我们处在一个大数据的时代。利用传统机器学习算法（例如：人工神经网络、支持向量机等）和深度学习算法预测化合物的生物活性，从大型数据库中寻找潜在的针对某个靶标及用于治疗某一疾病的先导化合物，是药物化学信息学的一个重要研究内容。作为由病毒导致的传染病，艾滋病和丙型肝炎严重威胁着人类健康。本项目拟选取抗艾滋病的HIV-1整合酶、抗丙肝病毒的 HCV蛋白酶和聚合酶作为生物靶标,相应的抑制剂为研究对象，建立全面的抑制剂数据库，用传统机器学习和深度学习的方法建立抑制剂的分类和定量预测模型，并建立相应的抑制剂数据库和计算模型的计算平台网站。用建立的模型对ZINC、PubChem等大型数据库进行虚拟筛选；对得到的高预测活性值的化合物进行ADMET性质评价；接着进行分子对接和分子动力学模拟等研究，从而得到高预测活性的潜在抑制剂。最后再对这些化合物进行生物活性实验测试，期望得到具有新颖骨架结构的先导化合物。

We are in an era of big data. It is an important task for researchers in medical chemoinforactis field to search for lead compounds for a certain target or for a disease therapy using conventional machine learning methods (e.g. artificial neural networks, support vector machines) and deep learning methods. Acquired immunodeficiency syndrome (AIDS) and hepatitis C (HC), which are caused by virus, threat to human health seriously. We select HIV-1 integrase, HCV protease and polymerase as the biological targets and put the corresponding inhibitors as our main research objects. We will establish comprehensive databases for inhibitors. With conventional machine learning and deep learning methods, we plan to establish the classification and quantitative prediction models on the bioactivities of the inhibitors, respectively. Put the corresponding databases and calculation models in the computer platform on the web. We will do the virtual screening by our calculation models on the big databases, such as ZINC and PubChem. Then we will evaluate the ADMET properties on the compounds which are predicted as the high-active compounds. Afterwards, we will make molecular docking and molecular dynamics simulations on enzymes and the compounds obtained from the above, to get the potential inhibitors. Finally, we will use the bioassay experiments to verify the prediction results. We want to find new lead compounds for HIV-1 integrase, HCV protease and polymerase with new scaffolds.