Viruses are typically identified in metagenomic sequencing datasets via homologous inference from sequence alignment. Common tools for this purpose include BLAST, BLAT, Bowtie, or other pairwise sequence aligners. BLAST is quick and specific, and can find homologous protein pairs when the percent sequence identity of the alignment exceeds 30% over the length of a protein or protein domain. But as the pairwise sequence identity drops below 30% for full-length protein sequences, BLAST finds fewer true homologs, and this problem is further compounded by the shorter reads generated by second-generation sequencers. Because pairwise alignment is limited in detecting low percent identity homologs, researchers seeking more distant evolutionary relationships have shifted towards Nortriptyline so-called ����profile���� methods for the detection of remote homologs. Profile methods consider information across a family of evolutionarily related sequences, derived from a multiple sequence alignment. Profile search methods gain sensitivity by incorporating position-specific information into the alignment process and by quantifying variation across family members at each position. For example, a query sequence can match a family profile because it is evolving like other members of the family, even if it is not significantly similar to any one known member of the family. Of the profile methods, profile hidden Markov models typically outperform other profile methods in the detection of distant homologs. Because many viruses have more error-prone polymerases than typically found in cellular organisms, especially RNA viruses that rely on RNAdependent RNA-Polymerases for genome 11-hydroxy-sugiol replication, more distant viral homologs can arise on much shorter evolutionary time-scales than are generally observed in bacteria or macroorganisms. In light of this higher mutation rate, profile HMMs are particularly well suited for the detection of divergent viral sequences. Profile HMMs have been built for some viral proteins, but viral genes are not comprehensively covered in publicly available profile HMM databases.