摘要：随着化学信息学以及人工智能技术的发展,计算机辅助合成路线设计已开始崭露头角.早期逆合成的预测主要依赖人工定义的专家规则,近年来依赖数据驱动的人工智能模型取得诸多进展,不少工具已经应用在现今化学合成的实际工作中.本文就目前AI算法在逆合成路线设计中的应用展开系统的综述,主要论述逆合成所涉及的三个重要组成部分,包括用于存储与表征化学信息的技术、单步逆合成反应的预测模型与算法和合成路线的多步策略.最后还介绍了如今比较成熟的逆合成设计平台以及模型评价方法,从而能更系统地了解到逆合成系统的构成以及目前所面临的一些挑战.

摘要：Background： A comprehensive metabolism network of engineered E. coli is very important in systems biology and metabolomics studies. Many tools focus on two-dimensional space to display pathways in metabolic network. However, the usage of three-dimensional visualization may help to understand better the intricate topology of metabolic and regulatory networks. Methods： We manually curated large amount of experimental data （including pathways, reactions and metabolites） from literature related with different types of engineered E. coli and then utilized a novel technology of three dimensional visualization to develop a comprehensive metabolic network named SynBioEcoll. Results： SynBioEoli contains 740 biosynthetic pathways, 3,889 metabolic reactions, 2,255 chemical compounds manually curated from about 11,000 metabolism publications related with different types of engineered E. coil Furthermore, SynBioEcoli integrates with various informatics techniques. Conclusions： SynBioEcoli could be regarded as a comprehensive knowledgebase of engineered E. coli and represents the next generation cellular metabolism network visualization technology. It could be accessed via web browsers （such as Google Chrome） supporting WebGL, at http：//***/synbioecoli/.

摘要：以中心静脉导管挤出成型精度为研究对象,借助Polyflow软件通过反变形技术分析得出精确的口模截面形状及尺寸;通过压缩角、口模成型段长度、间隙角度等参数进行了3因素2水平的正交试验仿真,借助方差分析得到中心静脉导管最佳的挤出模具流道结构参数,并通过生产实际验证了仿真分析结果。结果表明,该挤出模具生产成品率高,对相关类型挤出模具的开发具有一定的借鉴意义。