A simple method to test the reproducibility of the phylogenetic reconstructions: the molecular systematics of cyanobacteria as a case study

Abstract

Molecular systematics uses currently available data to produce the best approximation to the true (un-observable) phylogeny of a taxon. Molecular phylogeny complements morphological identification and classification of organisms, in order to infer their evolutionary relationships. In the current era dominated by cultivation-independent surveys, testing the potential technical and analytical pitfalls and limitations of environmental DNA surveys appears crucial. Sequence-based phylogenetic reconstructions rely on three main steps: alignment, alignment curation and tree building. Several independent options and settings can be adopted at each step, but it is well known that their choice (or combination) can significantly affect the topology of the phylogenetic tree obtained and skew the reliability of the resultant systematics. For the present study, five alignment algorithms, two curation options and three tree-building methods were used to infer the phylogeny of three orders of cyanobacteria, based on four validated markers widely used for this phylum: 16S rRNA, 16S-23S ITS, cpcBA-IGS and rpoC1. Compared to the alignment algorithm or the curation stringency used, the tree-building method was found to have the greatest effect on the resultant tree topology. This result was consistent for all loci, including the genetically-constrained (protein-coding) locus rpoC1. The reproducibility of the tree topology was clearly visualized and measured for each locus. This paper presents pitfalls in cyanobacteria systematics and implements a simple and rapid method, applicable to any locus and organism, to identify aberrant results and assess the reproducibility of phylogenetic reconstructions.