| Popis: |
Experimental drug development is time-consuming, expensive and limited to a relatively small number of targets. However, recent studies show that repositioning of existing drugs can function more efficiently than experimental drug development to minimize costs and risks. In this paper, we use the relationship extraction method to obtain the disease-causing genes and proteins from SemMedDB(A PubMed-scale repository of biomedical semantic predications). Then, we calculate the influencing degree of every gene and protein by combining the relationships of gene-gene, gene-protein, protein-protein, protein-gene. Next, these different influences of each gene and each protein are used to calculate the possibility of drugs that can treat disease. Finally, we propose two methods to sort the prediction of PSPS. The experiments about the potential drug predictions of Parkinson disease are applied to verify the validity of PSPS. The results of the prediction of PSPS demonstrates that the linking of relevant structured semantic relationships in SemMedDB can support the discovery of potential disease drugs. In my results, 9 and 7 kind pharmacological substances in the TOP10 of our two rank results are confirmed by biomedical literature data respectively. Furthermore, some pharmacological substances in our prediction results are regarded as research objectives which may treat Parkinson disease. However, other top-ranking results may have the potential of treating Parkinson disease are not mentioned before. Moreover, our PSPS can be applied to drug repositioning studies in different diseases. |
