A computational strategy to investigate relevant similarities between virus and human proteins: Local high similarities between herpes and human proteins.

Investigating primary sequence and structural features of viral proteins/genes has revealed molecular mimicry and evolutionary relationship linking viruses to eukaryotes. The continuous improvement in sequencing-techniques makes available almost daily the whole genome/proteome of several microorganisms, making now possible systematic analyses of evolutionary correlations and accurate phylogeny investigations. In the present study we set up a methodology to identify significant and relevant similarities between viral and human proteomes. To this aim, the following steps were applied: i) identification of local similarity corresponding to continuous identity over at least 8-residues long fragments; ii) filtering results for statistical significance of the identified similarities, according to BLAST parameters for short sequences; iii) additional filters applied to the BLAST outputs, to select specific viruses. The present study indicates a novel accurate methodology to find relevant similarities among virus and human proteomes, useful to further investigate pathogenic mechanisms underlying infectious and non-infectious diseases.