基于部首嵌入和注意力机制的病虫害命名实体识别

为了解决农业病虫害命名实体识别过程中存在的内在语义信息缺失、局部上下文特征易被忽略和捕获长距离依赖能力不足等问题，以农业病虫害文本为研究对象，提出一种基于部首嵌入和注意力机制的农业病虫害命名实体识别模型(Chinese agricultural diseases and pests named entity recognition with joint radical embedding and self attention, RS-ADP)。首先，该模型将部首嵌入集成到字符嵌入中作为输入，用以丰富语义信息。其中，针对部首嵌入设计了3种特征提取策略，即卷积神经网络(Convolutional neural network, CNN)、双向长短时记忆网络(Bidirectional long short term memory network, BiLSTM) 和CNN-BiLSTM；其次，采用多层不同窗口尺寸的CNNs层提取不同尺度的局部上下文信息；然后，在BiLSTM提取全局序列特征的基础上，采用自注意力机制进一步增强模型提取更长距离依赖的能力；最后，采用条件随机场(Conditional random field, CRF)联合识别实体边界和划分实体类别。在包含11个类别和24715条标注样本的农业病虫害自制语料上进行了实验。结果表明，本文模型RS-ADP在该数据集上精确率、召回率和F1值分别为94.16%、94.47%和94.32%；在具体实体类别上，RS-ADP在作物、病害、虫害等易识别实体上F1值高达95.81%、97.76%和97.23%。同时，RS-ADP在草害、病原等难以识别实体上F1值仍保持86%以上。实验结果表明，本文所提模型能够有效识别农业病虫害命名实体，其识别精度优于其他模型，且具有一定的泛化性。     

全双工双通路中继网络干扰处理与信号检测算法

全双工双通路连续中继(FD-TPSR)网络频谱效率高,状态控制简单,但全双工中继存在残留自干扰(RSI),发射中继对接收中继也会形成中继间干扰(IRI),会导致通信性能下降.这里提出一种中继间干扰处理方法,中继节点消除部分IRI,以提高端到端信干噪比;保留部分IRI,在目的节点构成延迟转发编码结构,以提供分集增益.同时提出一种基于并行软干扰消除的匹配滤波(MF-PSIC)信号检测算法,匹配滤波器结构简单,实现复杂度低,基于软输出的并行干扰消除能同时检测所有时隙的符号,处理时延小.仿真结果表明,中继间干扰处理方法兼顾了分集增益和累积干扰与噪声因素,相比于无IRI消除和完全IRI消除,误比特率最低; MF-PSIC信号检测算法实现复杂度低,相比于ML信号检测算法仅有很少的性能损失.     

基于网络安全等级保护2.0的农业科研单位网络安全体系研究——以中国农业科学院为例

[目的/意义]大数据时代,中国农业科研单位持续加强信息化建设,给科学研究和科研管理带来极大的便利,但相应的信息安全问题也开始逐渐显现。依托信息等级保护工作,深入分析农业科研单位信息系统建设中各个薄弱环节,提高信息系统安全,避免发生网络安全事件已被摆在农业科研单位工作的突出地位。[方法/过程]按照等级保护制度要求,论文主要围绕加强农业科研单位信息安全管理的内容进行探讨,介绍了信息等级保护的重要性,对比分析了等级保护2.0标准与等级保护1.0标准的不同;详细分析中国农业科学院的网络安全管理制度体系和网络安全管理技术等内容,最后对加强农业科研单位信息安全建设的有效策略提出了相关建议。[结果/结论]论文以中国农业科学院为例详细论述其相关做法,希望为各农业科研机构开展网络安全等级保护安全建设提供借鉴。     

面向养殖水体分布差异的COD光谱法检测研究

面向养殖水体，传统光谱法对化学需氧量（Chemical Oxygen Demand，COD）检测模型构建的基础：源域（现有样本库）与目标域（检测地水体）间光谱数据独立同分布。但是当源域与目标域分布间存在差异时，由源域得到的低误差模型常在目标域上表现下滑。针对该问题，提出面向UV Vis光谱的域对抗训练网络（DAUVwpNet），将分布不同的源域和目标域数据映射至相同分布的特征空间中，使其在该空间的分布距离尽可能接近，从而在特征空间中对源域训练的目标函数也可以迁移至目标域上，以降低模型在目标域的误差。试验表明：面向同一批测试数据，DAUVwpNet的预测误差为0.78，要低于传统模型的预测误差（0.85）；DAUVwpNet预测值与实测值间相关系数为0.95，要高于传统模型的相关系数（0.89）。表明了该网络能够较好对齐两域特征空间数据分布，降低因分布差异带来的COD检测误差。     

An overview of animal science research 1945–2011 through science mapping analysis

The conceptual structure of the field of Animal Science (AS) research is examined by means of a longitudinal science mapping analysis. The whole of the AS research field is analysed, revealing its conceptual evolution. To this end, an automatic approach to detecting and visualizing hidden themes or topics and their evolution across a consecutive span of years was applied to AS publications of the JCR category ‘Agriculture, Dairy & Animal Science’ during the period 1945–2011. This automatic approach was based on a coword analysis and combines performance analysis and science mapping. To observe the conceptual evolution of AS, six consecutive periods were defined: 1945–1969, 1970–1979, 1980–1989, 1990–1999, 2000–2005 and 2006–2011. Research in AS was identified as having focused on ten main thematic areas: ANIMAL‐FEEDING, SMALL‐RUMINANTS, ANIMAL‐REPRODUCTION, DAIRY‐PRODUCTION, MEAT‐QUALITY, SWINE‐PRODUCTION, GENETICS‐AND‐ANIMAL‐BREEDING, POULTRY, ANIMAL‐WELFARE and GROWTH‐FACTORS‐AND‐FATTY‐ACIDS. The results show how genomic studies gain in weight and integrate with other thematic areas. The whole of AS research has become oriented towards an overall framework in which animal welfare, sustainable management and human health play a major role. All this would affect the future structure and management of livestock farming.     

硒蛋白X基因克隆、定点突变、原核表达及多克隆抗体制备

本试验旨在克隆鉴定猪硒蛋白X(Sel X)基因,将其定点突变后进行原核表达,获得重组蛋白,并制备猪Sel X多克隆抗体,为以猪为模型Sel X功能研究奠定基础。利用c DNA末端快速克隆(3'-RACE)技术从猪肝脏总RNA扩增出含开放阅读框(ORF)至poly(A)共1 215 bp的Sel X基因c DNA片段,其348 bp的ORF编码区和对应的氨基酸残基与人相应序列分别有88%和91%序列同源性,猪Sel X基因提交NCBI Gen Bank数据库,序列号为EF113597。ORF编码氨基酸残基中硒代半胱氨酸(Sec)位于C-端第95个残基,其编码密码子TGA经定点突变为半胱氨酸(Cys)的TGT后,将其插入载体p ET30,转入大肠杆菌BL21(DE3)中,在0.5 mmol/L异丙基硫代-β-D-半乳糖苷(IPTG)诱导2 h获得融合表达产物,经His-tag亲和层析后,获得分子质量为18.0 ku的纯化蛋白。将纯化获得的Sel X基因突变体融合蛋白(Sel Xm)免疫兔子,得到效价高达1∶20 000的多克隆抗体,免疫印迹(Western-blot)结果显示该抗体能特异性识别纯化的Sel Xm和猪肝脏中相应Sel X。本试验成功克隆并鉴定了猪Sel X基因,并制备了其多克隆抗体。     

基于无人机高光谱影像的玉米地上生物量反演

【目的】研究以玉米地上干生物量为研究对象,探讨基于无人机高光谱数据利用人工神经网络法反演生物量的可行性。【方法】在吉林省蔡家镇开展玉米氮肥梯度试验,并进行无人机高光谱数据和地上干生物量获取,共获数据30组。随机选22组数据用于建模,剩下8组用于模型的外部验证。分别基于光谱指数法和BP神经网络算法构建反演模型,比较分析各种方法反演玉米生物量的优劣。【结果】结果表明：和基于光谱指数构建的生物量反演模型相比,BP神经网络模型取得了更好的反演结果。其建模时决定系数为0.99均方根误差为0.08 t/ha,相对均方根误差为3.39%；外部验证时,决定系数为0.99,均方根误差为0.15 t/ha,相对均方根误差为8.56%。【结论】BP神经网络模型可有效提高无人机高光谱遥感反演玉米地上生物量的精度。     

Meta-analysis of fungicide efficacy on soybean target spot and cost–benefit assessment

Target spot of soybean has spread in Brazil, the southeastern United States and Argentina in the last decade. A collaborative network of field Uniform Fungicide Trials (UFT) in Brazil was created in 2011 to study the target spot control efficacy of fungicides, including azoxystrobin + benzovindiflupyr (AZ_BF), carbendazim (CZM), fluxapyroxad + pyraclostrobin (FLUX_PYRA), epoxiconazole + FLUX_PYRA (EPO_FLUX_PYRA), mancozeb (MZB) and prothioconazole + trifloxystrobin (PROT_TRIF). Network meta-analysis was used to conduct a quantitative synthesis of UFT data collected from 2012 to 2016 and to evaluate the effects of disease pressure (DP, low ≤ 35% target spot severity in the nontreated control < high) and year of experiment on the overall mean efficacy and yield response to each of the tested fungicides. Based on mean percentage control of target spot severity, the tested fungicides fall into three efficacy groups (EG): high EG, FLUX_PYRA (76.2% control relative to the nontreated control) and EPO_FLUX_PYRA (75.7% control); intermediate EG, PROT_TRIF (66.5% control) and low EG, MZB (49.6% control), AZ_BF (46.7% control) and CZM (32.4% control). DP had a significant effect on yield response. At DP_Low, the highest response was due to PROT_TRIF (+342 kg ha⁻¹, +12.8%) and EPO_FLUX_PYRA (+295.5 kg ha⁻¹, +11.2%), whereas at DP_High, EPO_FLUX_PYRA and FLUX_PYRA outperformed the other treatments, with yield responses of 503 kg ha⁻¹ (+20.2%) and 469 kg ha⁻¹ (+19.1%), respectively. The probability of a positive return on fungicide investment ranged from 0.26 to 0.56 at DP_Low and from 0.34 to 0.66 at DP_High.     

A data transfer method for improving seed identification of maize (Zea mays) haploid breeding based on genetic similarity

Maize haploid breeding technology is able to identify haploid seeds non‐destructively, rapidly and at low cost with the help of Near‐infrared (NIR) spectral analysis. However, due to the hybridization of numerous parents and the low production rate of haploid, the haploid data collection becomes a burden for engineering this technology. Biologically, there are considerable similarities between the progeny of the same female parent and different male parents. Based on this advantage, similar spectral data can be transferred when the NIR technology is employed. A revised method of Transfer adaptive boost (TrAdaBoost) is proposed to improve identifying for the backpropagation neural network (BPNN) classifier. To avoid the negative transfer, a screening thresh is used to select out similar data, and the amount of these data are limited to join current training. The results show that the identification performances are improved significantly when the data amount is small. This method shows a high ability to make the seed identification more convenient for engineering maize haploid breeding.     

利用WGCNA进行玉米花期基因共表达模块鉴定

权重基因共表达网络分析(weighted gene co-expression network analysis, WGCNA)是系统生物学的一种研究方法,在挖掘生物学数据与特定性状之间的生物学关系方面具有十分重要的作用。本研究利用玉米(Zea mays L.)自交系B73的14份不同发育阶段的转录组数据,筛选掉低表达丰度的基因,最终得到了22,426个高表达的基因用于创建基因表达矩阵;利用不同组织作为性状,创建表型矩阵。然后利用R软件中的WGCNA包建立了共表达网络,共得到20个模块。本研究将与组织相关性高于0.65的模块定义为组织特异性模块,最终鉴定到14个组织特异性模块。利用在线网站Agrigo对组织特异性模块中的基因进行GO (gene ontology)富集分析,发现14个模块中均可以得到富集种类。开花作为玉米生育周期中的一个重要生理过程,不仅代表着植物从营养生长到生殖生长的转变,也关系到产量、株高和抗逆性等农艺性状。本研究发现8个组织特异性模块中的基因可以富集到与开花调控的代谢通路。此外,有17个已经报道过的开花时间调控基因存在于共表达模块中,并且主要分布在Blue模块和Darkmagenta模块,因此本研究重点关注了这2个模块内部的基因调控网络。本研究通过计算不同组织中的基因表达丰度,并联合权重基因共表达网络分析的方法,鉴定到了具有生物学意义的共表达基因模块,挖掘到了数个开花相关的模块,有助于揭示玉米开花调控的遗传机制。