Despite the dramatic growth of online genomic data, their understanding is still in early stage. To have meaningful interpretation, it requires the integration of various types of data that are relevant, such as gene sequence, protein interaction and metabolic pathway. Gene Ontology, an ontology proposed by molecular biologist community, is a possible tool to help address some of the difficulties in such integration. In this paper, we exam the formality of Gene Ontology, and study the possibilities and potential problems in applying Gene Ontology to both structured (such as database) and semi-structured (such as the publications in the literature) data integration for online genomic information.