基于一维卷积神经网络和拉曼光谱的肺炎支原体菌株分类

肺炎支原体是造成人类呼吸系统疾病的主要原因.临床中,患者感染不同肺炎支原体症状极为相似,很难根据症状判别肺炎支原体类型并对症给药.因此,准确判别肺炎支原体菌株类型对于发病机理和疾病流行病学研究以及临床精准治疗具有重要意义.拉曼光谱具有快速、高效、无污染等优点,在生物医学领域逐渐得到越来越多研究者们的关注.一维卷积神经网...     

Raman labeled nanoparticles: characterization of variability and improved method for unmixing

Raman spectroscopy can differentiate the spectral fingerprints of many molecules, resulting in potentially high multiplexing capabilities of Raman‐tagged nanoparticles. However, an accurate quantitative unmixing of Raman spectra is challenging because of potential overlaps between Raman peaks from each molecule, as well as slight variations in the location, height, and width of very narrow peaks. If not accounted for properly, even minor fluctuations in the spectra may produce significant error that will ultimately result in poor unmixing accuracy. The objective of our study was to develop and validate a mathematical model of the Raman spectra of nanoparticles to unmix the contributions from each nanoparticle allowing simultaneous quantitation of several nanoparticle concentrations during sample characterization. We developed and evaluated an algorithm for quantitative unmixing of the spectra called narrow peak spectral algorithm (NPSA). Using NPSA, we were able to successfully unmix Raman spectra of up to seven Raman nanoparticles after correcting for spectral variations of 30% intensity and shifts in peak locations of up to 10 cm⁻¹, which is equivalent to 50% of the full width at half maximum (FWHM). We compared the performance of NPSA to the conventional least squares (LS) analysis. Error in the NPSA is approximately 50% lower than in the LS. The error in estimating the relative contributions of each nanoparticle with the use of the NPSA are in the range of 10–16% for equal ratios and 13–19% for unequal ratios for the unmixing of seven composite organic–inorganic nanoparticles (COINs); whereas, the errors from using the traditional LS approach were in the range of 25–38% for equal ratios and 45–68% for unequal ratios. Here, we report for the first time the quantitative unmixing of seven nanoparticles with a maximum root mean square of the percentage error (RMS%) error of less than 20%. Copyright © 2012 John Wiley & Sons, Ltd.     

Robust analysis of multiplexed SERS microscopy of Ag nanocubes using an alternating minimization algorithm

To drive the application of surface‐enhanced Raman spectroscopy (SERS) mapping in ex vivo diagnostic imaging and non‐biological material characterization, we have designed a robust and accurate multiplex spectral fitting method using an alternating minimization algorithm to extract individual constituent Raman spectra with very small overall fitting error (as low as 2%). For each mixed Raman signal, constituent spectra and mixture coefficients were estimated jointly based on reference spectra that were measured in the lab. Our method is based on a Poisson model to reflect the photon counting nature of Raman signals and includes the nonlinear noise in the measured data, making our method robust against data containing relatively large random noise. In our method, we minimized a cost function consisting of two terms: (1) the overall fitting error between the measured and modeled spectra and (2) the sum of the individual error between each reference spectrum and its corresponding constituent. This method inherently guarantees that the estimates will approach the global minimum with monotonic convergence. The accuracy of our method was validated by applying it to a SERS spectral fitting problem and comparing our results to those from existing methods. Copyright © 2013 John Wiley & Sons, Ltd.     

Research on the Raman spectral character and diagnostic value of squamous cell carcinoma of oral mucosa

Early diagnosis of oral carcinomas is essential for successful treatment. The purpose of the present study is to apply near‐infrared Raman spectroscopy to detect oral squamous cell carcinoma (SCC) and leukoplakia (OLK), in order to establish the diagnostic model of the Raman spectra of oral diseases. We collected Raman spectra of normal, OLK and SCC by near‐infrared Fourier transform Raman spectroscopy. The biochemical variations between different lesions were analyzed by the characteristic bands in the subtracted mean spectra. Gaussian radial basis function support vector machines (SVM) were used to classify spectra and establish the diagnostic models. Major differences were observed in the range between 800 and 1800 cm⁻¹. Compared with normal mucosa, high contents of protein, DNA and lipid in SCC and OLK were observed, but the difference between OLK and normal tissue was not as much as that between normal and SCC. SVM displayed a powerful ability in the classifying of normal and SCC, and the specificity, sensitivity and accuracy were 100, 97.56 and 98.75%, respectively. In discriminating between the OLK and normal groups, the three parameters were 85, 68 and 72.5%. The algorithm showed good ability in grouping and modeling of OLK and SCC, and the three parameters were 95, 97.43 and 96.25%. Combined with SVM, near‐infrared Raman spectroscopy can detect biochemical variations in oral normal mucosa, OLK and SCC, and establish diagnostic models accurately. Copyright © 2009 John Wiley & Sons, Ltd.     

Range‐independent background subtraction algorithm for recovery of Raman spectra of biological tissue

An important requirement for the use of Raman spectroscopy for tissue diagnostic applications is an appropriate algorithm that can faithfully retrieve weak tissue Raman signals from the measured raw Raman spectra. Although iterative modified polynomial‐fitting‐based automated algorithms are widely used, these are sensitive to the choice of the fitting range, thereby leading to significantly different Raman spectra for different start and stop wavenumber selection. We report here an algorithm for automated recovery of the weak Raman signal, which is range independent. Given a raw Raman spectrum and the choice of the start and the stop wavenumbers, the algorithm first truncates the spectrum to include the raw data within this wavenumber range, linearly extrapolates the truncated raw spectrum beyond the points of truncation on the two sides by using coefficients of linear least‐square fit, adds two Gaussian peaks of appropriate height and width on the extrapolated linear wings on either side and then iteratively smoothens the data with all these add‐ons such that the smaller of the ordinate values of the smoothed and the starting raw data serve as the input to each successive round of iterative smoothing until the added Gaussian peaks are fully recovered. The algorithm was compared with the modified polynomial‐based algorithms using mathematically simulated Raman spectrum as well as experimentally measured Raman spectra from various biological samples and was found to yield consistently range‐independent and artifact‐free Raman signal with zero baseline. Copyright © 2012 John Wiley & Sons, Ltd.     

基于匹配追踪的拉曼光谱信号重构算法

拉曼光谱技术是一种高灵敏度、无损伤、振动分子光谱技术,在医药、生物、分析化学等诸多领域有着重要的作用。然而,由于拉曼散射强度低,实际测得的拉曼信号容易被噪声所污染。特别是在较短的曝光时间,收集到的拉曼光谱的信噪比很低。因此,提出了一种基于匹配追踪算法的信号重构方法,用于提取低信噪比的拉曼信号。该方法首先通过阈值循环迭代的方法在平均谱上找出特征峰的位置、估计峰的区间。根据峰的位置区间等信息,用高斯密度函数生成字典。在噪声谱上,根据特征峰位置和区间,将其区分为有信号区间和无信号区间,在有信号区间上利用匹配追踪算法重构被噪声所掩盖的拉曼信号。该算法不仅能够很好的逼近掩盖在噪声中的拉曼信号,且在重构信号的过程中也会对基线进行扣除,无须作基线校正处理。在仿真和实验中对该算法与常规算法进行了比较,结果证明,该算法在低信噪比条件下能够较好的恢复拉曼信号。该算法不同于传统光谱去噪算法,能同时对拉曼光谱进行了基线扣除以及噪声的处理,且能取得较为理想的结果,不需要使用不同的算法对基线和噪声分别处理。其次,在算法上我们创造性地将匹配追踪算法用于拉曼光谱信号的稀疏逼近求解。     

Multipeak fitting analysis of Raman spectra on DLCH film

The hydrogenated diamond‐like carbon (DLCH) film with 1‐µm thickness is deposited by direct hydrocarbon gas ion beam method on silicon wafer and annealed at 400 °C. Detailed Raman spectra feature are fitted from nine sets of different peak fitting functions, including Gaussian, Lorentzian and Breit‐Wigner‐Fano (BWF) functions. These fitting results obtained from a two‐peak combination show some specific variances on the G peak position, FWHM_G and I_D/I_G ratio for as‐deposited and as‐annealed DLCH films. The most popular two‐peak fitting method with full Gaussian function tends to exhibit a higher ratio of the G peak position shift and higher I_D/I_G ratio than others fitting methods, the drastic difference among the most popular G (G) & G (D) and B (G) & L (D) schemes also have brought out in I_D/I_G ratio. However, for a more complex four‐peak Gaussian function fitting Raman spectra, the I_D/I_G ratio is close to that of a two‐peak fitting function with a mixture functions of BWF (G) and L (D). Furthermore, a series of systematic peak fitting procedures and comparisons of Raman spectra have been discussed in this study. Copyright © 2009 John Wiley & Sons, Ltd.     

Analysis of G‐quadruplex conformations using Raman and polarized Raman spectroscopy

G‐quadruplexes (G4s) are four‐stranded DNA structures formed within nucleic acid sequences that are rich in guanines. G4 formation within DNA strands is believed to have significant biological relevance for the control of cell replication and gene expression. Therefore, the development and validation of experimental techniques that can easily and reliably characterize G4 structures under biologically relevant measurement conditions, like Raman spectroscopy, are desirable for G4‐targeted structure based drug design. Here we report Raman and polarized Raman studies of solutions of three oligonucleotides, thrombin binding aptamer (TBA) 5′‐GGTTGGTGTGGTTGG‐3′, human telomeric (HT) 5′‐(TTAGGG)₄‐3′, and a modified c‐Myc NHE‐III₁ sequence (MycL1) 5′‐TGAGGGTGGGTAGGGTGGGTAA‐3′, which were previously reported to form four distinct intramolecular G4 structures in the presence of Na⁺ or K⁺, as determined by NMR. Our results support the previously proposed antiparallel (TBA), antiparallel and hybrid (HT), and parallel with double‐chain reversal (DCR) loop (MycL1) structures. Large sample‐dependent variations in the intensity of bands associated with deoxyribose backbone modes in the 840–930 cm⁻¹ and 1420–1460 cm⁻¹ spectral regions were observed. Most notably, a highly polarized deoxyribose ring symmetric stretch (~930 cm⁻¹) appeared strongly in the solution spectra for HT and TBA, but was very weak or absent in the solution spectrum for MycL1 and the drop deposition (dried sample) spectra for all three oligonucleotides. It is hypothesized that the intensity of this band is likely controlled by furanose ring structure uniformity and/or solvent accessibility to certain nucleotide binding sites. Raman depolarization ratios measured for the G4s in solution were generally very similar to those previously reported for canonical B DNA, with the possible exception of base ring modes that consistently yielded slightly lower depolarization ratios for G4s compared to B DNA. The results further underscore the utility of Raman and polarized Raman spectroscopy for G4 structure elucidation under biologically relevant solution conditions. Copyright © 2015 John Wiley & Sons, Ltd.     

OFDM水声通信CS限幅失真补偿与LS信道估计优化算法*

正交频分复用技术应用于水声通信系统时,会引起较高的峰均比,当采用限幅法对峰均比进行抑制时,会产生非线性失真。另外,系统采用最小二乘法进行信道估计受噪声的影响较大。针对以上问题,提出了一种基于压缩感知技术的补偿限幅非线性失真与最小二乘信道估计相组合的新算法,在接收端利用导频数据采用压缩感知算法对限幅失真进行补偿,同时对最小二乘信道估计进行优化运算,以减小噪声对其影响。理论分析和仿真结果均表明,新算法不但能够有效抑制限幅法造成的系统非线性失真问题,而且能够降低高斯白噪声对于信道估计的影响。     

An intelligent background‐correction algorithm for highly fluorescent samples in Raman spectroscopy

Fluorescent background is a major problem in recoding the Raman spectra of many samples, which swamps or obscures the Raman signals. The background should be suppressed in order to perform further qualitative or quantitative analysis of the spectra. For this purpose, an intelligent background‐correction algorithm is developed, which simulates manual background‐correction procedure intelligently. It basically consists of three aspects: (1) accurate peak position detection in the Raman spectrum by continuous wavelet transform (CWT) with the Mexican Hat wavelet as the mother wavelet; (2) peak‐width estimation by signal‐to‐noise ratio (SNR) enhancing derivative calculation based on CWT but with the Haar wavelet as the mother wavelet; and (3) background fitting using penalized least squares with binary masks. This algorithm does not require any preprocessing step for transforming the spectrum into the wavelet space and can suppress the fluorescent background of Raman spectra intelligently and validly. The algorithm is implemented in R language and available as open source software ( http://code.google.com/p/baselinewavelet ). Copyright © 2009 John Wiley & Sons, Ltd.     

Fast Poissonian image segmentation with a spatially adaptive kernel

The variational models with the goal of minimizing the local variation are widely used for the segmentation of the intensity inhomogeneous images recently. Local variation is a good measure for the images corrupted by Gaussian noise. However, in many applications such as astronomical imaging, electronic microscopy and positron emission tomography, Poisson noise often occurs in the observed images. To deal with this kind of images, we develop a novel segmentation model based on minimizing local generalized Kullback–Leibler (KL) divergence with a spatially adaptive kernel. A fast algorithm based on the split-Bregman method is proposed to solve the corresponding optimization problem. Numerical experiments for synthetic and real images demonstrate that the proposed model outperforms most of the current state-of-the-art methods in the present of Poisson noise.     

Rapid prediction of fatty acid composition of vegetable oil by Raman spectroscopy coupled with least squares support vector machines

The fatty acid composition of vegetable oil plays a significant role in a nutrition‐balanced diet, which makes this industry more quality conscious. A set of store‐purchased vegetable oils and their binary mixtures were characterized by Raman spectra in a region of 800–2000 cm⁻¹. The obtained Raman spectral data were pretreated, and intensities of eight characteristic peaks were extracted as the eigenvalues of an entire spectrum. A prediction model of fatty acid content based on least squares support vector machines (LS‐SVM) were established for multivariate analysis between the Raman spectral eigenvalues and the fatty acid composition measured by gas chromatography (GC) method. The performance of the model was evaluated by comparing the predicted values to the reference values from GC analysis. The correlation coefficient for the prediction of oleic acid, linoleic acid and α‐linolenic acid was 0.9972, 0.9982 and 0.9854, respectively. Raman spectroscopy based on LS‐SVM can be a promising technique for predicting the fatty acid composition of vegetable oil with the advantages of being simple and time‐effective while not requiring any sample preprocessing. In particular, a portable Raman system is suitable for on‐site detection in practical applications. Copyright © 2013 John Wiley & Sons, Ltd.     

Na_2SO_4溶液激光拉曼/激光诱导击穿光谱联合探测

光谱技术应用于海底极端环境下多参数、多相态、无接触探测已成为深海化学传感器发展的一个重要方向,尤其是水下激光拉曼光谱技术和水下激光诱导击穿光谱技术正成为目前研究开发的热点。该工作旨在探索一项水下激光诱导击穿光谱与激光拉曼光谱(LIBS-LRS)联合探测技术,以实现LIBS和拉曼两种检测技术在检测系统上的整合,在信息获取上的互补。在实验室搭建了一套LIBS-LRS联合探测装置,该装置对于拉曼和LIBS采用同样的激发光源、光谱仪和探测器,前置光路分为两部分:拉曼光路和LIBS光路,分别收集Na_2SO_4溶液的拉曼信号和LIBS信号。前置光路收集的拉曼和LIBS信号由Y型光纤导入光谱仪,分别在面阵CCD不同区域进行探测。利用该装置对配置的Na_2SO_4溶液进行探测,同时获得了Na元素的LIBS信号和SO~(2-)_4拉曼信号。另外,随着激光能量的提高,在532nm脉冲激光能量超过3.6mJ时,在拉曼光路同时获得了Na元素的LIBS信号和SO~(2-)_4拉曼信号,这样采用同一光路即可实现两种光谱技术的联合,然而实验发现,随着激光能量的增加,激光在溶液中击穿产生的轫致辐射造成了光谱探测基线整体的抬升,对拉曼光谱弱信号的探测是不利的。实验结果初步证明了在拉曼和LIBS在水下联合探测的可行性。     

基于拉曼光谱高斯分峰拟合的硫化锑含量检测方法

针对硫化锑含量采用化学分析方法检测存在操作复杂、检测时间长的问题,提出了一种基于拉曼光谱高斯分峰拟合的硫化锑含量快速检测方法。采用拉曼光谱系统测定锑矿样品的拉曼谱图,对原始拉曼谱图进行平滑去噪、背景扣除、谱段选择及归一化等预处理。基于高斯曲线的高斯峰、峰面积、半高宽和峰位置等信息,建立单个高斯峰的数学模型,提出高斯分峰拟合算法对光谱进行拟合,采用状态转移算法对高斯峰模型进行优化求解,得到表征光谱信息的特征参量,结合偏最小二乘回归方法,确定特征参量和硫化锑含量之间的关系,从而建立模型,实现对硫化锑含量的预测。实验中通过训练样本建立校正模型,对测试样本进行预测,同时从训练样本中随机挑选出检验样本,利用已建立的模型,对其硫化锑含量进行预测,以检验模型的正确性和外推性。实验结果表明：与预处理后全光谱建模相比,采用高斯分峰拟合后建立的预测模型的预测效果更好,证明了模型的正确性和良好外推性。因此,拉曼光谱结合高斯分峰拟合算法应用于锑矿中硫化锑含量的检测是可行的,且测量过程更简单,适用于矿物成分的快速分析。     

SVD‐based method for intensity normalization,background correction and solvent subtraction in Raman spectroscopy exploiting the properties of water stretching vibrations

A semiautomated method combining intensity normalization with effective elimination of the solvent signal and non‐Raman background is presented for Raman spectra of biochemical and biological analytes in aqueous solutions. The method is particularly suitable for rapid and effortless preprocessing of extensive datasets taken as a function of gradually varied physicochemical parameters, e.g. analyte and/or ligand concentration, temperature, pH, pressure, ionic strength, time, etc. For intensity normalization, the strong Raman OH stretching band of water in the range of 2700–3900 cm⁻¹ recorded together with the analyte spectrum in the fingerprint region below 1800 cm⁻¹ is employed as internal intensity standard. Concomitant dependences of the solvent Raman spectra are taken into account and, in some cases, turned into advantage. Once the Raman spectra of the solvent are acquired for a particular range of the parameter varied, solvent contribution can be subtracted correctly from any analyte spectrum taken within this range. The procedure presented can be efficiently applied only for the analytes having their own Raman signal in the range of OH stretching vibrations much weaker than that of the solvent. However, this is the case for a great number of biochemical and biological samples. Accuracy, reliability and robustness of the method were tested under the conditions of spontaneous Raman, resonance Raman and surface‐enhanced Raman scattering. Serviceability of the method is demonstrated by several real‐world examples. Copyright © 2011 John Wiley & Sons, Ltd.     

小波多尺度正交校正在近红外牛奶成分测量中的应用

光谱分析中,干扰信号的存在直接影响所建分析模型的质量。基于信号和干扰的不同特性,提出了一种扣除背景和噪声干扰的新方法——小波多尺度正交校正(WMOSC)法。首先将原始光谱进行小波变换(DWT),消除噪声及背景信息,然后采用正交信号校正(OSC)滤除与待测组分浓度无关的全部信息。与单纯的小波变换及正交信号校正相比,WMOSC能有效地扣除背景和噪声干扰,使模型具有更强的抗干扰能力,提高了模型的预测精度。利用该方法对牛奶样品的近红外光谱进行处理,采用偏最小二乘法建立校正模型,其牛奶中脂肪、蛋白质和乳糖的预测均方根误差(RMSEP)分别为0.101 6%,0.087 1%和0.110 7%。实验结果表明该方法能有效地去除干扰,保留有用信息。     

Surface‐enhanced Raman spectroscopy on novel black silicon‐based nanostructured surfaces

Two different black silicon nanostructured surfaces modified with thin gold layers were tested for analytical signal enhancement with Surface‐Enhanced Raman Spectroscopy (SERS). The relationship between the thicknesses of the gold layers and the analytical signal enhancement was studied. Also, effects of Ti and Ti/Pt adhesion layers underneath the gold layers on the analytical signal enhancement were tested. An enhancement factor of 7.6 × 10⁷ with the excitation laser 785 nm was achieved for the tested analyte, Rhodamine 6G, and non‐resonance SER spectra were recorded in a 5 s acquisition mode. Such an enhancement enables to achieve a detection limit down to 2.4 pg of Rhodamine 6G on a black silicon‐based nanosurface coated with a 400‐nm‐thin layer of gold. Copyright © 2009 John Wiley & Sons, Ltd.     

Improved blind‐source separation for spectra

A new method is proposed that improves the performance of independent component analysis (ICA) algorithms for separating overlapping spectra under certain circumstances. The method is designed for Raman spectra in particular, although it should be applicable to spectra with similar line shapes, such as nuclear magnetic resonance. In the zero‐noise case, conventional ICA fails to separate synthetic Raman spectra completely; by maximising smoothness, as opposed to other more traditional measures of statistical independence, better performance can be achieved. The new method is tested against artificial and real Raman spectra, and demonstrates improved performance for each. The algorithm is fully automated, requiring no user input or judgement in order to separate the spectra. Copyright © 2011 John Wiley & Sons, Ltd.     

小波包熵的复杂体系近红外光谱信息提取

有用信息提取是复杂体系近红外检测的重点和难点之一。由于复杂体系光谱中存在各种噪声、基线漂移、谱带重叠及复杂背景的干扰,常规方法不能准确地从光谱中获得有用信息。为此,将小波包变换(DWPT)和信息熵理论相结合--小波包熵(EWPIE)提取复杂体系光谱中的有用信息。思路是采用小波包变换对光谱信号进行多频带分解,根据有用信号与噪声的频带分布特点,基于信息熵理论滤除干扰的频率分量,采用正交校正法(OSC)剔除与被测组分无关的信息,然后对处理后的频率分量进行重构,从而实现复杂体系有用信息的准确提取。通过对复杂体系光谱数据建立多元校正模型来验证该方法的效果。采用牛奶的近红外光谱数据,以牛奶中脂肪和蛋白质浓度为研究对象,建立了偏最小二乘法(PLS)模型。结果显示,牛奶中脂肪和蛋白质的预测均方根误差(RMSEP)分别为0.132%和0.121%,与单纯的DWPT和OSC相比,EWPIE能够有效地提取有用信息,避免了无用信息的干扰,明显提高了模型的预测精度,对复杂体系的准确检测具有一定的理论意义和实际应用价值。     

Fabrication and characterization of SERS‐active silver clusters on glassy carbon

In this article, a novel technique for the fabrication of surface enhanced Raman scattering (SERS) active silver clusters on glassy carbon (GC) has been proposed. It was found that silver clusters could be formed on a layer of positively charged poly(diallyldimethylammonium) (PDDA) anchored to a carbon surface by 4‐aminobenzoic acid when a drop containing silver nanoparticles was deposited on it. The characteristics of the obtained silver clusters have been investigated by atomic force microscopy (AFM), SERS and an SERS‐based Raman mapping technique in the form of line scanning. The AFM image shows that the silver clusters consist of several silver nanoparticles and the size of the clusters is in the range 80–100 nm. The SERS spectra of different concentrations of rhodamine 6G (R6G) on the silver clusters were obtained and compared with those from a silver colloid. The apparent enhancement factor (AEF) was estimated to be as large as 3.1 × 10⁴ relative to silver colloid, which might have resulted from the presence of ‘hot‐spots’ at the silver clusters, providing a highly localized electromagnetic field for the large enhancement of the SERS spectra of R6G. The minimum electromagnetic enhancement factor (EEF) is estimated to be 5.4 × 10⁷ by comparison with the SERS spectra of R6G on the silver clusters and on the bare GC surface. SERS‐based Raman mapping technique in the form of line scanning further illustrates the good SERS activity and reproducibility on the silver clusters. Finally, 4‐mercaptopyridine (4‐Mpy) was chosen as an analyte and the lowest detected concentration was investigated by the SERS‐active silver clusters. A concentration of 1.6 × 10⁻¹⁰ M 4‐Mpy could be detected with the SERS‐active silver clusters, showing the great potential of the technique in practical applications of microanalysis with high sensitivity. Copyright © 2006 John Wiley & Sons, Ltd.