Integration Of Heterogeneous Biological Data With Semantic Web Technologies

Currently, the biomedical field is generating a large amount of gene, protein,disease and other related biological data at an amazing rate, which are characterizedby autonomy, heterogeneity and diversity. Biologist waste a lot of time and effort insearching and integrating for all the available biological information. Thus, theintegration of heterogeneous bioinformatics data has become an important topic.In this paper, the integration process of heterogeneous biological data is studiedand improved， including data gathering, data conversion, ontology alignment, dataintegration, data retrieval and output. In the data converting phase, an approach isutilized to establish domain ontology with relational databases, realizing the accessionof relational databases by using Semantic web technologies. For ontology matching,we use a variety of string-based matching algorithms for establishing a multiple linearregression model. After model evaluation, it is verified that the model can improve theontology matching results. For data integrating, a semantic integration framework forheterogeneous biological data is proposed which consists of semantic layer, datamanipulation layer, application service layer and data layer. For data retrieval andoutput, we respectively explain the process of ontology retrieval with SPARQL, theretrieval of relational databases with dynamically generated HQL queries, and theprocess of combing the result of both retrievals.