TranSMART Translational Medicine Data Platform’ Localization And Further Development

Currently, the Chinese market of translational medicine data platforms is still empty, and none of the translational medicine data platforms have been researched and popularized in China. TranSMART is a translational medicine data platform with an outstanding ability in data management that has been developed by Imperial College and Johansson & Johansson pharmaceutical consortium. This paper localizes and further develops tranSMART. This can effectively popularize the developing process of translational medicine data platforms in China.This paper introduces the technical architecture and related techniques of tranSMART, and analyzes the requirement of localization of the original tranSMART application. On this basis, localization programs of user interface, data, the user manual and advanced workflow are made, then, the localization of tranSMART platform have been completed. After studying mining algorithms and cloud computing platform, I explore the method of adding new advanced workflow to tranSMART, implemente a decision tree by R and add it successfully to tranSMART. Because of the increase of medical data, a RHadoop platform must be set up, allowing tranSMART platform to combine with big data. Then from parallelizing the original K-Means algorithm by MapReduce, the result demonstrates that improved algorithm’s rate has greatly increased, thus verifying that tranSMART platform when combined with RHadoop has the ability to deal with large amounts of data.