张拉整体结构的智能化找形研究进展

近年来,“未来的结构体系”张拉整体结构得到学术界的广泛关注。其中,找形是张拉整体结构设计的关键步骤,即确定结构的平衡状态的过程。随着人工智能逐渐应用到各个领域,张拉整体结构的智能找形方法也应运而生,通过使用人工智能技术改进传统的找形方法,以达到简化找形流程的目的。首先介绍人工智能在建筑领域的应用;其次,阐述使用人工智能技术改进张拉整体结构找形方法的研究意义;然后介绍张拉整体结构几种常用的传统的找形方法及其优缺点,再通过调研大量文献,对现在最新的张拉整体结构智能找形方法,特别是优化算法和神经网络方法进行详细介绍和分析;最后,预测并分析总结该领域未来可能的研究方向及相应的发展趋势。

城市空气质量感知方法综述

城市空气质量信息对于控制空气污染和保护大众健康都是尤为重要的。城市空气质量感知方法按传感器位置是否发生改变可划分为静态感知方法和动态感知方法两种。其中静态感知方法的数据是基于空气质量监测站、卫星遥感和固定位置的传感器进行采集的。再按成本高低可进一步划分为低成本静态感知和高成本静态感知。动态感知方法按是否以参与者为感知中心划分为参与式方法和非参与式方法。随着感知技术和计算能力的发展,将多源异构的城市数据,如气象数据、交通数据等进行融合,可进一步提高感知的准确性。文中首先对当前空气质量感知方法进行综述,然后分类介绍了各种方法的感知框架和数据处理方法,最后讨论了其面临的问题和挑战。

miRNA与疾病关联关系预测算法

microRNAs(miRNAs)在生命进程中发挥着重要作用.近年来,预测miRNAs与疾病的关联关系成为一个研究热点.当前,计算方法整体上可以分为两大类:基于相似度度量的方法和基于机器学习的方法.前者通过度量网络中节点之间的关联强度预测miRNA-疾病关联,但需要构建高质量的生物网络模型;后者将机器学习相关算法应用到这个问题中,但需要构建高可信度的负例集合.基于以上困难和不足,提出了一种计算模型BNPDCMDA,用于预测miRNAs-疾病关联关系.该方法首先构建miRNA-疾病双层网络模型,然后利用miRNA的功能相似度对其进行基于密度的聚类,进而将二分网络投影应用于聚类后的miRNAs及疾病集合构成的miRNA-疾病双层子网中,最终完成对miRNA与疾病关联关系的预测.实验结果表明,采用留一交叉验证法得到的AUC值可达99.08%,明显优于当前其他高效方法.最后,采用BNPDCMDA方法对某些常见疾病所关联的miRNAs进行预测,实验结果获得了文献的支持,进一步表明了该方法的有效性.

基于RNA-Seq的转录组分析方法

RNA-Seq技术凭借测序成本低、精度高、覆盖范围广等优点,已经成为了转录组分析的重要方法,为研究基因表达模式、疾病的生物标志物探测、作物抗逆性研究和分子育种等提供了新的手段。然而,RNA-Seq产生的海量数据也给数据分析带来了挑战,如何有效地对RNA-Seq数据进行处理和分析成为了生物信息学研究的热点。文中对基于RNA-Seq技术的转录组分析流程进行介绍,包括RNA-Seq数据预处理、差异表达分析和高层分析。其中,RNA-Seq数据预处理即对原始测序数据进行质控和定量计算;差异表达分析则是对基因进行筛选,通常基于统计学或机器学习两种方法;高层分析是对差异基因进一步处理,通过富集分析等手段确定基因功能和调控网络。最后,对基于RNA-Seq的转录组分析方法的发展进行了探讨。

基于时空卷积残差网络的空气质量预测

城市计算的数据具有鲜明的时空特性,即时间维度相似性与空间维度相近性的耦合关系,因此,对时空数据的分析和处理,已成为城市计算中亟需解决的热点问题。面向城市污染中的空气质量问题,提出时空卷积残差网络(spatio-temporal convolution residual network,ST-ResNet),通过分析空气质量指数(air quality index,AQI),实现预测和预警。时空卷积残差网络子组件是由以卷积层为基础的单元通过全等映射残差连接构成,将AQI数据通过空间转换组件转换成AQI像素图,利用卷积运算捕获其空间特性;而时间趋势性、周期性和时间接近度等属性分别被三个子组件捕获,将三者的输出加权连接得到时空卷积残差网络并输出AQI的预测结果。最后,将ST-ResNet网络与经典的长短时记忆网络(long short-term memory,LSTM)进行对比,结果表明ST-ResNet网络在准确率上比LSTM网络提高了7%,有望对城市环境监测预测和精细化管理提供理论依据和技术支撑。

Genomic Analysis of MicroRNA Promoters and Their Cis-Acting Elements in Soybean

MicroRNAs （miRNAs） are derived from distinct loci in the genome and play crucial roles in RNA-mediated gene silencing mechanisms that regulate cellular processes during development and stress responses of plants. The miRNAs are approximately 21 nucleotides long and code for the complementary strand to a larger genic mRNA. They are often found within the complementary primary transcript （pri-miRNAs）. In the past few years, a growing number of soybean miRNAs have been discovered, however, little is known about the transcriptional regulation of these miRNAs. In this study, promoters and cis-acting elements of soybean miRNAs were analyzed using the genomic data for the first time. A total of 82 miRNAs were located among 122 loci in genome, some were present as double or multiple copies. Five clusters that included ten miRNAs were found in genome, and only one cluster share the same promoter. A total of 191 promoters from 122 loci of the soybean miRNA sequences were found and further analyzed. The results indicated that the conserved soybean miRNA genes had a greater proportion of promoters than that of non-conserved ones, and the distribution of the transcript start sites （TSSs） and TATA-boxes found had different motif styles between conserved and non-conserved miRNA genes. Furthermore, the cis-acting elements 5＇ of the TSSs were analyzed to obtain potential function and spatiotemporal expression pattern of miRNAs. The data obtained here may lead to the identification of specific sequences upstream of pre-miRNAs and the functional annotation of miRNAs in soybean.

In silico Detection of Novel MicroRNAs Genes in Soybean Genome

The importance of microRNAs （miRNAs） at the post-transcriptional regulation level has recently been recognized in both animals and plants. In this study, the simple and most effective method of comparative genomic approach was used. First known plants miRNAs BLAST against the soybean genome, and then the located candidates were searched for novel miRNAs by RNA folding method in the vicinity （±400 nt） of the candidates. The results showed that a total of 521 novel soybean miRNA genes, including 236 mature miRNAs, were identified. All these mature miRNAs were grouped into 58 families, of which 21 of them were novel family in soybean. The upstream 2 000 nt of potential pre-miRNAs was used for promoter prediction, in order to investigate prediction of miRNAs and detect transcript unit and clustering. In this study, rniRNA genes less tend to be present as clusters in soybean. Only 9 clusters, containing 2l miRNA genes （accounted for 4.0% of the total）, were observed as part of polycistronic transcripts. Detailed analysis of sequence characteristics of novel miRNAs in soybean and all previous known plants miRNAs, were carried out. These results of this study provide a reference point for further study on miRNAs identification in plants, and improve the understanding of genome in soybean.

MicroRNA Primary Transcripts and Promoter Elements Analysis in Soybean (Glycine max L. Merril.)

The importance of microRNA （miRNA） at the post-transcriptional regulation level has recently been recognized in both animals and plants. In recent years, many studies focused on miRNA target identification and functional analysis. However, little is known about the transcription and regulation of miRNAs themselves. In this study, the transcription start sites （TSSs） for 11 miRNA primary transcripts of soybean from 11 miRNA loci （of 50 loci tested） were cloned by a 5＂ rapid amplification of cDNA ends （5＂ RACE） procedure using total RNA from 30-d-old seedlings. The features consistent with a RNA polymerase II mechanism of transcription were found among these miRNA loci. A position weight matrix algorithm was used to identify conserved motifs in miRNA core promoter regions. A canonical TATA box motif was identified upstream of the major start site at 8 （76%） of the mapped miRNA loci. Several cis-acting elements were predicted in the 2 kb 5＂ to the TSSs. Potential spatial and temporal expression patterns of the miRNAs were found. The target genes for these miRNAs were also predicted and further elucidated for the potential function of the miRNAs. This research provides a molecular basis to explore regulatory mechanisms of miRNA expression, and a way to understand miRNA-mediated regulatory pathways and networks in soybean.

改进选择策略的烟花算法

烟花算法(FWA)中的选择策略直接影响其收敛效率、收敛精度、对初值敏感性以及能否跳出局部最优,对此,提出一种改进选择策略的烟花算法(ISSFWA).ISSFWA建立峰值火花和探索火花的概念,并提出基于N-1朵峰值火花和一朵探索火花充当下一代N朵烟花的选择策略.峰值火花兼顾了火花的适应度值及相对位置,保证选择全局最优火花及峰值火花邻域内的局部最优火花,同时避免重复选择搜索能力相似的火花,而基于最远距离的探索火花可以增强全局探索能力.在10次标准及增加位置偏移的测试函数实验中,ISSFWA在最优适应度值方面优于PSO、GA、FWA;在平均适应度值方面优于PSO和FWA,略劣于GA.这一结果表明,ISSFWA能够增强寻找最优解的能力,降低对初值的敏感性,并提升搜索效率.