Annotation of miRNAs in COVID-19 coronavirus

The COVID-19 coronavirus is a new strain of coronavirus that had not been previously detected in humans. As its severe pathogenicity is concerned, it is important to study it thoroughly to aid in the discovery of a cure. In this study, the microRNAs (miRNAs) of COVID-19 were annotated to provide a powerful tool for the study of this novel coronavirus. We obtained 16 novel coronavirus genome sequences and the mature sequences of all viruses in the microRNA database (miRbase), and then used the miRNA mature sequences of the virus to perform the Basic Local Alignment Search Tool (BLAST) analysis in the coronavirus genome, extending the matched regions of approximately 20 bp to two segments by 200 bp. Six sequences were obtained after deleting redundant sequences. Then, the hairpin structures of the mature miRNAs were determined using RNAfold. The mature sequence on one hairpin arm was selected into a total of 4 sequences, and finally the relevant miRNA precursor prediction tools were used to verify whether the selected sequences are miRNA precursor sequences of the novel coronavirus. The miRNAs of the novel coronavirus were annotated by our newly developed method, which will lay the foundation for further study of this virus.     

适应多领域多来源文本的汉语依存句法数据标注规范

近十年来,依存句法分析由于具有表示形式简单、灵活、分析效率高等特点,得到了学术界广泛关注。为了支持汉语依存句法分析研究,国内同行分别标注了几个汉语依存句法树库。然而,目前还没有一个公开、完整、系统的汉语依存句法数据标注规范,并且已有的树库标注工作对网络文本中的特殊语言现象考虑较少。为此,该文充分参考了已有的数据标注工作,同时结合实际标注中遇到的问题,制定了一个新的适应多领域多来源文本的汉语依存句法数据标注规范。我们制定规范的目标是准确刻画各种语言现象的句法结构,同时保证标注一致性。利用此规范,我们已经标注了约3万句汉语依存句法树库。     

基于矢量瓦片的点状要素注记处理技术

矢量瓦片体积小,可高度压缩,受网络带宽开销和存储空间的限制较小。地图瓦片化对桌面软件点注记处理带来挑战,不仅是点注记的处理存在重要注记被次要注记压盖、同级注记之间互相压盖、注记与要素压盖等问题,同时瓦片化也带来了注记被瓦片边缘截断显示不完全的问题,这些问题严重影响了地图的可读性和信息传递功能。本文针对以上问题总结了矢量瓦片点注记处理原则,通过设计矢量瓦片的组织结构,确定注记搭配表的JSON组织形式,明确矢量瓦片点注记的绘制流程,根据矢量瓦片特点使用四叉编码进行目标过滤,使用R树作为高效空间索引,并采用基于注记优先级的避障技术解决上述点注记处理带来的问题。     

基于决策树C4.5集成算法的图像自动标注

集成学习逐渐成为机器学习的研究热点。针对决策树C4.5集成算法中基分类器多样性差的问题,提出了修正矩阵Correction Matrix –C4.5(CMC4.5)集成学习算法,并将其应用于图像自动标注。算法首先对特征子集进行多样性处理,然后通过构造修正矩阵依次得到基分类器C4.5全新训练数据集,实现训练数据集之间的多样性和属性特征完整性,完成集成算法。对比实验表明,CMC4.5集成学习方法大大提高了分类准确率。将CMC4.5集成学习与图像标注相结合,实现了基于CMC4.5的图像自动标注。     

心理学视角下的自动表情识别

自动表情识别是心理学与计算机科学等深度交叉的前沿领域。情绪心理学、模式识别、情感计算等领域的研究者发展表情识别相关的理论、数据库和算法,极大地推动了自动表情识别技术的进步。文章基于心理学视角,结合我们前期开展的相关工作,首先梳理自动表情识别的心理学基础、情绪的面部表达方式、表情数据的演化、表情样本的标注等方面的理论观点与实践进展,然后分析指出自动表情识别面临的主要问题,最后基于预测加工理论的建构观点,提出注重交互过程中的表情“理解”,有望进一步提高自动表情识别的有效性,并预期这可能是自动表情识别研究的未来发展方向。     

Smart Spaces for Adaptive Information Integration in Bioinformatics

Bioinformatics is reaping the benefits of advances in Semantic Web technology thanks to the growing number of available biomedical web resources and portals. Although positive in general, this abundance poses practical challenges to researchers who must be skilled in techniques for retrieving and integrating sparse and complex contents, and thereof calls for more “intelligent” and user friendly ways of interaction to easily get information. With the aim of making available the intelligent functionality of smart systems, this paper presents SSAIIB (Smart Spaces for Adaptive Information Integration in Bioinformatics), a reference framework for designing bioinformatics smart applications that support discovering, aggregating and delivering contents from web resources according to user’s goals, tasks and concerns. The “Smart Spaces” are software environments whose smartness lies in their ability to adaptively accomplish specific user’s activities such as the exploiting content from biomedical resources, integrating data captured from different sources, supporting data analytics, etc. SSAIIB is structured around two main technologies: service oriented architectures and software agents. In particular, it relies on mechanisms for dynamically assembling suitable services and the use of agents as a natural metaphor for both modelling user’s activities and accessing web resources. A case study is presented that shows the application of SSAIIB to the design and the implementation of a smart space for annotating biomedical texts.     

Phylogenetic Studies,Gene Cluster Analysis,and Enzymatic Reaction Support Anthrahydroquinone Reduction as the Physiological Function of Fungal 17β‐Hydroxysteroid Dehydrogenase

17β‐Hydroxysteroid dehydrogenase (17β‐HSDcl) from the filamentous fungus Curvularia lunata (teleomorph Cochliobolus lunatus) catalyzes NADP(H)‐dependent oxidoreductions of androgens and estrogens. Despite detailed biochemical and structural characterization of 17β‐HSDcl, its physiological function remains unknown. On the basis of amino acid sequence alignment, phylogenetic studies, and the recent identification of the physiological substrates of the homologous MdpC from Aspergillus nidulans and AflM from Aspergillus parasiticus, we propose an anthrahydroquinone as the physiological substrate of 17β‐HSDcl. This is also supported by our analysis of a secondary metabolite biosynthetic gene cluster in C. lunata m118, containing 17β‐HSDcl and ten other genes, including a polyketide synthase probably involved in emodin formation. Chemoenzymatic reduction of emodin by 17β‐HSDcl in the presence of sodium dithionite verified this hypothesis. On the basis of these results, the involvement of a 17β‐HSDcl in the biosynthesis of other anthrahydroquinone‐derived natural products is proposed; hence, 17β‐HSDcl should be more appropriately referred to as a polyhydroxyanthracene reductase (PHAR).     

CALL模式下电子词典多种注释信息对词汇附带习得的分析

随着现代技术的发展,许多学者开始关注将二语习得与新型技术相结合的研究领域。主要研究了在计算机辅助教学软件的帮助下,学生对电子词典中各种不同注释信息的选择方式及对词汇记忆的影响。实验中的CALL软件记录下学生对电子词典中注释信息的选择,和点击次数,通过前测和后测,得出：母语＋二语,母语＋二语的注释是最有利于词汇记忆的查阅方式。     

用语义联想支撑基于内容的视频检索

视频检索的传统方法之一是首先从视频中摘取出文本信息（如标题,关键词等等）,然后基于一关键字集上回答用户的查询,由于自动摘取文本信息的过程至今尚未自动化,因而从视频中摘取信息主要由人工来完成,这在实际应用上证明是不现实的,另一种方法则是上一情形的极端,即它是利用低层的视频内容,诸如颜色,纹理,形状,运动特征等等,目的在于克服人工摘取关键字所涉及的困难,文中提出了基于ＴｏＣ视频结构的语义表达,从视频的     

用C语言在PCB中加中文注释的方法

论述了用Ｃ语言在ＰＣＢ中加中文注释的方法