CAA-PPI: A Computational Feature Design to Predict Protein–Protein Interactions Using Different Encoding Strategies

Protein–protein interactions (PPIs) are involved in an extensive variety of biological procedures, including cell-to-cell interactions, and metabolic developmental control. PPIs becoming one the most important aims system biology. act as a fundamental part predicting protein function target drug ability molecules. An abundance work has been performed to develop methods computationally predict this supplements laboratory trials offers cost-effective way likely set at entire proteome scale. This article presents innovative feature representation method (CAA-PPI) extract features from sequences using two different encoding strategies followed by ensemble learning method. The random forest methodwas used classifier for PPI prediction. CAA-PPI considers role trigram bond given amino acid with its nearby ones. proposed model achieved more than 98% prediction accuracy scheme 95% another diverse datasets, i.e., H. pylori Yeast. Further, investigations were compare approach existing sequence-based revealed proficiency both strategies. To further assess practical competence, blind test was implemented on five other species’ datasets independent training set, obtained results ascertained productivity schemes.