RECOVIR: An application package to automatically identify some single stranded RNA viruses using capsid protein residues that uniquely distinguish among these viruses

BackgroundMost single stranded RNA (ssRNA) viruses mutate rapidly to generate large number of strains having highly divergent capsid sequences. Accurate strain recognition in uncharacterized target capsid sequences is essential for epidemiology, diagnostics, and vaccine development. Strain recognition based on similarity scores between target sequences and sequences of homology matched reference strains is often time consuming and ambiguous. This is especially true if only partial target sequences are available or if different ssRNA virus families are jointly analyzed. In such cases, knowledge of residues that uniquely distinguish among known reference strains is critical for rapid and unambiguous strain identification. Conventional sequence comparisons are unable to identify such capsid residues due to high sequence divergence among the ssRNA virus reference strains. Consequently, automated general methods to reliably identify strains using strain distinguishing residues are not currently available.ResultsWe present here RECOVIR ("recognize viruses"), a software tool to automatically detect strains of caliciviruses and picornaviruses by comparing their capsid residues with built-in databases of residues that uniquely distinguish among known reference strains of these viruses. The databases were created by constructing partitioned phylogenetic trees of complete capsid sequences of these viruses. Strains were correctly identified for more than 300 complete and partial target sequences by comparing the database residues with the aligned residues of these sequences. It required about 5 seconds of real time to process each sequence. A Java-based user interface coupled with Perl-coded computational modules ensures high portability of the software. RECOVIR currently runs on Windows XP and Linux platforms. The software generalizes a manual method briefly outlined earlier for human caliciviruses.ConclusionThis study shows implementation of an automated method to identify virus strains using databases of capsid residues. The method is implemented to detect strains of caliciviruses and picornaviruses, two of the most highly divergent ssRNA virus families, and therefore, especially difficult to identify using a uniform method. It is feasible to incorporate the approach into classification schemes of caliciviruses and picornaviruses and to extend the approach to recognize and classify other ssRNA virus families.