Drug Repositioning Based On Side Effects Similarity Measures And Content Based Recommendations Algorithm

Drug repositioning is aimed at finding new indications for existing drugs.Drug repositioning can significantly reduce time and money compared to traditional drug development processes.With the rapid development of bioinformatics,researchers have begun to try to solve this problem by using various calculation methods to construct related computational models.The side effects of drugs as a direct clinical manifestation of drugs in the human body have attracted the attention of researchers and have been applied to drug repositioning.This paper starts with the quantitative analysis of the side effects of drugs,instead of the qualitative approach used by most of the existing methods that simply consider whether a drug has some side effects,and builds a DSCB(Drug repositioning based on side effects similarity measures and Content Based Recommendations algorithm)model.The main idea is that similar side effects between drugs suggest that they may have similar therapeutic effects.This study represents the drug side effects information from clinical trials and market surveillance into a space vector model for text word frequency construction.The use of content-based recommendation algorithms to predict new indications for existing drugs provides researchers with a new track constructing drug repositioning models and reveals potential new therapeutic relationships between drugs and diseases.The predictions of the DSCB model achieved 62% accuracy in the existing dataset,and 78% of the drug-disease pairs were mentioned in the literature.Finally,this paper investigated the potential efficacy of dofetilide on Stomach Carcinoma and the new use of sildenafil for the treatment of Alzheimer's disease,and further confirms the reliability of the model from the perspective of molecular mechanism.