海洋天然产物研究新进展

海洋天然产物化学是目前天然产物化学中最活跃的研究领域之一。近年来,从海洋生物中分离到非常多的化学结构和生理活性上令人注目的新化合物,引起了有机化学家和药物化学家的关注。本文按萜类、太环内酯、聚醚类、生物碱,环肽、含氰化合物、甾醇、聚丙酸酯类化合物等化学结构来概述海洋天然产物研究近年来的进展。     

Determination of dissolved manganese in natural waters by differential pulse cathodic stripping voltammetry

A procedure based on differential pulse cathodic stripping voltammetry on the graphite electrode is described for the determination of dissolved manganese in natural waters buffered at pH about 6.5 with acetate solution. In order to avoid interference of iron(II) the addition of fluoride is used. The limit of detection is 3 g/l for a deposition time of 6 min. Acidification and UV-irradiation are recommended for samples containing dissolved organic matter. Results of manganese determination in table mineral waters are reported and the possibility of manganese speciation is discussed.     

Total synthesis of the marine alkaloid halitulin

The structure of the strongly cytotoxic marine alkaloid halitulin (1) has been confirmed by total synthesis and its absolute configuration determined as (15S). The synthesis follows a strategy previously reported by one of us and uses an efficient preparation of the quinoline-7,8-diol unit by modified Baeyer-Villiger and Skraup reactions. The O-benzyl protecting groups were removed in the last step of the synthesis by transfer hydrogenolysis without concomitant reduction of the quinoline ring. The method can be applied for the synthesis of halitulin analogues.     

分子内Diels-Alder反应在有机合成方面的新进展

本文按合成产物的结构类型,综述分子内 Diels-Alder 反应在有机合成方面的新进展。     

Coulombic Basis for Relative Hydrogen-Bonding Basicities and Conformational Energies of Tertiary Amine Oxides and Phosphine Oxides

The structures, energies, and natural atomic charges of 2-dimethylaminophenol oxide, 2-Me₂N-(O)C₆H₄OH, and 2-dimethylphosphinylphenol, 2-Me₂P(O)C₆H₄OH, in three different conformations were computed at the ab initio MP2/6-31G* level. Computed natural charges indicate distributions of electron density in amine oxides and phosphine oxides that are quite different from what is normally assumed on the basis of the formal charges in the usual representations of these compounds. The charges on nitrogen and phosphorus in these compounds are typically computed to be approximately zero on nitrogen and +2 on phosphorus, and the oxygen is considerably more negative in the phosphine oxide than in the amino oxide. Electronegativity differences thus play a larger role and formal charges a smaller one in determining atomic charges in these compounds than is generally believed. Despite the more negative oxygen in phosphine oxides, amine oxides are computed to be considerably more basic when participating in hydrogen bonding. Calculations treating the computed natural charges on these six conformations as point charges for classical approximations of the coulombic energies support the idea that the quantum mechanically computed relative energies are largely determined by coulombic interactions.     

链霉菌中3个新脂肪酰胺类化合物

特殊生境放线菌具有较大的开发前景,因其独特的生存环境可以产生新颖且具有生物活性的次生代谢产物.研究Streptomyces plumbidurans发酵液的次生代谢成分,并寻找具有抗肿瘤及抗菌活性的小分子化合物.利用凝胶Sephadex LH-20柱色谱、D101大孔树脂柱色谱、硅胶柱色谱等分离技术,并采用现代波谱学方法鉴定化合物的结构.采用噻唑蓝比色法测试化合物1～3抗肿瘤细胞毒活性;采用微量稀释法检测体外抑菌活性.从Streptomyces plumbidurans发酵液中确定3个新脂肪酰胺类化合物,并依次命名为：kcacylamide A(1)、kcacylamide B(2)和kcacylamide C(3),生物活性结果显示3个化合物在50μmol/L均无明显抗肿瘤细胞毒活性及抑菌活性.研究首次发现Streptomyces plumbidurans会产生脂肪酰胺类化学物质.     

联合实体边界检测的命名实体识别方法

针对传统命名实体识别方法无法有效利用实体边界信息的问题,提出一种联合实体边界检测的命名实体识别方法,即将实体边界检测作为辅助任务,增强模型对实体边界的判断能力,进而提升模型对实体的识别效果。首先,利用Bert预训练语言模型对原始文本进行特征嵌入获取词向量,并引入自注意力机制增强词对上下文信息的利用;其次,在命名实体识别任务的基础上,添加实体边界检测辅助任务,增强模型对实体边界的识别能力;再次,对比联合实体边界检测的命名实体识别方法与基线方法的有效性,并对测试结果进行消融实验;最后,进行样例分析,分析损失权重β对实体边界检测的影响。实验结果表明,在英文社交媒体数据集Twitter-2015上,联合实体边界检测的命名实体识别方法相较于基线模型取得了更高的精准率、召回率和F1值,其中F1值达到了73.57%;并且,边界检测辅助任务提升了基线方法的检测效果。所提方法能有效利用实体边界信息,从而获得更好的实体识别效果,促进了人机交互系统的发展,对自然语言处理下游任务有重要意义。     

基于Bert股吧舆情分析的特征融合股价走势预测研究

传统股价预测模型往往只考虑时序性数据且局限于模型自身机制,而忽略舆情对股价的影响,导致预测精度不高,针对该问题,提出基于 Bert 股吧舆情分析的特征融合预测模型对股价收盘价进行涨跌幅预测。 首先,采用Bert 自然语言处理对股吧舆情以及公司公告政策进行情感分类,并转化为虚拟变量,构建金融舆情情感特征库;然后将金融舆情特征库和时序性数据合并构建特征融合矩阵;最后输入长短期记忆网络模型(LSTM)进行股价收盘价预测,并得出股价的涨跌结果。 以华银电力(600744. SH)为例进行实证分析,实验结果表明:引入股票情感特征后的模型,得到的股价走势准确率上升了 8. 63%,预测收盘价的回归指标 FMAPE FRMSE 分别下降了 23. 59%、22. 9%,R2 提高了 8. 11%,证明引入新的舆情情感特征在实际预测中能提高股价预测的准确率,可以作为精准预测股价走势的手段。     

木芙蓉叶抗菌活性及其有效部位研究

采用乙醇回流法对木芙蓉叶粉末进行活性成分提取,用不同极性溶剂（石油醚、乙酸乙酯和水）对粗提物分别进行萃取。采用连续稀释法评价粗提物及其萃取组分对多种细菌的抗菌活性。结果显示乙酸乙酯和石油醚层萃取组分的抑菌活性良好,对各种菌株的最低抗菌浓度（MIC）在1.25~5 mg/mL之间。其中,石油醚萃取组分对枯草芽孢杆菌CMCC 63501的最小抗菌浓度达到0.625 mg/mL,抗菌效果最佳。乙酸乙酯萃取组分对四种耐药菌株也表现出较好的抗菌活性,其MIC值达到1.25或者2.5 mg/mL。本研究初步确定木芙蓉叶乙醇提取物具备广谱的抗菌活性,其有效部位为石油醚与乙酸乙酯萃取相,这为木芙蓉叶抗菌成分的提取分离及抗菌产品的开发提供了科学的理论及实验依据。     

空间机器人组装超大型结构的动力学分析北大核心CSCD

超大型航天结构具有超大柔性、超低固有频率的特点,空间机器人在轨组装时应尽可能避免激起超大型结构的柔性振动.空间机器人组装超大型结构模块的过程分成抓捕阶段、位姿调整与稳定阶段、安装阶段和爬行阶段.通过对安装阶段的动力学与控制研究,提出共线安装的轨迹规划方法,有效避免了柔性结构振动.首先,采用自然坐标法和绝对节点坐标法建立主结构-空间机器人-待组装结构的在轨组装系统动力学模型.然后,将共线安装的要求转化为空间机器人的轨迹规划约束,要求空间机器人质心到主结构/待组装结构的距离保持不变,实现共线安装的轨迹规划.数值仿真表明:提出的组装方法在组装过程中可有效避免超大型结构的横向运动,降低夹持力矩.最后,分析了系统参数对组装过程动力学响应的影响,为超大型航天器的在轨组装提供了参考.