Solving the species problem in viral taxonomy: recommendations on non-Latinized binomial species names and on abandoning attempts to assign metagenomic viral sequences to species taxa.

Abstract

Properties useful for defining virus species are phenotypic properties of viruses that can be altered by a few mutations. Such properties include the natural host range, cell and tissue tropism, symptomatology, pathogenicity and mode of transmission. All these properties are not necessarily present in identical form in all the members of a species; therefore, a virus species is a polythetic class of viruses defined by a variable combination of several properties rather than by a single conserved property present in all the members of the species. This review will discuss current controversies about what virus species actually are as well as which names should be given to them. It will be emphasized that most species-defining properties are so-called relational properties that arise because viruses necessarily interact with biological partners such as vectors, hosts and immune systems. Although these relational properties are of utmost importance to laboratory and clinical virologists, they remain unknown if only the viral genome is available and the relational partners of the virus have not yet been identified. Since the International Committee on Taxonomy of Viruses (ICTV) in 2013 ratified a new definition of virus species, which no longer accepts that species are polythetic classes but instead are monophyletic groups, the implications of this new definition for viral taxonomy and nomenclature will be analyzed. In my private capacity, I also make the following recommendations regarding current debates on proposed new names for virus species as well as on the feasibility of assigning viral sequences found in metagenomic databases to individual species taxa in the current ICTV classification. 1) The ICTV should abandon the current rule that the names of virus species (for instance Measles virus) should differ from the virus name (measles virus) only by typography. 2) Non-Latinized binomial species names based on familiar virus and genus names should become the norm. This would obviate the need to create about 5000 hard-to-memorize Latinized species names. 3) Virus species are defined not by the intrinsic properties of virions and viral genomes but by the relational properties of viruses that arise from their interactions with host and vector partners. Since the hosts and vectors associated with nearly all viral sequences found in metagenomic databases are unknown, the phenotypic properties of the putative viruses also remain unknown, and these viral sequences cannot be allocated to established species in the ICTV classification.