Dr. Mark Holder

My research revolves around statistical phylogenetics and its applications to evolutionary biology. In particular, I focus on Bayesian techniques for inferring phylogenies. I have contributed to development of Markov chain Monte Carlo methods used to implement Bayesian tree inference, but my primary interest is in the evolutionary models and prior assumptions that underlie these methods. Improvements to models allow us to estimate trees more accurately and assess the error in our estimates. More importantly, the development of richer models lets us use the comparative approach to a wide range of biological problems. Current Activities/Research Program Currently, I am working on a collaborative effort to improve the techniques available for multiple sequence alignment. My research group, along with collaborators at the University of Texas, University of Nebraska, University of Georgia, and Penn State University, will focus aligning sequences for the purposes of phylogenetic analysis. In particular, we will try to extend the realm of data set sizes for which it is feasible to use methods that simultaneously align sequences while searching for trees that best explain the data. The focus of the work here at KU will be on fast ways to approximate the maximum likelihood estimate of a phylogeny and history of insertions and deletions. In the next phase of my research program, I will be building on the emerging field of context-dependent evolutionary models. Most phylogenetic models of sequence evolution make the unrealistic assumption that different sites evolve completely independently of each other. Building on recent Markov chain Monte Carlo techniques (Jensen and Pedersen, Advances in Applied Probability, 32, 2000), researchers have begun to explore models that consider constraints on the entire sequence. For example, the requirement that a protein must fold into a particular three-dimensional structure in order to function, constraints the amino acids that are allowed in a sequence. A mutation in one site may change the state-space of residues allowed at its neighbor (or an interacting site in the folded configuration). Initial efforts to construct phylogenetic models to explicitly accommodate the influence of protein tertiary structure (for examples see Robinson et al., Molecular Biology and Evolution, 20, 2003; Rodrigue et al., Gene, 347,2005; but also see Thorne et al., Molecular Biology and Evolution, 24, 2007). My work will focus on modeling the constraints on protein evolution more accurately. I am also interested in applying this class of context-dependent model to the analysis of morphological character evolution.