Michael D. Nowak

Coffea

Generally, I am interested in the evolutionary processes that generate diversity both within and among species in plants. My dissertation research will test a series of hypotheses about the factors involved in lineage divergence and breeding system evolution of self-incompatible species presumed to have colonized islands in the Indian Ocean via long-distance dispersal (LDD). The self-incompatible genus Coffea (Rubiaceae: Coffeeae) in Madagascar and the surrounding islands shows a clear history of repeated island colonization and will be used as a model system in this research. The results of my research will be of general significance to the population genetics of island radiations and the evolution of self-incompatibility (SI) systems in scenarios of LDD. Intensive genetic sampling and field experiments of island and mainland populations of Coffea will be employed to test for the role of ecological factors in lineage divergence on islands, and test for the breakdown of SI coincident with island colonization.

Simulation of Fossil Data and Divergence Time Estimation

I am working in collaboration with Dr. David Hearn on a simulation study of the biasing effects of fossil constraints on molecular divergence time estimates. Fossils inherently exhibit only a fraction of the morphological characters that define extant clades. This means that the necessary assignment of a fossil taxon to an extant clade is often based on very few (or one) characters. Of course, any tree describing the relationships of extant taxa is naturally missing extinct lineages that likely represent a majority of the diversity through time. We hope that perhaps the extant tree represents a suitable proxy for the true tree, but we have no way of assessing this. We do know that the process of fossilization occurred on the true tree, and that the unique character state combinations that fossils often represent can provide significant insights into the morphological evolution of a group. One would therefore expect that the potential for erroneous placement of a fossil taxon on an extant reconstructed tree is elevated when there is considerable extinct diversity in the true evolutionary history of the group. Erroneous placement of a fossil taxon that is to be employed as a temporal constraint can lead to significant errors in the divergence time estimates of all nodes in the analysis. Our simulation mimics the process of fossilization on true trees (i.e. extant and extinct lineages) and subsequently uses these fossils to constrain nodes in a reconstructed extant tree to estimate divergence times.

Constructing Informative Divergence Time Priors

Under the mentorship of Dr. Derrick Zwickl, I am working this summer (2007) on a NESCent (National Evolutionary Synthesis Center) Phyloinformatics Summer of Code (sponsored by Google) software development project, which will employ fossil occurrence data from the Paleobiology Database to construct meaningful divergence time priors for application in the software package BEAST.

PUBLICATIONS