Models and Algorithms for Phylogenetic Marker Analysis

Phylogenetic markers are widely used in DNA barcoding, DNA taxonomy, and amplicon based metagenomics. In this thesis, we address problems with phylogenetic marker analysis by developing novel models and algorithms. We present the Poisson Tree Processes (PTP) model for species delimitation using single-locus phylogenetic markers. We develop and test algorithms for maximum likelihood inference under the PTP model, and extend the PTP model using a Bayesian framework. Further, employing the species delimitation method, we develop a new algorithm PhyloMap, for visualizing large phylogenetic marker data sets. We also describe and make available an accurate and robust paired-end reads merger for the Illumina Next-Generation Sequencing (NGS) platform. Finally, we integrate PTP with the Evolutionary Placement Algorithm (EPA) to delimit species in amplicon based metagenomic data.