Spotted hyenas (Crocuta crocuta) are large mammalian carnivores, but their societies, called 'clans', resemble those of such cercopithecine primates as baboons and macaques with respect to their size, hierarchical structure, and frequency of social interaction among both kin and unrelated group-mates. However, in contrast to cercopithecine primates, spotted hyenas regularly hunt antelope and compete with group-mates for access to kills, which are extremely rich food sources, but also rare and ephemeral. This unique occurrence of baboon-like sociality among top-level predators has favoured the evolution of many unusual traits in this species. We briefly review the relevant socioecology of spotted hyenas, document great demographic variation but little variation in social structure across the species' range, and describe the long-term fitness consequences of rank-related variation in resource access among clan-mates. We then summarize patterns of genetic relatedness within and between clans, including some from a population that had recently gone through a population bottleneck, and consider the roles of sexually dimorphic dispersal and female mate choice in the generation of these patterns. Finally, we apply social network theory under varying regimes of resource availability to analyse the effects of kinship on the stability of social relationships among members of one large hyena clan in Kenya. Although social bonds among both kin and non-kin are weakest when resource competition is most intense, hyenas sustain strong social relationships with kin year-round, despite constraints imposed by resource limitation. Our analyses suggest that selection might act on both individuals and matrilineal kin groups within clans containing multiple matrilines.

Evolvability is the capacity of an organism or population for generating descendants with increased fitness. Simulations and comparative studies have shown that evolvability can vary among individuals and identified characteristics of genetic architectures that can promote evolvability. However, little is known about how the evolvability of biological organisms typically varies along a lineage at each mutational step in its history. Evolvability might increase upon sustaining a deleterious mutation because there are many compensatory paths in the fitness landscape to reascend the same fitness peak or because shifts to new peaks become possible. We use genetic marker divergence trajectories to parameterize and compare the evolvability—defined as the fitness increase realized by an evolving population initiated from a test genotype—of a series of Escherichia coli mutants on multiple timescales. Each mutant differs from a common progenitor strain by a mutation in the rpoB gene, which encodes the β subunit of RNA polymerase. Strains with larger fitness defects are proportionally more evolvable in terms of both the beneficial mutations accessible in their immediate mutational neighborhoods and integrated over evolutionary paths that traverse multiple beneficial mutations. Our results establish quantitative expectations for how a mutation with a given deleterious fitness effect should influence evolvability, and they will thus inform future studies of how deleterious, neutral, and beneficial mutations targeting other cellular processes impact the evolutionary potential of microorganisms.

We introduce a Bayesian approach to discovering patterns in structurally complex processes. The proposed method of Bayesian structural inference (BSI) relies on a set of candidate unifilar hidden Markov model (uHMM) topologies for inference of process structure from a data series. We employ a recently developed exact enumeration of topological -machines. (A sequel then removes the topological restriction.) This subset of the uHMM topologies has the added benefit that inferred models are guaranteed to be -machines, irrespective of estimated transition probabilities. Properties of -machines and uHMMs allow for the derivation of analytic expressions for estimating transition probabilities, inferring start states, and comparing the posterior probability of candidate model topologies, despite process internal structure being only indirectly present in data. We demonstrate BSI's effectiveness in estimating a process's randomness, as reflected by the Shannon entropy rate, and its structure, as quantified by the statistical complexity. We also compare using the posterior distribution over candidate models and the single, maximum a posteriori model for point estimation and show that the former more accurately reflects uncertainty in estimated values. We apply BSI to in-class examples of finite-and infinite-order Markov processes, as well to an out-of-class, infinite-state hidden process.