Novel function discovery through sequence and structural data mining.

摘要: Large-scale sequence and structural data is a goldmine of novel proteins, but how can this be effectively mined for new functions? Here, we review protein function prediction methods recent studies that apply these to discover functionality. Core approaches include sequence-based homology detection, phylogenetic analysis, bioinformatics, inference functional associations using genomic context related methods. With such wide range approaches, sequences may reveal functionality regardless their similarity characterized reference. Homologs known identified in unexpected species or associations. Detection shifts activities specificities. New functions also predicted uncharacterized structures. Finally, integrated applied at increasingly large scales due improved domain knowledge coverage, which amplifies the ability predict functions.