Correlation networks provide a powerful tool for analyzing large sets of biological information. This method of high-throughput data modeling has important implications in uncovering novel knowledge of cellular function. Previous studies on other types of network modeling (protein-protein interaction networks, metabolomes, etc.) have demonstrated the presence of relationships between network structures and organization of cellular function. Studies with correlation network further confirm the existence of such network structure and biological function relationship. However, correlation networks are typically noisy and the identified network structures, such as clusters, must be further investigated to verify actual cellular function. This is traditionally done using Gene Ontology enrichment of the genes in that cluster. In this study a novel method to identify common cluster functions in correlation networks is proposed, which uses annotations of edges as opposed to the traditional annotation of node analysis. The results obtained using proposed method reveals functional relationships in clusters not visible by the traditional approach.