基于径向基神经网络的叶轮轴面投影图优化

为了提高余热排出泵的效率,采用拉丁超立方试验设计方法对叶轮轴面投影图上的前盖板圆弧半径、后盖板圆弧半径、前盖板倾角和后盖板倾角4个几何变量进行35组叶轮方案设计,应用ANSYS CFX 14.5软件对余热排出泵进行定常数值模拟,得到设计工况下的效率,应用径向基神经网络建立效率与轴面投影图的4个几何变量之间的近似模型,最后采用遗传算法对近似模型进行极值寻优,获得最优的轴面投影图几何参数组合。研究结果表明:对比原始泵的数值模拟性能曲线和试验外特性能曲线,两者吻合较好;径向基神经网络能较好地预测泵设计点效率;优化的轴面投影图使得余热排出泵的水力效率提高了6.18个百分点,改善了叶轮内流场特性。因此,叶轮轴面投影图的优化设计方法是可行的。     

基于PPC模型的农机选型与优序关系研究

将投影寻踪（PP）与实数编码的遗传算法（RAGA）相结合，应用到农业机械的选型与优序关系研究中，通过寻求各种机型评价指标的最佳投影方向，将高维数据转换成一维投影指标值，实现对机器样本的分类与排序，从而克服 以往二阶模糊综合评判法及灰色系统评价法中权重取值的人为因素，取得了良好的效果。     

遗传神经网络在温室黄瓜霜霉病预测中的应用

在神经网络结构的优化、连续权重的训练中采用遗传算法的方法对日光温室黄瓜霜霉病进行预测，并对其发生程度进行了拟合，计算值与实际值接近，表明该模型具有良好的预测效果。     

克隆文库分析红树林土壤中放线菌的多样性

直接提取8个红树林土壤的总DNA,用放线菌门的特异性PCR引物进行扩增,通过构建16S rRNA文库,测序及系统发育分析,获取新颖放线菌菌株.结果发现,有酸微菌亚纲(Acidimicrobidae) (30.2％)、放线菌亚纲(Actinobacteridae) (68.0％)、红蝽菌亚纲(Coriobacteridae)(1.1％)和去腈基菌亚纲(Nitriliruptoridae)( 0.7％)成员,没有红色杆菌亚纲( Rubrobacteridae)成员.仙农指数和Chao1物种多样性指数(Schao1)表明,红树林土壤样品中含有丰富的放线菌多样性.以16S rDNA基因的相似率≥97％作为临界点,45.1％可能代表新候选种(candidate species)(未培养、分离).本研究为进一步获取新颖的红树林放线菌资源提供了基础.     

OsGA20ox2不同长度RNAi片段对水稻株高等农艺性状的遗传效应

利用OsGA20ox2基因序列构建不同长度的RNAi片段并导入水稻，获得不同高度的矮化植株。将这些矮化植株与野生型植株回交获得B₁F₂群体，卡方检测表明B₁F₂群体矮秆植株数和高秆植株数符合3∶1比例，表现为矮秆显性的遗传规律。对矮化植株的F₅和B₁F₂群体株高、各茎节间长度和一些主要农艺性状方差分析显示, OsGA20ox2基因的RNAi能显著缩短株高和各节间长度(P<0.05)，RNAi干扰片段越长，使植株株高和节间长度缩短程度越大，可使株高降低24~42 cm，矮化22%~39%。在同一长度的RNAi干扰片段下，倒一节节间长度平均缩短与倒二节节间长度平均缩短非常相近，倒三节和倒四节节间长度平均缩短非常相近，总的缩短程度是倒四节>倒三节>倒二节>倒一节。这种近基部节间长度缩短幅度和比例较大的特点，利于提高水稻的抗倒伏能力，同时上部节间缩短幅度和比例较小，有效地保持合理株高，不使生物产量明显降低，有利于水稻的稳产和高产。OsGA20ox2基因的RNAi不影响如千粒重、结实率、穗长等其他主要农艺性状或影响很小。     

黑色/长梨形樱桃番茄种质的研究利用

以樱桃番茄黑褐色圆形自交系Z1、红黑色长梨形自交系Z2、红色圆形、浅粉红色扁圆形、
橙黄色长椭圆形和黄色圆形等自交系为试材，进行不完全双列杂交配制组合。通过对杂交组合后代的农艺
性状、品质特性和抗病性等的调查研究，探究黑色樱桃番茄种质品质改良和长梨形果形的利用，为今后特
殊果形果色番茄种质的创新利用提供参考。结果表明：樱桃番茄果色黑色性状为隐性性状；亲本为圆形果
或扁圆形果×长梨形果、长椭圆形果×长梨形果时，F₁ 果形分别是椭圆形与长指形。     

Plant monocultures produce more antagonistic soil Streptomyces communities than high-diversity plant communities

Plant–soil feedbacks are important to productivity and plant community dynamics in both natural and managed ecosystems. Among soil bacteria, the Streptomyces possess particularly strong antagonistic activities and inhibit diverse plant pathogens, offering a clear pathway to involvement in plant–soil feedbacks. We hypothesized that feedback effects and the ability of individual host plant species to foster antagonistic Streptomyces populations may be modified by the richness of the surrounding plant community. To test this, we collected soil associated with four different plant species (two C4 grasses: Andropogon gerardii, Schizachyrium scoparium; and two legumes: Lespedeza capitata, Lupinus perennis), grown in communities that spanned a gradient of plant species richness (1, 4, 8, 16, or 32 species). For each of these soils, we characterized the potential of soil Streptomyces to antagonize plant pathogens, using an in vitro plate assay with indicator strains to reveal inhibition. We cultivated each plant species in each conditioned soil to assess feedback effects on subsequent plant growth performance. Surrounding plant richness modified the impacts of particular plant species on Streptomyces antagonistic activity; A. gerardii supported a higher proportion of antagonistic Streptomyces when grown in monoculture than when grown in 32-spp plant communities, and L. capitata supported more strongly antagonistic Streptomyces when grown in 4- or 32-spp plant communities than in 8-spp plant communities. Similarly, the feedback effects of particular plant species sometimes varied with surrounding plant richness; aboveground biomass production varied with plant species richness for A. gerardii in L. perennis-trained soil, for L. capitata in A. gerardii-trained soil, and for L. perennis in L. capitata-trained soil. Streptomyces antagonist density increased with overall Streptomyces density under low but not under high plant richness, suggesting that plant diversity modifies selection for antagonistic phenotypes among soil Streptomyces. This work highlights the complexity of feedback dynamics among plant species, and of plant–microbiome interactions in soil.     

贵州野生枇杷资源遗传多样性的ISSR分析

本文采用ISSR标记，构建了贵州39份野生枇杷种质的DNA指纹图谱，并对其遗传多样性进行了评价。结果表明，以筛选出的15条引物进行PCR扩增，获得清晰、重复性好的谱带125条，其中多态性谱带108条，多态性比例为86.4%；应用引物815、825和841的组合，构建了39份枇杷种质的ISSR指纹图谱，可以有效区分所有供试材料；15条引物的谱带数据聚类分析表明，供试种质的遗传相似性系数为0.40~0.98，在相似性系数为0.70时，可将供试种质分成五大类。聚类结果与种质表型相关，与地理分布的相关性不明显。     

Allozyme Variation of Picea abies in Poland

The aim of this study was to determine the levels and patterns of allozyme variation among 29 populations of Norway spruce [ Picea abies (L.) Karst.] from Poland. Thirteen investigated isoenzyme systems were encoded by 27 gene loci and, on average, 71% of the loci per population were polymorphic. The average and effective numbers of alleles per locus were 2.17 and 1.26, respectively, while the expected heterozygosity was 0.156. A relatively low allozyme differentiation among populations from north-eastern and southern Poland was observed ( F ST = 0.028, mean genetic distance D = 0.005). The results suggest that historical events and extensive gene flow played an important role in the distribution of the observed allozyme differentiation of Norway spruce in Poland.     

Multivariate analysis and genetic diversity for morphometric and root textural quality traits in ashwagandha (Withania somnifera Dunal)

Ashwagandha (Withania somnifera) is an important medicinal plant and its dried roots are used in traditional systems of medicine. The market price of roots is determined by physical (textural) quality. Brittle roots with high starch and low fiber are considered to be superior because of ease in grinding. Genetic diversity studies based on root textural parameters have not been done so far. So the present study was designed to assess genetic diversity for morphometric traits and root textural quality parameters among two morphologically distinct groups: Poshita and Nagore. The PCA separated the morphometric and root texture variables distinctly into two different principal components: PC-1 and PC-2 respectively, indicating that both are negatively associated. All the morphotypes in Poshita group showed high positive loadings in PC-1 indicating that component genotypes are high root yielding. Nagore morphotypes were low yielding but the root texture was good. Clustering of morphotypes grouped Poshita and Nagore separately with high inter-cluster distances indicating that both groups are highly divergent from each other, suggesting that there is sufficient scope for varietal improvement through hybridization.