现代优化计算方法在蛋白质结构预测中的应用

现代优化计算方法在蛋白质结构预测中占有重要地位.简要地介绍了模拟退火算法,遗传算法,人工神经网络和图论算法在蛋白质结构预测中的应用.对国内外近年来应用这些算法,特别是在蛋白质构象搜索问题中,解决蛋白质结构预测的研究作了回顾,并分析、比较了这几种算法的效果和特点.     

Recognition of the Finite Simple Groups F 4(2m) by Spectrum

The spectrum of a finite group is the set of its element orders. A finite group G is said to be recognizable by spectrum, if every finite group with the same spectrum as G is isomorphic to G. The purpose of the paper is to prove that for every natural m the finite simple Chevalley group F ₄(2 ^m ) is recognizable by spectrum.     

Kinetics of proton transfer in a green fluorescent protein: A laser-induced pH jump study

The sub-millisecond protonation dynamics of the chromophore in S65T mutant form of the green fluorescent protein (GFP) was tracked after a rapid pH jump following laser-induced proton release from the caged photolabile compoundo-nitrobenzaldehyde. Following a jump in pH from 8 to 5 (which is achieved within 2 μs), the fluorescence of S65T GFP decreased as a single exponential with a time constant of ∼90 μs. This decay is interpreted as the conversion of the deprotonated fluorescent GFP chromophore to a protonated non-fluorescent species. The protonation kinetics showed dependence on the bulk viscosity of the solvent, and therefore implicates bulk solvent-controlled protein dynamics in the protonation process. The protonation is proposed to be a sequential process involving two steps: (a) proton transfer from solvent to the chromophore, and (b) internal structural rearrangements to stabilize a protonated chromophore. The possible implications of these observations to protein dynamics in general is discussed     

组织协同进化分类算法用于雷达目标一维像识别

针对雷达目标一维像识别问题，提出了一种基于组织协同进化分类算法的识别方法．该方法与现有进化分类方法的不同之处在于它的进化操作直接作用于样本而不是规则，采用了一种自下而上的搜索机制，即先使若干样本的集合得到进化，再从进化结果中提取规则．这样有利于避免在进化过程中产生无意义的规则．该方法不需要进行特征提取；对于高维数据，不需要预先进行降维处理；没有复杂的运算，训练和识别的速度都很快．对3种飞机微波暗室实测数据的识别实验表明，该方法性能稳定，优于基于支撑矢量机与子波核函数的方法，识别率均达到了96％以上．实验中还对算法的抗噪能力进行了测试，获得了良好的效果．     

机器人头部的视觉跟踪系统研究

对运动目标进行识别和跟踪是机器人视觉的一项关键技术。提出了一种可以快速跟踪运动目标的系统设计方案,采用了双计算机处理和Memolink通信方式的机器人控制结构,以图像处理为基础,在图像经过校正之后,通过区域分割和目标识别两个步骤来确定目标的位置,并控制机器人对目标进行跟踪。实验表明,该方法具有较高的速度和准确性。     

Intrinsic Fluorescence of Actin

In this work actin is used to illustrate connection of protein fluorescence characteristics with its structure. On one hand, it has been demonstrated what kind of information about the contribution of each tryptophan residues to the bulk fluorescence spectrum can be obtained from the special analysis of protein three-dimensional structure. On the other hand, potentials of intrinsic fluorescence for elucidation of proteins structure, dynamics and processes of folding-unfolding are shown. In particular, using this method a new essentially unfolded kinetic intermediate state of actin was detected and characterized, and the place of inactivated actin and its kinetic predecessor in the process of folding-unfolding was determined. It has been revealed that inactivated actin is not intermediate state between the native and completely unfolded states, as it has been accepted before, but a result of protein misfolding. On the basis of the obtained data a new model of actin folding-unfolding pathway has been proposed.     

江浙蝮蛇蛇毒蛋白C激活因子基因的克隆及测序

从江浙蝮蛇毒腺中抽提总RNA，RT－PCR进行体外扩增，获得江浙蝮蛇蛇毒蛋白C激活因子基因，克隆至pGEX-5X-3载体中，对3个重组克隆分别作DNA全序列分析，通过遗传密码推导出相应的氨基酸序列，与其它已知的丝氨酸复白酶蛇毒蛋白的氨基酸作比较，其中许多上氨基酸有很强的同源性，该基因的成功克隆，不仅推导出江浙蝮蛇蛇毒蛋白C激活因子的蛋白质序列，也为进一步开展江浙蝮蛇蛇毒蛋白C激活因子蛋白质工程的研究工作打下良好的基础。     

酵母蛋白质网络的动力学性质

蛋白质-蛋白质、蛋白质-DNA相互作用网络决定了细胞中各种关键功能的执行．基于芽殖酵母(budding yeast saccharomyces cerevisiae)的蛋白质-蛋白质相互作用网络数据和相关的实验文献，我们建立了调控细胞周期和生命周期(cell cycle and life cycle)的蛋白质网络，并利用离散模型研究了该网络的动力学性质，研究表明：细胞周期网络的动力学性质具有很强的稳定性，约94％的蛋白质初态将演化到对应于生物学G1基态的稳定态，使其成为惟一的全局吸引点；同时，绝大多数的初态的演化路径都通过由G1激发态到G1基态的细胞周期演化路径，使细胞周期路径成为全局性的“吸引”路径。     

基于组织协同进化分类算法的遥感图像目标识别

针对遥感图像目标识别问题,提出了一种基于组织协同进化分类算法的识别方法。它没有复杂的运算,训练和识别速度都很快。对实测遥感图像的实验表明,本文方法性能稳定,优于文献[3]和[4]中基于支撑矢量机的方法,识别率均达到了95％以上,且训练时间非常短,不到1秒钟。