Insect-virus relationships: sifting by informatics.

Several groups of large DNA viruses successfully utilise the rich resource provided by insect hosts. Defining the mechanisms that enable these pathogens to optimise their relationships with their hosts is of considerable scientific and practical importance, but our understanding of the processes involved is, as yet, rudimentary. Here we describe an informatics-based approach that uses comparison of viral genomic sequences to identify candidate genes likely to be specifically involved in this process. We hypothesise that such genes should satisfy two essential criteria, namely, that they should be (i) present in those members of a virus family that infect insects, but absent from those that infect other hosts, and (ii) found in at least two unrelated taxa of insect viruses. These criteria currently identify six groups of viral genes, including one that encodes the fusolin/gp37 proteins. Demonstration that the fusolin/gp37 proteins can enhance oral infectivity of insect viruses provides a primary validation of this approach to the examination of insect-virus relationships.