Comparative and Genetic Approaches to Placental Evolution and Disease

The placenta is an important locus of theory and empirical research in mammalian evolution, physiology, health and disease. Novel comparative approaches to the study of placentation hold great promise in bridging the gap between applied placental research and evolutionary theoretical biology, potentially providing insights into intractable medical conditions affecting the placenta in human beings. This thesis describes genetic and comparative approaches designed for the study of placentation, but which will also prove broadly useful in research at the intersection of human health and evolutionary biology. The thesis begins with a comprehensive investigation into the historical course of evolution of the eutherian placenta, with special focus on identifying the polarity of transformation of interhemal relations, fetal-maternal interdigitation and shape. A range of statistical approaches appear to concur on an early origin of invasive, hemochorial placentation and the existence of repeated independent transitions toward less invasive forms. Tests for positive selection, and assessment of positively selected genes for substitutions of major phenotypic effect, are used to identify genes involved in the evolution of spiral arteries at the origin of the great apes and in the evolution of reduced placental invasion in three independent branches of the euarchontogliran phylogeny. It is shown to be possible to prioritize such genes for investigation into their involvement in diseases of placental vasculature including preeclampsia. The thesis continues with elaboration and discussion of statistical models for the evolution of biological traits that are known to deviate from the neutral, gradualistic assumptions of standard approaches - such as independent contrasts and phylogenetic generalized least squares - that are based on a Brownian motion model of evolutionary change. First, I discuss the use of stable models of continuous character evolution and provide a methodology for estimating ancestral states and characterizing the evolutionary process operating on traits exhibiting occasional rapid bursts of change. Second, I discuss the incorporation of directional tendency into phylogenetically independent contrasts. Simulation studies and application to real biological datasets suggest that such methods may be superior under conditions that deviate markedly from Brownian motion.