Repliclade : a simulator of sequence evolution in gene families

Abstract

An ever-increasing need in phylogenetic inference entails selection of suitable analytic methods for a given lineage. It also includes identifying appropriate evolutionary models to perform forwards (or backwards) simulation on that lineage. These analytic methods and evolutionary models are prone to error and are known to bias phylogenetic simulation results. This can lead to erroneous results in an inferred phylogeny or phyletic simulation. It is of paramount importance to allow for selection of multiple methods of sequence simulation to perform phylogenetic inference with minimal prejudice. In this thesis, I present Repliclade. Repliclade is a software package written in the Python programming language that utilizes statistical methods to evaluate and compare models of molecular evolution and estimate values for associated parameters to facilitate robust phylogenetic inference on a monophyletic lineage of interest. It will then execute a simulation of a single inferred ancestral sequence generated via the Coalescent Theory once various crucial sequence parameters have been extracted. A phylogenetic reconstruction of the derived sequences are generated as a result of the simulation. The resulting progeny sequences and performance metrics along with numerous methods of phylogenetic reconstruction will provide insights into the efficacy of competing evolutionary models and methods of inference implemented by other software packages. These insights will assist researchers in identifying the strengths and weaknesses of varying approaches to simulation of an ancestral sequence and phylogenetic estimation. By corollary, Repliclade addresses an important and currently unmet need in the world of phylogenetic inference for informed model selection and sequence parameter extraction during DNA sequence analysis.