Prediction Of Drug-Target Network Based On Huge Data Of Drugs

Network pharmacology has brought a new theoretical guidance to drug research and development, which mainly build a comprehensive biological information network, and simulation analysis is conducted to the network by computer skills which takes multiple factors in to account. According to this method, the side effect is brought down and on the other hand the treatment performance improved. As a consequence, the success rate of the development of new drugs is increased and the cost is cut down.Virtual screening are adopted to simulate the lead compounds screening process using computer technology, to get recommended lead compounds from large drug database. At present, several kinds of virtual screening techniques are widely used which use the three-dimensional space of chemical to analysis, while a lot of target protein three-dimensional structure are measured. And that the mainstream sharing medicine information database on the international are limited. To solve these problem, this paper analyzed about the present situation of network pharmacology firstly, and build a drug biochemical information database which contains the international mainstream medicine database information and real-time information associated warehousing. Then we analyzes the researches on drug-target prediction at present, following the theoretical guidance of network pharmacology, and we build a pharmacology network based on biochemical information data, and put forward two kinds of hybrid similarity forecast method which analyze the roles of drugs comprehensively in the building pharmacology network, which can get better effect of prediction. Finally, this paper presents the distribution of Hadoop to implement the prediction algorithm proposed in this paper.