Algebraic statistics in phylogenetic reconstruction: incorporating invariable sites

In modern phylogenetics, when comparing sequences of DNA the traditional approach assumes that nu- cleotide substitutions follow a Markov model, that sites on a DNA sequence evolve independently and in the same way. However, recent studies have shown that some positions in DNA sequences remain invar...

Descripción completa

Detalles Bibliográficos
Autor: Mañez Fernandez, Davinia
Tipo de recurso: tesis de maestría
Fecha de publicación:2023
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/392846
Acceso en línea:https://hdl.handle.net/2117/392846
Access Level:acceso abierto
Palabra clave:Markov processes
Biomathematics
Phylogenetic tree
Markov process
flattening
invariable parameters
invariable sites
Markov, Processos de
Biomatemàtica
Classificació AMS::92 Biology and other natural sciences
Àrees temàtiques de la UPC::Matemàtiques i estadística
Descripción
Sumario:In modern phylogenetics, when comparing sequences of DNA the traditional approach assumes that nu- cleotide substitutions follow a Markov model, that sites on a DNA sequence evolve independently and in the same way. However, recent studies have shown that some positions in DNA sequences remain invariant. This thesis aims to investigate cases where certain regions of DNA sequences do not change throughout the evolutionary process. The approach, inspired by the work of Allman and Rhodes in [1], considers a model where a proportion of sites in the DNA sequences cannot vary, while the remaining sites are variable. This model is referred to as the general Markov model plus invariable sites (GM + I ). In [1] they obtain formulae to recover the called invariable parameters involved in the GM + I model. We implement a computational method based on this proposition using Python and evaluate its performance using simulated data. The performance of the method is analyzed in different situations, and potential improvements and variations based on our findings are suggested.