Conformational stability and basal metabolic rate: Reexamination of the case of myoglobin

Summary The free energy of unfolding of several myoglobins from different animal species has been determined from their denatunation pattern by using the ligand binding model. The results indicate that no simple correlation exists between the free energy of unfolding of myoglobin and the basal metabolic rate of the animal species from which the myoglobin was isolated     

Structural and dynamical features contributing to thermostability in α-amylases

In recent years an increasing number of studies on thermophilic and hyperthermophilic proteins aiming to elucidate determinants of protein thermostability have yielded valuable insights about the relevant mechanisms. In particular, comparison of homologous enzymes with different thermostabilities (isolated from psychrophilic, mesophilic, thermophilic and hyperthermophilic organsims) offers a unique opportunity to determine the strategies of thermal adaptation. In this respect, the medium-sized amylolytic enzyme α-amylase is a well-established representative. Various studies on α-amylases with very different thermostabilities (melting temperature T_m = 40–110°C) report structural and dynamical features as well as thermodynamical properties which are supposed to play key roles in thermal adaptation. Here, results from selected homologous α-amylases are presented and discussed with respect to some new and recently proposed strategies to achieve thermostability.Received 28 February 2005; received after revision 29 April 2005; accepted 19 May 2005     

基于流形学习和流形高阶近似的图像距离度量

为克服切距离方法的不足,提出了流形高阶近似距离（HMD）.HMD度量方法通过最大差异延展方法（MVU）学习出非线性图像流形的内蕴变量,然后在原型图像处用基于流形高阶泰勒展开式的非线性曲面来局部近似图像流形.HMD定义为待识别图像与图像流形的高阶近似曲面间的最小距离,通过计算待识别图像与图像流形上多个基准图像间的HMD距离能够实现图像的分类和识别.人脸识别和手写数字识别的结果表明：HMD距离在识别精度和稳定性上要优于切距离和当前一些典型的图像距离度量方法.     

采用折叠-展开技术的一种并行排序算法

给出ｎ×ｎ网孔环接式阵列处理机上的一种并行排序算法,它将ｎ×ｎ阵列上的数据折叠成ｎ×ｎ／ｋ子阵列,排序后再展开到整个ｎ×ｎ阵列上,实现ｎ×ｎ项数据的行主序排序,其平均时间复杂度为（２＋１／ｋ）ｎ＋ｏ（ｎ）．若采用ｎ×ｎ／ｋ阵列模型,且各处理器初始、结束状态允许有ｋ项数据时,该算法的平均时间复杂度只有（１＋２／ｋ）ｎ＋ｏ（ｎ）．     

钣金冲压件CAPP系统中展开图的研究

通过对饭金零件几何信息描述模型和数据结构的分析,提出了利用参数化图形技术和按“父子”关系对零件各型面进行展开,解决了饭金冲压件CAPP系统中展开图自动生成的关键技术。     

基于Pro/E二次开发的钣金件设计

基于Pro／E二次开发,以三通管为例,详细介绍了参数的建立、MFC与Pro／E接口的实现、菜单的生成及MFC对话框的实现,得到了一种只需要修改MFC对话框中的参数,而不需要修改源程序代码便可以实现相似三通管设计的方法。     

天门冬氨酸对精氨酸激酶的作用研究

应用酶活测定、荧光发射光谱等方法分析了精氨酸激酶在不同浓度天门冬氨酸中活性变化和去折叠动力学。实验结果表明：天门冬氨酸引起精氨酸激酶失活与去折叠,并且过程中没有蛋白质的聚沉;精氨酸激酶的去折叠遵循一级反应动力学的二相过程,包括快相和慢相;失活与天门冬氨酸浓度呈正相关;当去除天门冬氨酸的影响,精氨酸激酶的活力能得到恢复。通过对比实验结果暗示,天门冬氨酸是通过其解离状态的H＋改变了机体pH值而影响精氨酸激酶活性的,并且就天门冬氨酸对精氨酸激酶影响及其意义进行探讨。     

Elman神经网络在中子解谱中的应用

人工神经网络由于其优良的自我调节能力及学习能力,已经被广泛地应用在各领域的非线性分析中.在中国锦屏极深地下实验室(CJPL)中的低本底液闪中子探测器一直在记录着中子的本底数据,探测器输出的能谱实际上是核反冲能谱,与输入能谱可一一对应,并随着输入能谱的改变而发生改变;因此可以将探测器输出信号输入到训练过的神经网络中判断输入能谱.本论文采用的神经网络为Elman神经网络,训练神经网络采用的数据为Geant4模拟所得.将实验获取的核反冲能谱输入到训练过的神经网路进行反解,最后Elman网络反解出的Am-Be中子源能谱与真实谱误差在0.1%～11.8%,反解出的~(252)Cf中子源能谱与真实谱误差在0.1%～8.9%.     

着色Petri网的结构展开方法

为提升着色Petri网的设计分析与模型检验能力,讨论了着色Petri网的结构化展开技术.以着色Petri网的令牌单元和绑定单元为基元,通过对着色Petri网展开为普通Petri网的等价性证明,提出了基于着色Petri网关联矩阵和标准元语言的展开规则和规范化步骤.研究结果为着色Petri网到普通Petri网的自动转换过程和着色Petri网验证提供了有力支持.     

Mimicking unfolding motion of a beetle hind wing

This paper presents an experimental research aiming to realize an artificial hind wing that can mimic the wing unfolding motion of AIIomyrina dichotoma, an insect in coleopteran order. Based on the understanding of working principles of beetle wing folding/unfolding mechanisms, the hind wing unfolding motion is mimicked by a combination of creative ideas and state-of-art artificial muscle actuator. In this work, we devise two types of artificial wings and the successfully demonstrate that they can be unfolded by actuation of shape memory alloy wires to provide actuation force at the wing base and along the leading edge vein. The folding/unfolding mechanisms may provide an insight for portable nano/micro air vehicles with morphing wings.