Protein-protein interaction (PPI) is a topic of central interest for biologists. A large number of experimentally detected PPIs are mentioned in scientific literature and the scientists are interested in extracting them for their experimental work, which is a gigantic task given the enormous amount of literature. Therefore, automated methods were developed for the task using text mining technique. The recent BioCreative challenge revealed difficulties involved in the identification of correct interactor proteins from full text scientific articles. In order to overcome the problem, we use previously unexplored clues related to affixes of the target proteins. For example, experimental biologists often use specific fusion protein or protein-tags such as -GST, -His, -Myc, etc. for detecting and visualizing the interaction information.
We carried out detailed study of affixes and their association with interactor proteins. We also explored whether affixes can be used for detecting PPI related articles and interaction detection method.