Statistical Phylogenetic Tree Analysis Using Differences of Means

摘要: We propose a statistical method to test whether two phylogenetic trees with given alignments are significantly incongruent. Our compares the distributions of by input alignments, instead comparing point estimations trees. This approach can be applied gene tree analysis for example, detecting unusual events in genome evolution such as horizontal transfer and reshuffling. uses difference means compare trees, after mapping vector space. Bootstrapping alignment columns then obtain p-values. To compute distances between means, we employ "kernel method'' which speeds up distance calculations when mapped high-dimensional feature space, e.g. splits or quartets In this pilot study, first our on data sets simulated under coalescence model, generated congruent follow simulation results applications various gophers lice, grasses their endophytes, different fungal genes from same genome. A companion toolkit, Phylotree, is provided facilitate computational experiments.