Towards accurate avian tree of life

Species tree reconstruction was long ago based on morphological characters. In ۱۹۹۰s, a few universalorthologous genes (gene markers) including the small and large subunit ribosomal RNAs were utilised forthis purpose. This built the foundation of molecular phylogenetics. However, thanks to the advent of DNAsequencing technologies, sequence data are being widely used for species tree inference. Thus, nowadayshundreds of gene markers could be used in this regard. Nevertheless, inferring these marker loci iscomputationally demanding and requires complicated pipelines to extract the single copy ortholog groups(OGs). Recently, a new phylogeny for a dataset of ۳۶۳ birds has been proposed. Here, we discuss the speedand accuracy of their approach compared to the NCBI taxonomy using the Robinson–Foulds metric. Besides,we present our fast method for tree inference for arbitrary species sets. In this method, the OMAmer softwarewas used to place proteins of species of interest onto a database of hierarchical orthologous groups fromOMA. Then, each protein is mapped to an OG if possible. Next, the most informative OGs are selected forwhich multiple sequence alignment (MSA) matrices are computed. Finally, the species tree is inferred usingIQ-tree using the super-matrix. The super-matrix of orthologous characters was computed within ۵۰ CPUhours and IQ-Tree needs ۱۱h on ۴۸ CPUs to infer the avian tree.

Keywords:

Comparative genomics , Species tree , NCBI taxonomy , Orthologous groups