Dictionary and Gene Ontology Based Similarity for Named Entity Relationship Protein-protein Interaction Prediction from Biotext Corpus

摘要

Protein-protein interactions functions as a significant key role in several biological systems. These involves in complex formation and many pathways which are used to perform biological processes. By accurate identification of the set of interacting proteins can get rid of new light on the functional role of various proteins in the complex surroundings of the cell. The ability to construct biologically consequential gene networks and identification of the exact relationship in the gene network is critical for present-day systems biology. In earlier research, the power of presented gene modules to shed light on the functioning of complex biological systems is studied. Most of modules in these networks have shown small link with meaningful biological function, because these methods doesn't exactly calculate the semantic relationship between the entities. In order to overcome these problems arid improve the PPI results in the biotext corpus a new method is proposed in this research. The proposed method which directly incorporates Gene Ontology (GO) annotation in construction of gene modules and Dictionary-based text is proposed to extract biotext information. Dictionary-Based Text and Gene Ontology (DBTGO) approach that integrates with various gene-gene pairwise similarity values, protein-protein interaction relationship obtained from gene expression, in order to gain better biotext information retrieval result. A result analysis has been carried out on Biotext Project at UC Berkley. Testing the DBTGO algorithm indicates that it is able to improve PPI relationship identification result with all previously suggested methods in terms of the precision, recall, F measure and Normalized Discounted Cumulative Gain (NDCG). The proposed DBTGO algorithm can facilitate comprehensive and in-depth analysis of high throughput experimental data at the gene network level.