Deep mining of the Sequence Read Archive reveals bipartite coronavirus genomes and inter-family Spike glycoprotein recombination

Abstract

Genetic variation in RNA viruses is generated by point mutation and recombination as well as reassortment in the case of viruses with segmented genomes. While point mutation concerns only few sites per genome copy, recombination and reassortment can affect large genome regions, possibly facilitating the sudden emergence of novel traits. The contribution of recombination and reassortment to genomic plasticity and their rates remain poorly understood and might be underappreciated because of the lack of a comprehensive description of the virosphere. Here we employed a computational approach that directly queries primary sequencing data in a highly parallelized way and involves a targeted viral genome assembly strategy. By screening more than 213,000 data sets from the Sequence Read Archive repository and using two metrics that quantitatively assess assembly quality we discovered 25 novel nidoviruses from a wide range of vertebrate hosts. These include eight fish coronaviruses with bipartite genomes, a giant 36.1 kilobase coronavirus genome with a duplicated Spike glycoprotein (S) gene, and 16 additional so far undescribed vertebrate nidoviruses. Some of these novel virus genomes encode protein domains that have not been described for nidoviruses. We provide evidence for a possible inter-family homologous recombination event involving S between ancestral bipartite coronaviruses and unsegmented tobaniviruses and report a case example of an individual fish simultaneously infected with members from both virus families. Our results shed light on the evolution and genomic plasticity of coronaviruses and identify recombinants with a possibly improved ability to cross species barriers, which might elevate their pandemic potential.