Predicting protein function is one of the most challenging problems of the post-genomic era. The development of experimental methods for genome scale analysis of molecular interaction networks has provided new approaches to inferring protein function. There are various approaches available for deducing the function of proteins of unknown function using protein information. In this paper, the reliable methods for assigning protein function are given based on the network of physical interactions. The characteristics of the method are: Function assignment is proteome-wide and is determined by the global connectivity pattern of the protein network. To validate the method, the yeast Saccharomyces cerevisiae protein-protein interaction network is analyzed. Comparing with the current protein function prediction based on network, our method can improve the quality of prediction substantially with multiple data sources. The precision has achieved 82% in the stringent functional classification and 96% in the less detailed classification.