The VIROME web-application was developed using Adobe Flex and runs on any web-browser that supports the Flash plug-in. This architecture ensures that the VIROME web-application is platform and browser independent. Communication between the back-end MySQL database and Seliciclib mw the VIROME web-application is handled by an Adobe ColdFusion server. Discussion The one consistent finding among viral metagenomics studies has been the high proportion of sequences having no significant homology to a known sequence within one of the large sequence databases (e.g., GenBank, UniRef etc.). Those viral metagenome libraries having the highest frequency of hits to known sequences typically come from marine environments where the hit frequency for longer Sanger reads is around 30% (at a BLAST e-score of ��0.001) [12].

Sanger libraries from soils show even lower hit rates at ~20%. The lack of homology to known sequences is only exacerbated by the shorter read lengths of next-generation sequencing technology [33] where libraries sequenced using the longest average read length next generation sequencing technology (i.e., 450 bp for the 454 pyrosequencing Ti FLX chemistry) yield hit rates to known sequence databases of less than 20%. In contrast, microbial shotgun metagenome libraries analyzed using the same databases and approaches will yield hit frequencies of ca. 80% [33]. These trends indicate that most viral genes are not represented within the major sequence databases. Viral metagenome hit frequencies are even lower when considering smaller, better annotated databases such as SEED and KEGG.

As a result, most metagenomic analyses of the genetic and taxonomic composition of viral communities have been based on small sub-populations of sequences within viral shotgun libraries. This subset of sequences has alerted researchers to the ubiquitous presence of metabolic genes within viruses, genes once thought to exist only in cellular life [34,35], and supported the development of approaches such as Maxi�� for examining the taxonomic diversity of viral communities [36]. However, relying solely on known sequence homologs ultimately stymies the discovery of novel viral genes that likely encode unique and important biological features of viruses found in nature. Thus, a key motivation behind developing the VIROME pipeline and ORF classification scheme has been to add some level of information to the majority viral metagenome ORFs having no significant hit to a known sequence.

This Dacomitinib objective has been accomplished through BLAST analysis of predicted viral metagenome ORFs against the ~ 49 million peptides within the Metagenomes On-Line database. To our knowledge, the only other metagenomics analysis pipeline to include analysis against environmental peptides is the Viral Metagenome Annotation Pipeline (VMGAP) [37].