基于并行群体模拟退火算法的蛋白质折叠研究

蛋白质折叠研究对于揭示蛋白结构和功能关系,进而了解相关疾病的致病机理意义重大。蛋白质折叠已被证明是 NP-完全问题。本文针对蛋白质折叠研究中的能量最小化问题,提出了一种新的并行群体模拟退火算法（Parallel Group Simulated Annealing,PGSA）及其改进型算法（PGSA_1/K)。该算法使用了降温因子加速收敛精度,并采用 MPI 消息传递并行编程技术加快蛋白质结构空间搜索以及能量最小化寻找速度。以 Met_Enkephalin 蛋白为对象的计算机模拟仿真结果表明,我们提出的算法及其改进型有很好的扩展性,可以高效搜索蛋白结构空间,从而找到相关蛋白的最小能量结构。     

Molecular interaction networks in the analyses of sequence variation and proteomics data

Protein–protein interaction networks are typically generated in standard cell lines or model organisms as it is prohibitively difficult to record large interaction datasets from specific tissues or disease models at a reasonable pace. Although the interaction data are of high confidence, they thus do not reflect in vivo relationships as such. A wealth of physiologically relevant protein information, obtained under different conditions and from different systems, is available including information on genetic variation, protein levels, and PTMs. However, these data are difficult to assess comprehensively because the relationships between the entities remain elusive from the measurements. Here, we exemplarily highlight recent studies that gained deeper insight from genetic variation, protein, and PTM measurements using interaction information pointing toward the importance and potential of interaction networks for the interpretation of sequencing and proteomics data.     

Appropriating an architectural design tool for musical ends

NURBS (non-uniform rational b-spline) modelling has become a ubiquitous tool within architectural design praxis. In this article I examine three projects that utilise NURBS modelling as a means for which a musical system's inherent spatiality is visualised. There are numerous precedents for which architectural form is a derivation of a musical system, or a musical system is proportionally informed by architectonic gesture. I propose in this article three NURBS modelling methodologies: for the spatial analysis of Karlheinz Stockhausen's sound projection geometries in Pole für 2; for a spatial realisation of John Cage's indeterminate work Variations III; and for the generation of a surface manifold informed by musically derived soundscape data from the Japanese garden Kyu Furukawa Teien. Rather than seeking to translate music into inhabitable architecture, or architectonic form into music, I highlight an approach that produces an interstitial territory between discourses on architecture and music analysis.     

CLARA算法聚类蛋白质序列的实现

随着基因工程产生大量新序列,导致蛋白质序列数据库的迅速增长,巨量蛋白质数据的功能组和族谱分析使蛋白质序列聚类分析成为结构和功能基因组学重要的研究目标,应用数据挖掘技术对生物数据进行聚类分析成为生物信息学研究的热点。聚类分析算法中的CLARA划分算法已广泛应用于其它领域,但在大数据量蛋白质序列聚类分析中应用很少,文章应用CLARA算法对在基准数据库中选取的蛋白质序列进行聚类分析,并将结果与其它几种蛋白质聚类算法进行了比较。     

Forecasting binary longitudinal data by a functional PC-ARIMA model

In order to forecast time evolution of a binary response variable from a related continuous time series a functional logit model is proposed. The estimation of this model from discrete time observations of the predictor is solved by using functional principal component analysis and ARIMA modelling of the associated discrete time series of principal components. The proposed model is applied to forecast the risk of drought from El Niño phenomenon.     

Neural networks in process life cycle profit modelling

Changes in operational environment of the process industry such as decreasing selling prices, increased competition between companies and new legislation, set requirements for performance and effectiveness of the industrial production lines and processes. For the basis of this study, a life cycle profit (LCP) model of a pulp process was constructed using different kind of process information including chemical consumptions and production levels of material and energy flows in unit processes. However, all the information needed in the creation of relevant LCP model was not directly provided by information systems of the plant. In this study, neural networks was used to model pulp bleaching process and fill out missing information and furthermore to create estimators for the alkaline chemical consumption. A data-based modelling approach was applied using an example, where factors affecting the sodium hydroxide consumption in the bleaching stage were solved. The results showed that raw process data can be refined into new valuable information using computational methods and moreover to improve the accuracy of life cycle profit models.     

暂态混沌神经网络在蛋白质关联图预测中应用研究

蛋白质结构预测是生物信息学的一个主要研究方向,而蛋白质关联图预测是其中的一个重要内容.针对蛋白质关联图预测问题,提出一种暂态混沌神经网络实现方法,所提出的方法具有暂态混沌特性和平稳收敛特性,能有效避免传统Hopfield 神经网络极易陷入局部极值的缺陷.它通过短暂的倒分叉过程,能很快进入稳定收敛状态.仿真结果表明:暂态混沌神经网络解决蛋白质关联图预测问题时,总能收敛到全局最优或几乎接近全局最优,同时具有更高的搜索效率.这种方法预测精度达到0.27,比随机预测器高9倍.     

蛋白质在硅烷膜表面吸附特性的压电传感监测

在压电石英晶片表面自组装修饰末端为胺基的硅烷膜,该膜在酸性和中性pH条件下带正电荷,可通过静电引力吸附带负电荷的牛血清白蛋白(BSA),吸附过程中的质量变化可通过石英晶片的谐振频率变化实时监测。结果表明:该吸附过程虽然为部分不可逆,但吸附等温线符合Langmuir方程,吸附速率常数为0.0117 L.g-1.s-1,吸附平衡常数为2.87 L.g-1,表观饱和吸附量Γ∞=0.954μg.cm-2。     

多特征融合的蛋白质相互作用位点预测

蛋白质相互作用位点预测为蛋白质功能和药物设计的理解提供重要线索。而蛋白质的各种特征为蛋白质相互作用位点预测提供了大量有用信息,特别是进化信息、残基序列邻近和空间邻近性。不同的蛋白质特征对蛋白质间的相互作用的贡献也不一样。通过提取蛋白质序列谱、保守性和残基熵,提出了特征融合技术对蛋白质相互作用位点进行研究,采用SVM构建三种预测器,分别对各种不同的特征加以验证,实验结果表明了基于特征融合方法的有效性和正确性。     

二次拟合：一种提高软件可靠性模型精度的新方法

针对软件可靠性模型精度不高的问题,提出了"一次建模 + 二次拟合"的方法,通过误差补偿机制来提高模型精度,并将该方法应用于航天软件失效数据集.实践结果表明,提出的方法对于提高软件可靠性模型精度是有效的.