Multi-label learning algorithm with SVM based association

Multi-label learning is an active research area which plays an important role in machine learning. Traditional learning algorithms, however, have to depend on samples with complete labels. The existing learning algorithms with missing labels do not consider the relevance of labels, resulting in label estimation errors of new samples. A new multi-label learning algorithm with support vector machine(SVM) based association(SVMA) is proposed to estimate missing labels by constructing the association between different labels. SVMA will establish a mapping function to minimize the number of samples in the margin while ensuring the margin large enough as well as minimizing the misclassification probability. To evaluate the performance of SVMA in the condition of missing labels, four typical data sets are adopted with the integrity of the labels being handled manually. Simulation results show the superiority of SVMA in dealing with the samples with missing labels compared with other models in image classification.

SNP site-drug association prediction algorithm based on denoising variational auto-encoder

Single nucletide polymorphism(SNP)is an important factor for the study of genetic variation in human families and animal and plant strains.Therefore,it is widely used in the study of population genetics and disease related gene.In pharmacogenomics research,identifying the association between SNP site and drug is the key to clinical precision medication,therefore,a predictive model of SNP site and drug association based on denoising variational auto-encoder(DVAE-SVM)is proposed.Firstly,k-mer algorithm is used to construct the initial SNP site feature vector,meanwhile,MACCS molecular fingerprint is introduced to generate the feature vector of the drug module.Then,we use the DVAE to extract the effective features of the initial feature vector of the SNP site.Finally,the effective feature vector of the SNP site and the feature vector of the drug module are fused input to the support vector machines(SVM)to predict the relationship of SNP site and drug module.The results of five-fold cross-validation experiments indicate that the proposed algorithm performs better than random forest(RF)and logistic regression(LR)classification.Further experiments show that compared with the feature extraction algorithms of principal component analysis(PCA),denoising auto-encoder(DAE)and variational auto-encode(VAE),the proposed algorithm has better prediction results.

Characteristics and advantages of adeno-associated virus vector-mediated gene therapy for neurodegenerative diseases

Common neurodegenerative diseases of the central nervous system are characterized by progressive damage to the function of neurons, even leading to the permanent loss of function. Gene therapy via gene replacement or gene correction provides the potential for transformative therapies to delay or possibly stop further progression of the neurodegenerative disease in affected patients. Adeno-associated virus has been the vector of choice in recent clinical trials of therapies for neurodegenerative diseases due to its safety and efficiency in mediating gene transfer to the central nervous system. This review aims to discuss and summarize the progress and clinical applications of adeno-associated virus in neurodegenerative disease in central nervous system. Results from some clinical trials and successful cases of central neurodegenerative diseases deserve further study and exploration.

序列稀疏自回归方法及其在美股做空数据分析上的应用

采用序列稀疏回归的思路来处理向量自回归模型,并设计适用于大规模时间序列数据分析的序列稀疏自回归方法。研究表明:从因子角度刻画向量自回归模型可以有效地将稀疏矩阵估计问题分解成稀疏奇异向量的估计问题,从而极大地提高了计算效率。以1523家美股上市公司1973年1月—2014年12月的做空数据为例,利用此方法探索公司之间的大规模做空关联网络。研究发现:此方法可以有效地恢复股票做空份额(即某一公司的空头股份数量)与股票收益率之间隐藏的关联网络,对于股票风险溢价研究具有一定启发意义。

考虑关联波段特性的光谱相似图像分类方法

光谱相似图像分类性能过差会增加光谱信息冗余度,降低地物勘探与军事防御等多种领域的光谱探测效率。为了多元素匀质区分光谱信息与光谱曲线,提出考虑关联波段特性的光谱相似图像分类方法。该方法首先利用光谱匹配消除光谱相似图像白色光源过曝现象。然后提取优化图像的关联波段,并将其作为聚类特征输入支持向量机中。最后根据支持向量机的输出结果,实现光谱相似图像分类。实验结果表明,所提方法分类结果清晰度较高,分类误差或像素块填色错误小,混淆矩阵中同行同列矩形块的分类精度较高。

Support Vector的几何解释及其在联想记忆中的应用

利用闭凸集上的投影解释 support vector的几何意义 ,利用支持超平面讨论线性分类器的设计问题 .对线性可分情形 ,Support vector由一类数据集合闭凸包在另一类数据集合闭凸包上投影的非零系数向量组成 ,SVM所决定的超平面位于两投影点关于各自数据集合支持超平面的中间 .作为应用 ,文中给出一种设计理想联想记忆前馈神经网络的方法 ,它是 FP算法的一般化 .

基于SVDD和改进K-Means的变压器故障诊断模型

变压器状态对于智能配电房的安全稳定运行具有重要意义。为实现对变压器故障的准确诊断,在变压器油中溶解气体分析(DGA)的基础上,提出了一种联合使用支持向量数据描述(SVDD)和改进K-Means聚类的变压器故障诊断方法。首先利用SVDD构造闭合分类曲面实现“正常”和“故障”两类判断,然后对“故障”类样本进行K-Means聚类分析,自动将其划分为低能放电、中低温过热、高能放电、高温过热和局部放电5种故障类型,同时针对K-Means初始聚类中心选取难题,提出局部密度概念自动确定K-Means初始聚类中心,提升聚类性能。最后利用变压器故障真实数据开展实验,结果表明,相较于支持向量机(SVM)和BP神经网络模型,所提方法的故障诊断准确率分别提升9.8%和8%。

混合属性网络多维多层关联数据智能挖掘算法

针对传统关联数据挖掘算法,强项集挖掘后产生大量候选项集,导致挖掘耗时长、挖掘精度低等问题,提出一种混合属性网络多维多层关联数据智能挖掘算法(Multidimensional Multilayer Associative Data Intelligent Mining Algorithm,MMAD-IM)。计算混合属性网络中随机数据到簇中心的距离,将目标数据分配到距离簇中心最近的簇中,使簇中心固定,完成混合属性网络数据的聚类分析。从聚类完成的数据中提取出有效的基本频繁向量,同时计算数据的候选项集,对哈希表进行扫描,利用改进Apriori算法完成强项集挖掘。以此为基础构建空间关系,获取近似区域与近似点之间的距离,形成待挖掘数据并计算数据的隶属度数值,完成智能挖掘。实验结果表明,所提算法具有较好的数据聚类效果,强项集挖掘后剩余的候选项集数量较少,整体数据挖掘耗时远低于传统算法,挖掘精度高达90%。

AAV基因治疗产品的免疫反应评估

随着生命科学技术的进步,基因治疗因其特异性和精准性等优势已成为新药研发的重点领域。其中,腺相关病毒(adeno-associated virus,AAV)载体是一种应用前景广阔的体内基因治疗载体,也是基因治疗产品载体研究的热点。目前,全球范围内已有7款AAV相关的基因治疗产品获批上市,并有多项临床试验正在进行。AAV载体的免疫原性较低,但部分人群体内存在针对AAV的预存免疫,其进入体内后也会诱导机体产生先天免疫反应,随后可能触发针对AAV衣壳和转基因产物的适应性免疫反应,从而影响药物的有效性和安全性。本综述旨在概述AAV基因治疗引起的免疫反应特征,并简要总结在临床前和临床研究中对AAV基因治疗免疫反应的评估,以期为AAV基因治疗药物评价提供参考。

基于PCA多模型融合的滚动轴承性能退化指标构建

单模型构建的滚动轴承性能健康指标仅能从本身的“单角度”来描述滚动轴承的性能退化状态,具有一定的局限性。为解决这个问题,提出一种基于主成分分析(principal component analysis,PCA)多模型融合的滚动轴承健康指标构建方法。该方法分别采用支持向量数据描述(support vector data description,SVDD)模型、自联想核回归(auto-associative kernel regression,AAKA)模型和高斯混合模型(gaussian mixture module,GMM)构建相应单模型的健康指标,再将3个单模型的健康指标经主成分分析(PCA)融合,并选取第一主成分作为能够包含“多角度”性能退化信息的健康指标(SAG-HI)。试验结果表明,相比于各单模型的健康指标,SAG-HI与滚动轴承保持可靠度的灰置信水平达到98.38%,其相关性、单调性和鲁棒性也均表现为最优,且通过包络谱分析验证了其能够准确且及时监测到早期故障发生时刻。

A Modified Moist Ageostrophic Q Vector

The quasi-geostrophic Q vector is an important rainfall associated with large-scale weather systems diagnostic tool for studying development of surface and is calculated using data at single vertical level. When ageostrophic Q vector was introduced, it required data at two vertical levels. In this study, moist ageostrophic Q vector is modified so that it can be calculated using data at a single vertical level. The comparison study between the original and modified moist ageostrophic Q vectors is conducted using the data from 5 to 6 July 1991 during the torrential rainfall event associated with the Changjiang-Huaihe mei-yu front in China. The results reveal that divergences of original and modified moist ageostrophic Q vectors have similar horizontal distributions and their centers are almost located in the precipitation centers. This indicates that modified moist ageostrophic Q vector can be used to diagnose convective development with reasonable accuracy.

Targeting adeno-associated virus and adenoviral genetherapy for hepatocellular carcinoma

Human hepatocellular carcinoma(HCC)heavily endangers human heath worldwide.HCC is one of most frequent cancers in China because patients with liver disease,such as chronic hepatitis,have the highest cancer susceptibility.Traditional therapeutic approaches have limited efficacy in advanced liver cancer,and novel strategies are urgently needed to improve the limited treatment options for HCC.This review summarizes the basic knowledge,current advances,and future challenges and prospects of adeno-associated virus(AAV)and adenoviruses as vectors for gene therapy of HCC.This paper also reviews the clinical trials of gene therapy using adenovirus vectors,immunotherapy,toxicity and immunological barriers for AAV and adenoviruses,and proposes several alternative strategies to overcome the therapeutic barriers to using AAV and adenoviruses as vectors.

Adenovirus vector expressing mda-7 selectively kills hepatocellular carcinoma cell line Hep3B

BACKGROUND: Melanoma differentiation associated gene-7 (mda-7) is a novel tumor suppressor gene, which has suppressor activity in a broad spectrum of human cancer cells both in vitro and in vivo through activation of various intracellular signaling pathways. In this study, we investigated the potential effect of mda-7 on human hepatocellular carcinoma (HCC) in vitro. METHODS: Cells from the human HCC cell line Hep3B and the human liver cell line L-02 were assigned to three groups. One was cultured in Dulbecco's modified Eagle's medium without serum (control). The others were transfected with adenovirus expressing the mda-7 gene (Ad.mda-7) or adenovirus vector serving as negative control (Ad.vec). The expression of MDA-7 and Bcl-2 proteins in Hep3B and L-02 cells was confirmed by the reverse transcriptase-polymerase chain reaction and enzyme-linked immunosorbent assay. The methyl thiazolyl tetrazolium colorimetric assay and flow cytometry were used to assess tumor cell proliferation and the cell cycle. Hoechst and Annexin-V/propidium iodide staining were used to study mda-7 gene expression in Hep3B and L-02 cells. The expression of MDA-7, Bcl-2 and Bax proteins were detected by Western blotting. RESULT S: The mda-7 gene was expressed in Hep3B and L-02 cells. The protein concentrations of MDA-7 in supernatants were 790 and 810 pg/ml, respectively. mda-7 induced Hep3B growth suppression and apoptosis, compared with Ad.mda-7 and control (P<0.01). In addition, cell block in G2/M was identified by exposure of HCC cells to secreted MDA-7 protein, but this was not found in L-02. The gene expression of Bcl-2 was markedly decreased in Hep3B but not in L-02. CONCLUSIONS: mda-7 selectively induces growth inhibition and apoptosis in the HCC cell line Hep3B but not in the normal liver cell line L-02 via downregulating the antiapoptosis protein Bcl-2. It could be an ideal gene for gene therapy in HCC.

ADENO-ASSOCIATED VIRUS INTRODUCED INTEGRATION AND EXPRESSION OF FOREIGN GENES IN PC12 CELLS

Objective To investigate integration and expression of adeno-associated virus (AAV) vectors in neuronal PC12 cells,the result of which can be applied in further gene therapy of diseases of the central nervous sys- tem. Methods Human neurotrophin-3(hNT3)genes were inserted into AAV vectors. Then the recombinat AAV plas- mids were encapsidated as recombinant virions. PCl2 cells were transfected with the virions and the positive cells were selected by G418. The transfection positive (hNT3 modified)PC12 cells were cultured for several generations and the cellular genomic DNA and total RNA were extracted. We investigated the integration locus or AAV vectors by South- ern blot and transcript situation or foreign genes by dot blot. Results The hybridization tests showed that AAV in- troduced foreign genes were stably integrated, but at random locus, and robustly transcribed in hNT3 modified PCl2 cells. Conclusion AAV vectors can serve as high efficiency vectors or target genes in neuronal PC12 cells.

MicroRNA-regulated viral vectors for gene therapy

Safe and effective gene therapy approaches require targeted tissue-specific transfer of a therapeutic transgene.Besides traditional approaches, such as transcriptional and transductional targeting, micro RNA-dependent posttranscriptional suppression of transgene expression has been emerging as powerful new technology to increase the specificity of vector-mediated transgene expression. Micro RNAs are small non-coding RNAs and often expressed in a tissue-, lineage-, activation- or differentiation-specific pattern. They typically regulate gene expression by binding to imperfectly complementary sequences in the 3' untranslated region(UTR) of the m RNA. To control exogenous transgene expression, tandem repeats of artificial micro RNA target sites are usually incorporated into the 3' UTR of the transgene expression cassette, leading to subsequent degradation of transgene m RNA in cel s expressing the corresponding micro RNA. This targeting strategy, first shown for lentiviral vectors in antigen presenting cells, has now been used for tissue-specific expression of vector-encoded therapeutic transgenes, to reduce immune response against the transgene, to control virus tropism for oncolytic virotherapy, to increase safety of live attenuated virus vaccines and to identify and select cell subsets for pluripotent stem cell therapies, respectively. This review provides an introduction into the technical mechanism underlying micro RNA-regulation, highlights new developments in this field and gives an overview of applications of micro RNA-regulated viral vectors for cardiac, suicide gene cancer and hematopoietic stem cell therapy, as well as for treatment of neurological and eye diseases.

Anatomical distributional defects in mutant genes associated with dominant intermediate Charcot-Marie-Tooth disease type C in an adenovirus-mediated mouse model

Dominant intermediate Charcot-Marie-Tooth disease type C（DI-CMTC） is a dominantly inherited neuropathy that has been classified primarily based on motor conduction velocity tests but is now known to involve axonal and demyelination features.DI-CMTC is linked to tyrosyl-t RNA synthetase（YARS）-associated neuropathies,which are caused by E196 K and G41 R missense mutations and a single de novo deletion（153-156 del VKQV）.It is well-established that these YARS mutations induce neuronal dysfunction,morphological symptoms involving axonal degeneration,and impaired motor performance.The present study is the first to describe a novel mouse model of YARS-mutation-induced neuropathy involving a neuron-specific promoter with a deleted mitochondrial targeting sequence that inhibits the expression of YARS protein in the mitochondria.An adenovirus vector system and in vivo techniques were utilized to express YARS fusion proteins with a Flag-tag in the spinal cord,peripheral axons,and dorsal root ganglia.Following transfection of YARS-expressing viruses,the distributions of wild-type（WT） YARS and E196 K mutant proteins were compared in all expressed regions; G41 R was not expressed.The proportion of Flag/green fluorescent protein（GFP） double-positive signaling in the E196 K mutant-type mice did not significantly differ from that of WT mice in dorsal root ganglion neurons.All adenovirus genes,and even the empty vector without the YARS gene,exhibited GFP-positive signaling in the ventral horn of the spinal cord because GFP in an adenovirus vector is driven by a cytomegalovirus promoter.The present study demonstrated that anatomical differences in tissue can lead to dissimilar expressions of YARS genes.Thus,use of this novel animal model will provide data regarding distributional defects between mutant and WT genes in neurons,the DICMTC phenotype,and potential treatment approaches for this disease.

基于词嵌入技术的心理学研究:方法及应用

词嵌入是自然语言处理的一项基础技术。其核心理念是根据大规模语料中词语和上下文的联系,使用神经网络等机器学习算法自动提取有限维度的语义特征,将每个词表示为一个低维稠密的数值向量(词向量),以用于后续分析。心理学研究中,词向量及其衍生的各种语义联系指标可用于探究人类的语义加工、认知判断、发散思维、社会偏见与刻板印象、社会与文化心理变迁等各类问题。未来,基于词嵌入技术的心理学研究需要区分心理的内隐和外显成分,深化拓展动态词向量和大型预训练语言模型(如GPT、BERT)的应用,并在时间和空间维度建立细粒度词向量数据库,更多开展基于词嵌入的社会变迁和跨文化研究。我们为心理学专门开发的R语言工具包PsychWordVec可以帮助研究者利用词嵌入技术开展心理学研究。

不同注射剂量9型腺相关病毒载体在小鼠脑和脊髓表达效果分析

腺相关病毒(AAV),尤其是从轴突末端转移到神经元的AAV9型,是研究神经环路和解析大脑功能的有力工具。利用囊泡型谷氨酸神经元(VGlut2)基因敲入小鼠联合脑立体定向注射,通过将AAV9-DIO-ChR2-EGFP病毒注射于VGlut2转基因小鼠的蓝斑核(LC),探究不同注射剂量对病毒侵染范围和小鼠体重的影响。结果表明,与高注射量(1μL)相比,低注射量(0.5μL)可将AAV9-DIO-ChR2-EGFP限定至蓝斑核(LC)。在AAV9-DIO-ChR2-EGFP注射3周后,侵染范围由LC延伸至初级运动皮质(primary motor cortex,M1)、下丘脑、中缝苍白核(araphe pallidus nucleus,Rpa)和颈髓(cervical spinal cord,CSC)。随后检测了低注射量的AAV9(0.5μL)是否影响小鼠体重,统计结果表明,0.5μL AAV9注射对小鼠体重无影响(P=0.9795)。结果表明,低注射量(0.5μL)可以在一定程度上限制AAV的表达范围,并且脑内注射AAV9不会改变小鼠体重。

非病毒纳米载体递送CRISPR/Cas9的研究进展

成簇规则间隔的短回文重复序列及其相关蛋白9(clustered regularly interspaced short palindromic repeat/CRISPR-associated protein 9,CRISPR/Cas9)技术目前广泛应用于生命医学领域的基础研究及临床应用研究。由于载体在CRISPR/Cas9技术中发挥了重要的作用,如何进一步开发和优化载体系统具有重要的意义。传统的载体大多以病毒载体为主,其递送的效率高,但亦存在插入片段的大小有限、免疫反应、致癌、难以大规模生产甚至脱靶等缺陷;而非病毒纳米载体在一定程度上可解决基因编辑过程中由病毒载体所带来的潜在毒性和容量限制等问题,可能具有更广阔的应用前景。本文主要综述了目前用于CRISPR/Cas9系统递送的非病毒纳米载体,探讨了非病毒纳米载体在递送CRISPR/Cas9系统时可能遇到的主要困难,提出了相应的解决方案和策略,以期为基因治疗和药物研发提供新的参考依据。

基于SVM-BP混合网络的输电线路安全评价模型研究

为了能够科学、客观地评价输电线路安全状态,提出了基于SVM-BP混合网络的输电线路安全分析模型。首先,对天气数据、输电线路标准文件和缺陷数据等进行处理和分析,利用关联规则和主成分分析法对输电线路缺陷数据进行整合,建立输电线路安全状态评价体系;然后,提出了SVM-BP混合网络模型挖掘各因素与缺陷状态间的关联,评价输电线路的安全状态;最后,以某地区输电线路为数据集进行验证,该评价模型的准确率达到97.4%,可以快速准确地评价输电的安全状态,并在灾害性天气下提前做好应对措施,保证输电线路的稳定运行。