CLONING, EXPRESSION AND CHARACTERIZATION OF Pro3 GENE OF YEAST Saccharomyces cerevisiae

The proline biosynthetic pathway and Pro genes in Saccharomyces cerevisiae have just begun to be studied recently. In our laboratory, Pro2 gene of S. cerevisiae had been cloned in yeast. As described in this paper, yeast Pro3 gene was also cloned, which can complement yeast Pro3 mutants, and be expressed efficiently in E. coli. The high activities of this gene product, L-pyrroline-5-carboxylate (P5C) reductase, can be detected in both organisms. The activity of the Pro3 gene product in multiple copy plasmids is not higher than that of single copy genes in chromosomes in both yeast and E. coll. The preliminary characterization of the gene is also reported.     

荷电颗粒在旋转环形通道内分离性能研究

为研究带电旋转环形通道内荷电颗粒的运动和沉积特性,本文使用计算流体力学两相流离散颗粒法对带电旋转环形通道内的荷电颗粒的运动过程进行了模拟。根据模拟结果分析了不同粒径、电压、入口雷诺数和通道长径比等参数对荷电颗粒运动和沉积的影响,研究了荷电颗粒在旋转通道内离心力与电场力之间的竞争关系,探索了离心力和电场力导致的荷电颗粒运动和沉积变化的规律。结果表明,单个不同粒径颗粒具有不同的颗粒逃逸电压区间,区间的大小随着颗粒粒径的增大而增大,且区间的宽度随着通道长径比的增大将会明显变小;多个颗粒的逃逸率曲线,不同粒径的颗粒将会有不同程度的交叉,随着长径比的增大,颗粒逃逸率曲线的高度与交叉会有明显的减小,而随着转速的增大,颗粒逃逸率曲线的交叉会有一定程度的减小,且高度不会有明显变化。     

基于双向循环神经网络的汉语语音识别*

当前基于深度神经网络模型中,虽然其隐含层可设置多层,对复杂问题适应能力强,但每层之间的节点连接是相互独立的,这种结构特性导致了在语音序列中无法利用上下文相关信息来提高识别效果,而传统的循环神经网络虽然做出了改进,但是只能对上文信息进行利用。针对以上问题,该文采用可以同时利用语音序列中上下文相关信息的双向循环神经网络模型与深度神经网络模型相结合,并应用于语音识别。构建具有5层隐含层的模型,其中第3层为双向循环神经网络结构,其他层采用深度神经网络结构。实验结果表明:加入了双向循环神经网络结构的模型与其他模型相比,较好地提高了识别正确率;噪声对双向循环神经网络汉语识别有重要影响,尤其是训练集和测试集附加噪声类型不同时,单一的含噪声语音的训练模型无法适应不同噪声类型的语音识别;调整神经网络模型中隐含层神经元数量后,识别正确率并不是一直随着隐含层中神经元数量的增加而增加,神经元数量数目增加到一定程度后正确率出现了降低的趋势。     

Neutron powder diffraction and high-pressure synchrotron x-ray diffraction study of tantalum nitrides

Tantalum nitride(TaN) compact with a Vickers hardness of 26 GPa is prepared by a high-pressure and hightemperature(HPHT) method. The crystal structure and atom occupations of WC-type TaN have been investigated by neutron powder diffraction, and the compressibility of WC-type TaN has been investigated by using in-situ high-pressure synchrotron x-ray diffraction. The third-order Birch–Murnaghan equation of state fitted to the x-ray diffraction pressure–volume(P–V) sets of data, collected up to 41 GPa, yields ambient pressure isothermal bulk moduli of B'_0= 369(2) GPa with pressure derivatives of B 0= 4 for the WC-type TaN. The bulk modulus of WC-type TaN is not in good agreement with the previous result(B_0= 351 GPa), which is close to the recent theoretical calculation result(B_0= 378 GPa). An analysis of the experiment results shows that crystal structure of WC-type TaN can be viewed as alternate stacking of Ta and N layers along the c direction, and the covalent Ta–N bonds between Ta and N layers along the c axis in the crystal structure play an important role in the incompressibility and hardness of WC-type TaN.     

使用小型中子源的活化分析

(一)绪言 F.約里奥(Joliot)和1.居里(Curie)发現人工放射性同位素不久:赫弗西(lievesy)等用鐳鈹中子源照射含镝的钇,与已知量的样品比較,成功地測出镝的含量为0.1％。之后,由意大利罗耳拉(Rolla)教授的委託,他們又用100毫克的Ra-Be中子源六个,以石蜡为减速剂,測定了氧化钇中0.1％的铕。活化分析就是这样由赫弗西开始的。以后受中子流密度的限制,没有很大发展。自从带电粒子的加速器以及原子反应堆的出現和利用,才有了今日的盛况。近来,由于测定仪器的进展及各有特色的小型中子源也较易得到,如用手提式的中子源作非破坏的分     

Design and test of a Multi-gap Resistive Plate Chamber with Long readout-strip （LMRPC）

A new kind （two end readout） of Multi-gap Resistive Plate Chamber with long readout-strip （LMRPC） is developed to be used at the large-area Muon Telescope Detector （MTD） at mid-rapidity at RHIC/STAR experiment for Time-of-Flight （TOF） measurement. The LMRPC has an active area of 87 cm× 17 cm, 10 gas gaps of 250 μm arranged in 2 stacks, with readout strips of 2.5 cm wide and 90 cm long. The considerations in LMRPC design related to its performance are discussed in this paper. The cosmic ray test results of a prototype LMRPC show a detection efficiency 〉95% and the time resolution -70 ps.     

2×4双层结构MRPC的宇宙线测试结果

研制了一种新的2×4双层结构多气隙电阻板室探测器(MRPC),宇宙线测试结果表明,2×4双层结构MRPC有较低的工作电压和较长的工作坪区,探测效率90%以上,时间分辨可以达到100ps.     

一种基于比例因子的慢启动策略

TCP慢启动机制中拥塞窗口大小指数增长,会导致在慢启动前期发送速率偏低的现象。针对此问题,本文提出一种基于比例因子的慢启动策略,该策略实现了慢启动到拥塞避免的快速过渡,并保证了数据传输的总体效率。实验表明,新策略能明显提高网络吞吐量,利于网络资源利用。     

基于岸站二维投影采集的波浪观测补偿算法

针对海洋环境监测中视频采集设备因拍摄条件限制引起的梯形失真、倾斜失真和波高失真三类主要问题，提出一种以二维投影的三维还原为基础的视频采集处理的补偿算法。通过坐标和倾角偏移量补偿，还原失真的目标图像，并在此基础上进行目标测量信息处理，从而得到更加准确合理的测量结果。实验证明，该算法可根据任意位置参数实现失真图像的准确修复，精度高且易于实现。