In-depth phylogeny: methods and applications in evolution

Phylogeny is a quest for evolutionary clues. The aim of this module is to recall the existence of gene phylogenies within species phylogenies, the ways in which evolutionary histories can be represented in tree form, and the challenge of positional molecular homology through sequence alignment. The principles of phylogenetic inference methods are at the heart of this course. Distance methods highlight the difficulties of separating homology and homoplasy, and the need to build models of character evolution. The maximum parsimony cladistic approach illustrates the use of bootstrapping to estimate the strength of phylogeny nodes, and the impact of taxonomic sampling in detecting multiple substitutions.

Probabilistic approaches are presented and explored in greater depth. The attraction artifact of long branches leads to an introduction to probabilistic reasoning. The maximum likelihood method is used to calculate likelihood, to estimate model parameters by optimality, to construct different character evolution models, and to compare models. Bayesian inference introduces the distinction between density-based and optimality-based approaches. It then shows the a priori use of probability densities, the data-driven estimation of a posteriori distributions of model parameters, their approximation by Markov chains with Monte Carlo techniques and Metropolis coupling (MCMCMC), the ignition and convergence phases, and the calculation and interpretation of tree and clade posterior probabilities. The importance of DNA, RNA and protein sequence evolution models and their improvement is emphasized.

"In addition to the above description, knowledge also concerns applications of phylogenetic inference methods: analysis of matrix representation with parsimony of source trees to infer supertrees, trait conservatism, phylogenetic inertia, phylogenetic diversity indices, estimation of ancestral character states, comparison of trees and evolutionary scenarios, and co-phylogenies with evaluation of co-evolution and co-speciation hypotheses.

Skills include bioinformatics application of the above concepts: data assembly (reads, sequences); sequence alignment; tree inference using distance methods, maximum parsimony, maximum likelihood, and Bayesian inference, with detection of the attraction of long branches; estimation of model parameters; model comparison; and use of multi-gene corroboration. Software to be handled includes SeaView, NJPlot, PhyML, IQTree, PAUP, and MrBayes."

"Optional: take the M1 S7 course "Phylogeny and Evolution" (HAB708B).

Recommended: Sequence alignment; tree reading and reconstruction: distances, cladistics and parsimony, probabilistic approaches (likelihood)."

100% continuous assessment