江谷典子 医学研究科特定研究員は、薬剤やその副作用、疾患の原因となる遺伝子などのビッグデータを解析することで、副作用をほぼ確実に予測できるとの研究成果を発表しました。 加えて、既存の薬剤の中で、元々のターゲット以外の疾患に効果を発揮する可能性があるものについての予測も行い、 いままで治療薬が公開されていない疾患に対して300件以上の候補を発見しました。 本研究成果は8月7日、Springer社の学術雑誌Journal of Big dataに掲載されました。
論文の結論ね Conclusion This paper describes database application model and its service for drug discovery introducing our proposed software development process in MDA into our research pro- cess. The issue of veracity can be solved when pinpoint data are selected from drug properties in big data analytics with domain model. The datasets of pinpoint data are designed to predict side effect and its incidence even if some of pinpoint data are missing. The proposed prediction model of side effect and its incidence can attain almost 100 % accuracy in its performance. In addition to our first research goal of side effect predic- tion in drug discovery, we can discover 2 new services for drug discovery by new uses for old drugs in logical system model with big data analytics. The veracity can be strength- ened based on the number of disease gene directly interacted to chemicals (drugs) and the number of the disease gene and SNPs contributing to human disease. So far, it can be concluded that the reputation of data modeling with data analytics and implementation can strengthen the “veracity” of big data. Our approach of software development process in MDA will be useful for devel- oping a big data application and a new service by “veracity” and “value” of big data because MDA provides an approach for deriving value from models and architecture. Our proposal will contribute to the personalized drug discovery with drug screening of side effect prediction in the personalized medicine although our proposed system is comprehensive.