Constructing the Relationship Tree of All Viruses via Whole Genomic Sequences

This paper aims to construct two relationship trees of all viruses using two types of genomic sequences, DNA("deoxyribonucleic acid") and CDS ("coding sequence"), respectively, via a previously developed approach BBRD (BLAST-Based Relative Distance). The BBRD approach is capable to construct the relationship trees of different genomic sequences without identifying common conserved regions among these sequences for comparison in advance. The experimental resources of viruses, with complete genome sequences, were downloaded from NCBI(National Center for Biotechnology Information) at 2018/3/1, and there are 7,535 viruses with whole DNA sequences and 7,434 viruses with at least one CDS sequences. Experimental results show that the relationship tree constructed via DNA sequences seems to be more consistent with the taxonomy of viruses in ICTV (International Committee on Taxonomy of Viruses) than that constructed via CDS sequences. Furthermore, observing the neighbors of one unknown virus within the relationship trees can provide hints to determine or guess its taxonomic information for the biologist or virologist. This study may inspect the fitness of the structures (skeletons) of one existing taxonomy, e.g. ICTV, by observing the relationship tree and providing the parts of subtree without consistence.