Characterizing Proteins with Finer Functions: A Case Study for Translational Functions of Yeast Proteins

Based on high-throughput data, numerous algorithms have been designed for finding functions of novel proteins. However, the effectiveness of such algorithms is currently limited by some fundamental factors including the low a-priori probability of novel proteins participating in a detailed function and the lack of detailed functional knowledge for training algorithms. For such partially characterized proteins, we suggest an approach to find their finer functions based on protein-protein interaction sub-networks, which can efficiently find proteins' novel functions. As an application, we find that finer functions can be predicted for 18 and 15 proteins currently annotated in "protein biosynthesis" and "translation" with more than 90% precision, respectively. The predicted finer functions are highly valuable both for guiding the follow-up wet-lab validation and for providing the necessary data for training algorithms to learn other proteins.