When the Past Fades: Detecting Phylogenetic Signal with SatuTe

In phylogenetics, the phenomenon of saturation is well known, although its influence on tree reconstruction lacks a systematic and well-founded method. Here, we propose a new measure of the phylogenetic information shared between two subtrees connected by a branch in a phylogeny. This measure generalizes the concept of saturation between two sequences to a theory of saturation between subtrees, whose implementation we provide as the versatile program SatuTe. We describe different usages of SatuTe, identifying which branches in a tree are phylogenetically informative and which alignment regions support a given branch. As an example, we discuss the Tree of Life reconstruction from ribosomal proteins and the 16S rRNA gene, with emphasis on the two-domain versus three-domain hypotheses. For the branch leading to Eukaryota, we show that most ribosomal proteins contain a strong phylogenetic signal, whereas some regions of the 16S rRNA gene have lost phylogenetic information. Thus, SatuTe opens new insights into phylogenetic inference and complements standard phylogenetic analysis.