Using Machine Learning Techniques for Phylogenetics

Abstract

Phylogenetics is the study of analyzing genes to infer evolutionary relationships among a set of species. Maximum likelihood and maximum parsimony are two typical methods for phylogenetic inference, but each of these two methods could perform well or badly for different phylogenetic tree types, resulting in inconsistent phylogenetic trees. This thesis trained a neural network to infer the phylogenetic tree type from a four-taxon alignment, which potentially provides insight into which tree reconstruction method is suitable to the alignment. Show more