文摘
iFrag is a new computational approach to infer the interface region of the interaction between two proteins, based only in sequence information and with a competitive performance compared to the state-of-the-art methodologies. Despite the low general accuracy of state-of-the-art approaches on the prediction of protein–protein interactions purely based on sequence information, iFraG predictions are 1.5-fold better than random predictions (when comparing to domain-binding predictions) and can reach 155-fold improvement over random predictions when comparing to homolog searches. Thus, iFraG predictions are useful complements to experimental efforts aiming at, for example, defining interface regions in proteins, selection of residues for site-directed mutagenesis, etc. The length of the sequence fragment involved in the interface varies between the size of a domain and the length of a peptide. Graphically, the predictions are presented in a bidimensional heat map, thus providing a convenient representation to help define and select protein interfaces for subsequent analyses. A proof of concept included the experimental validation of a predicted peptide that modulates the aggregation of β-amyloids, which is responsible for Alzheimer's disease, hence proving the potential and value of iFraG predictions in a real case situation.
