The products of such transitions include eukaryotic cells, multicellular organisms, social insect colonies, and integrated multi-species communities such as lichens. The evolutionary dynamics of these transitions are not well understood. Standard evolutionary theory is based selection among individuals, but during a transition the very quality of “individuality” shifts from one level of organization to another. The theory of multilevel selection provides a powerful tool for investigating the dynamics that cause the focus of selection to shift to a higher level, thereby producing new levels of adaptation and functional integration.

How do allometric scaling laws affect the structure and function of ant colonies?
One example of a transition in individuality is the integration of individual insects into highly organized colonies.  Here individuals coordinate their behavior and physiology to create a new level of biological organization with its own emergent informational and metabolic functions.  Allometric scaling laws describe many ways that body size influences the structure and function of individual organisms.  Additional problems of scaling arise as individuals aggregate into colonies.  How do critical structures and functions of colonies vary with colony size, and how do these scaling relations affect the ecology and evolution of group-living organisms? I am working with colleagues from the Santa Fe Institute and the University of New Mexico to develop a mathematical model that addresses how metabolism and resource acquisition scale with colony size, how foraging activities are organized to maximize resource uptake, and what limits the size of colonies and the areas that they forage.