AnaGram: A computational tool for protein function prediction
AnaGram is able to propose a function to protein sequences based on finding correlations between short sequences signals and functional annotations in public databases.
AnaGram is a computational tool for protein function assignment based on detecting small significant fragments by identity that could act as the modular pieces in the peptide construction.
AnaGram assigns function to protein sequences based on finding correlations between short sequence signals and functional annotations in a protein database. The overall procedure is divided into two different successive stages. First, the query sequence is analyzed to find statistically significant subtle amino acid patterns in the database (Thode et al., 1996, Rodriguez et al., 2000), which are called protomotifs (because they do not constitute, separately, motifs with their own structural or functional organization).
In the second step, the protomotifs are associated with the functional annotations derived from the original SWISS-PROT entries that gave rise to them, and then used with the aim of assigning functions to the analyzed sequence. At present is used the Keywords, Features, and References fields as informative lines on function from SWISS-PROT. The Keywords are used in order to assign the functions and the Features locate these functions in a specific sequence position.