化学计量学方法用于蛋白质组学质谱数据的特征筛选

提出了一种基于偏最小二乘判别分析和F-score的特征筛选方法,并将其用于蛋白质组学质谱数据分析。方法主要包含3个步骤:(1)用LIMPIC算法对原始数据进行预处理;(2)计算每个变量的F-score值并将所有变量按F-score值降底的顺序排列;(3)采用偏最小二乘判别分析交互检验按前向选择法选择最佳变量子集。用本方法对一组卵巢癌数据进行分析,最终从原始的15154个质荷比变量中选择了12个特征变量作为潜在生物标记物,它们在训练集上交叉检验的特异性和灵敏度分别为98.36%和98.15%,在独立测试集上的特异性和灵敏度分别为96.67%和100%。用筛选出的变量作PCA所得的结果显示这些变量能够较好地将样本分类,说明能够反映出样本的类别信息。所提出的方法可用于蛋白质组学质谱数据的特征筛选及样本分类。     

基于气相色谱-质谱联用技术的痛风病人血清代谢特征分析

痛风是一组仅见于人类的异质性疾病,随着时间的推移,将导致慢性关节炎并逐渐致残。该研究将基于气相色谱-质谱联用技术(GC-MS)的代谢组学方法应用于痛风病人的血清代谢特征分析。首先利用GCMS获得痛风病人和健康人的血清代谢指纹图谱,采用多变量统计分析对所得数据进行分析。主成分分析(PCA)得分图显示,痛风病人与健康人的血清代谢谱有差异。通过偏最小二乘-判别分析(PLS-DA)对样品进行进一步分型,根据模型的变量重要性因子(VIP值)及非参数检验结果筛选差异代谢物。共筛选出43种可能与痛风相关的代谢物,并对其中22个变量进行结构鉴定,主要包括丙二醇、2,3-二羟基丁酸、2,4-二羟基丁酸、赤藓糖醇、苏糖醇、苏糖酸、阿拉伯糖醇、D-葡萄糖酸、肌醇、次黄嘌呤、尿酸、尿苷、3-羟基-3-甲基丁酸、鸟氨酸、吲哚-3-乳酸、单乙醇胺、甘油、甘油酸、月桂酸及亚油酸等代谢物。与健康人相比,痛风病人的糖代谢、核苷酸代谢、氨基酸代谢及脂类代谢均发生了明显的紊乱。这些结果将为痛风临床诊断及治疗提供重要依据。     

基于支持向量机的近红外特征变量选择算法用于树种快速识别

将稳定度自适应重加权采样特征变量选择算法用于支持向量机定性分析(Support vector machine-stability competitive adaptive reweighted sampling,SVM-SCARS)。该算法通过对数据多次采样建模计算各变量的稳定度值,稳定度值能更加客观准确地评估变量在建模中的作用,因此可作为变量重要性的评价依据。通过循环迭代方式,采用自适应重加权采样技术逐步筛选变量,然后以每次循环所得变量子集建立SVM模型,并以模型交叉验证分类正确率(Correct classification rate of cross validation,CCRCV)评估子集优劣,确定最优特征变量子集。将该算法结合漫反射近红外光谱技术建立了制浆造纸常用木材的树种识别模型,实现了对4种桉木和2种相思木的快速识别分类。最终共筛选出15个特征变量建立分类模型,模型对各树种分类的正确率达97.9%,具有较好的分类效果。与全光谱模型和递归特征消除支持向量机模型相比,SVM-SCARS能够筛选出更少的特征变量,且模型具有更好的预测性能和稳定性。研究结果表明,SVM-SCARS算法能够有效优化光谱特征变量,提高近红外在线分析模型在木材材性分析中的稳健性和适用性。     

基于QSAR模型预测有机化合物的生物富集因子

生物富集因子(BCF)是REACH法规要求的生物积累危害评价指标的关键参数，用实验的方法测定BCF值人力物力花费较大，而用计算的方法预测BCF值可以有效地替代昂贵的实验过程。该研究使用E-Dragon计算了数据集中每个分子的1 666种描述符，并用筛选后的描述符与lgBCF建立了QSAR模型：采用随机森林与支持向量机建立的分类预报模型，随机森林分类模型的准确率为0.89、敏感度为0.89;用基于准确率递减和基尼系数方法的随机森林分类模型筛选出对lgBCF值有重要影响的30个描述符，这些描述符包括ALOGPS＿lgP、MATS6v、TPSA.NO.、GATS7v等；ALOGPS＿lgP和ALOGPS＿lgS是用支持向量机分类模型筛选出的对lgBCF值有重要影响的描述符。     

基于结合位点的辅酶A结合蛋白家族的分类

发展了一种基于分子相互识别的蛋白质分类方法, 应用数据挖掘策略与统计学聚类, 根据辅酶A (coenzyme-A, CoA)结合蛋白的结合模式特征数据, 通过对比和分析多种分类方法对该体系的分类准确度, 对这类体内重要的蛋白进行了分类方法学研究, 选择了最优的两步聚类法. 本研究工作设计和建立了一个分类参数, 可以简洁有效地评价出各个结合特征的显著性与重要性, 并以此为依据从所有特征中筛选出决定性的特征变量. 研究结果所得到的CoA结合蛋白的三个分类, 都具有显著的氢键与疏水结合特征; CoA可以与多个生物活性关键氨基酸残基形成氢键作用. 这些相互作用的共性及分类上的差异, 说明了配体与不同受体相互作用过程中结合模式上的细微差别, 对于以CoA结合蛋白为靶点的选择性调控分子设计具有重要的参考意义与指导作用.     

主成分分析-支持向量机用于肝病分类模型

将主成分分析(PCA)用于肝功能检测数据特征提取,然后用支持向量机(SVM)对乙肝、丙肝、肝硬化、正常人样本建立分类模型。采用高斯径向基函数(RBF)为核函数,调节核函数参数C及σ以建立最佳支持向量机模型。该模型对训练集的识别率为99.3%,对预测集的预测率为96.4%。结果表明:PCA-SVM法建立的肝病分类模型能较好的区分乙肝、丙肝、肝硬化及正常人,且分类效果优于传统支持向量机及人工神经网络(ANN)分类模型。     

电感耦合等离子体质谱(ICP-MS)法测定地质样品中Cu和Zn的干扰研究

电感耦合等离子体质谱(ICP-MS)法标准模式测定地质样品中Cu和Zn时,样品溶液中Ti在高温下易形成TiO⁺和TiOH⁺产生离子重叠干扰,使测定值偏离真值而影响测定结果。采用氢氟酸-高氯酸-硝酸消解、王水复溶样品,通过多级在线校正⁴⁶Ti¹⁶O⁺,消除TiO⁺和TiOH⁺离子重叠干扰,研究了单质Ti与⁴⁶Ti¹⁶O⁺、质荷比63、64、65、66离子间存在的关系,建立了一种电感耦合等离子体质谱(ICP-MS)法测定地质样品中Cu和Zn的分析方法。结果表明：单质Ti对Cu和Zn产生的干扰值与其质量浓度呈非简单的正相关性,TiO⁺产生的干扰离子质荷比65离子强度比63高,质荷比64离子强度比66高;⁴⁶Ti¹⁶O⁺对质荷比63、66离子干扰增量值与单质Ti质量浓度变化无关;^6...     

基于显微拉曼检测蛋白核小球藻鉴别丁草胺及草甘膦

以蛋白核小球藻(Chlorella pyrenoidosa)作为鉴别载体,利用共聚焦显微拉曼光谱仪分别获取生长在除草剂草甘膦、丁草胺污染水体以及正常水体的蛋白核小球藻β-胡萝卜素的拉曼光谱信息,对2种除草剂进行了鉴别.利用预处理后的光谱信号,建立偏最小二乘回归(PLS)预测模型及线性判别分析(LDA)分类模型.当阈值为±0.3时,全波段建立的PLS模型对草甘膦和丁草胺的预测正确率高达83.33%,特征峰建立的LDA分类模型对2种除草剂的分类正确率均达到了100%.结果表明,利用蛋白核小球藻为载体对丁草胺和草甘膦2种除草剂进行鉴别是可行的,且LDA分类模型更适合除草剂的分类研究.     

基于人血浆荧光光谱与CARS-PLS-LDA的结直肠癌早期检测

该文利用竞争性自适应加权算法（CARS）筛选重要的人血浆荧光光谱变量,并结合偏最小二乘法判别分析（PLS－LDA）建立了结直肠癌患者与非癌患者的分类模型,同时与全波长模型和基于平行因子分析（PARAFAC）建立的模型进行比较。从模型评价指标看,CARS－PLS－LDA的性能显著优于全波长模型和基于PARAFAC的模型。高波未稀释组和低波稀释组的荧光光谱结合CARS－PLS－LDA分类模型的AUC（Area under curve）值均高于0.9,可有效地识别结直肠癌患者。结果表明,CARS变量筛选能够明显改善结直肠癌分类模型的性能,有助于后续癌症临床诊断工具的开发与研究。     

因子分析法在质谱成像数据分析中的应用

对因子分析法在质谱成像数据分析中的应用进行了研究。本方法分析的质谱成像数据来源于空气动力辅助离子源质谱成像技术,所用样品为含有3种不同颜料(红色、蓝色、黑色)的笔迹样品。对该样品的成像数据进行因子分析后,将成像数据分为了背景、黑色、蓝色和红色因子。分析结果显示,m/z 443.2,478.4,322.2(344.2)分别在红色、蓝色、黑色因子中的贡献值远大于其它质荷比,因此是3种颜料的特征质荷比。此结果与实际情况相符,证明使用因子分析方法对质谱成像数据进行分析和特征提取是可行的。对因子分析与主成分分析的成像数据处理结果进行了比较,结果显示,因子分析可以更简单和定量地对特征质荷比进行取舍,在生物标志物提取、疾病诊断、药理分析等方面有较大的应用潜力。     

A novel colorimetric PCR-based biosensor for detection and quantification of hepatitis B virus

Hepatitis B virus (HBV) can cause viral infection that attacks the liver and it is a major global health problem that put people at a high risk of death from cirrhosis of the liver and liver cancer. HBV has infected one third of the worldwide population, and 350 million people suffer from chronic HBV infection. For these reasons, development of an accurate, sensitive and expedient detection method for diagnosing, monitoring and assessing therapeutic response of HBV is very necessary and urgent for public health and disease control. Here we report a new strategy for detection of viral load quantitation of HBV based on colorimetric polymerase chain reaction (PCR) with DNAzyme-containing probe. The special DNAzyme adopting a G-quadruplex structure exhibited peroxidase-like activity in the presence of hemin to report colorimetric signal. This method has shown a broad range of linearity and high sensitivity. This study builds important foundation to achieve the specific and accurate detection level of HBV DNA with a low-cost and effective method in helping diagnosing, preventing and protecting human health form HBV generally all over the world and especially in developing countries.     

Mutation detection in the drug-resistant hepatitis B virus polymerase gene using nanostructured reverse micelles.

The emergence of drug-resistant hepatitis B virus (HBV) has been reported in patients with prolonged administration of lamivudine, which is a potent drug for the prevention of HBV infection. Lamivudine-resistant HBV has several types of mutations at the YMDD motif of its DNA polymerase. We successfully demonstrated that monitoring the hybridization behavior in nanostructured reverse micelles enables us to detect single nucleotide polymorphisms (SNPs). With the aid of reverse micelles, a model 40-mer oligonucleotide containing a single-base substitution was clearly distinguished from the normal, complementary oligonucleotide. In addition, we extended this technique to a high-throughput analysis. The results obtained with a 96-well micro-plate reader indicated the possibility of SNPs detection toward multiple samples of patients.     

血清代谢轮廓分析在评估乙肝肝硬化病程中的应用

采用高效液相色谱-质谱联用(HPLC-MS)作为代谢组学研究平台,分析不同Child-Pugh分级肝硬化病人和健康人群的血清标本,获取代谢轮廓.对数据进行主成分分析(PCA)和正交偏最小二乘判别分析(OPLS-DA),用各组病例的80%作为训练数据构建疾病的OPLS-DA区分模型,以剩余的20%作为检测数据,观察模型对...     

化学发光成像法测定人体血清中乙型肝炎病毒

基于对碘苯酚增强的luminol-H2O2-HRP化学发光反应,利用化学发光成像法检测乙肝病毒(HBV)。用该法对人体血清中的乙型肝炎表面抗原、表面抗体、e抗原、e抗体以及核心抗体进行测定,其结果与ELISA法所得结果一致,对表面抗原检测结果为阳性的病人血清测定9次,结果的相对标准偏差为4．2％。     

Current Progress in the Development of Hepatitis B Virus Capsid Assembly Modulators: Chemical Structure,Mode-of-Action and Efficacy

Hepatitis B virus (HBV) is a major causative agent of human hepatitis. Its viral genome comprises partially double-stranded DNA, which is complexed with viral polymerase within an icosahedral capsid consisting of a dimeric core protein. Here, we describe the effects of capsid assembly modulators (CAMs) on the geometric or kinetic disruption of capsid construction and the virus life cycle. We highlight classical, early-generation CAMs such as heteroaryldihydropyrimidines, phenylpropenamides or sulfamoylbenzamides, and focus on the chemical structure and antiviral efficacy of recently identified non-classical CAMs, which consist of carboxamides, aryl ureas, bithiazoles, hydrazones, benzylpyridazinones, pyrimidines, quinolines, dyes, and antimicrobial compounds. We summarize the therapeutic efficacy of four representative classical compounds with data from clinical phase 1 studies in chronic HBV patients. Most of these compounds are in phase 2 trials, either as monotherapy or in combination with approved nucleos(t)ides drugs or other immunostimulatory molecules. As followers of the early CAMs, the therapeutic efficacy of several non-classical CAMs has been evaluated in humanized mouse models of HBV infection. It is expected that these next-generation HBV CAMs will be promising candidates for a series of extended human clinical trials.     

Metabolite identification of small interfering RNA duplex by high-resolution accurate mass spectrometry

On-line liquid chromatography/electrospray ionization high-resolution mass spectrometry (LC/ESI-HRMS) using an LTQ-Orbitrap mass spectrometer was employed to investigate the metabolite profiles of a model siRNA duplex designated HBV263. The HBV263 duplex was incubated in rat and human serum and liver microsomes in vitro. The siRNA drug and its metabolites were then extracted using a liquid-liquid extraction followed by solid-phase extraction (LLE-SPE), and analyzed by LC/ESI-MS. High-resolution accurate mass data enabled differentiation between two possible metabolite sequences with a monoisotopic molecular mass difference of less than 1 Da. ProMass deconvolution software was used to provide semi-automated data processing. In vitro serum and liver microsome incubation samples afforded different metabolite patterns: the antisense strand of the duplex was degraded preferentially in rat and human serum, while the sense strand of the duplex was less stable in rat and human liver microsomes.     

Pattern Search Method for Determination of DAEM Kinetic Parameters from Nonisothermal TGA Data of Biomass

The most accurate and up-to-date approach to modeling biomass pyrolysis is to adopt the distributed activation energy model (DAEM). In this study, a pattern search method to be used for the determination of DAEM kinetic parameters from the nonisothermal thermogravimetric analysis (TGA) data of biomass has been introduced. The method has been applied to the nonisothermal TGA data of peanut shell sample, and DAEM kinetic parameters of biomass samples have been determined. Calculated model results from determined kinetic parameters have been compared with nonisothermal TGA data of biomass.     

The use of the NRTLmKW model for predicting VLE data of binary systems from the excess enthalpy data

A new local composition model NRTLmKW has been used for correlation of 15 binary excess enthalpy data. The data, binary systems formed by hydrocarbons and alkanols, have been selected to give a wide representation of various kinds molecular interactions in solution. Further the model, basing on the results of the correlation, has been used for prediction of excess enthalpy and VLE (vapour-liquid equilibria) in these systems. The obtained results have been discussed from the point of view of intermolecular interactions and some recommendations have been made on the use of the NRTLmKW model for such calculation.