Bayesian study of AFLP marker evolution in phylogenetic inference

Amplified Fragment Length Polymorphism (AFLP) markers are formed by selective amplification of DNA fragments from digested total genomic DNA. The technique is popular because it is a relatively inexpensive way to produce large numbers of reproducible genetic markers. In the dissertation, I describe two models for AFLP evolution based on an understanding of the genetic processes at the nucleotide level that directly cause marker gain and loss. One model assumes that nucleotide substitution is the only genetic mechanism affecting AFLP markers, and a second model, the Sub-ID model, involves both substitution and insertion/deletion processes. Novel Markov chain Monte Carlo methodologies specifically tailored for these models are built to implement Bayesian approaches to phylogenetic inference from AFLP marker data. We demonstrate the methods on species in Carex section Ovales, a group of sedges common in North America.; The Bayesian method based on the substitution-only model is the first approach specifically to model AFLP marker evolution, and has advantages over methods such as NJ and MrBayes in that it is closer to the genetic background and uses the available data more fully. Our method incorporates explicitly both the fragment length and marker presence/absence information. The Sub-ID model, incorporating both nucleotide substitution and insertion/deletion processes, relaxes the common assumptions used in the substitution-only model and MrBayes method that markers are independent and homologous. The Bayesian method based on the Sub-ID model allows us to infer both the phylogenies and the subset of markers that are possibly homologous. The inference of topologies was not sensitive to the prior settings and the Jukes-Cantor assumption for nucleotide substitution.