A continuous wavelet transform algorithm for peak detection

Contactless conductivity detector technology has unique advantages for microfluidic applications. However, the low S/N and varying baseline makes the signal analysis difficult. In this paper, a continuous wavelet transform-based peak detection algorithm was developed for CE signals from microfluidic chips. The Ridger peak detection algorithm is based on the MassSpecWavelet algorithm by Du et al. [Bioinformatics 2006, 22, 2059-2065], and performs a continuous wavelet transform on data, using a wavelet proportional to the first derivative of a Gaussian function. It forms sequences of local maxima and minima in the continuous wavelet transform, before pairing sequences of maxima to minima to define peaks. The peak detection algorithm was tested against the Cromwell, MassSpecWavelet, and Linear Matrix-assisted laser desorption/ionization-time-of-flight-mass spectrometer Peak Indication and Classification algorithms using experimental data. Its sensitivity to false discovery rate curve is superior to other techniques tested.     

Data processing in metabolic fingerprinting by CE-UV: application to urine samples from autistic children

Metabolic fingerprinting of biofluids such as urine can be used to detect and analyse differences between individuals. However, before pattern recognition methods can be utilised for classification, preprocessing techniques for the denoising, baseline removal, normalisation and alignment of electropherograms must be applied. Here a MEKC method using diode array detection has been used for high-resolution separation of both charged and neutral metabolites. Novel and generic algorithms have been developed for use prior to multivariate data analysis. Alignment is achieved by combining the use of reference peaks with a method that uses information from multiple wavelengths to align electropherograms to a reference signal. This metabolic fingerprinting approach by MEKC has been applied for the first time to urine samples from autistic and control children in a nontargeted and unbiased search for markers for autism. Although no biomarkers for autism could be determined using MEKC data here, the general approach presented could also be applied to the processing of other data collected by CE with UV-Vis detection.     

Metabolic fingerprinting of Schistosoma mansoni infection in mice urine with capillary electrophoresis

Schistosoma mansoni infection in mice has been fingerprinted using CE to study the capabilities of this technique as a diagnostic tool for this parasitic disease. Two modes of separation were used in generating the electrophoretic data, with each untreated urine sample the following methods were applied: (i) a fused-silica capillary, operating with an applied potential of 18 kV, in micellar EKC (MEKC) and (ii) a polyacrylamide-coated capillary, operating with an applied potential of -20 kV under zonal CZE conditions. By combining normal and reverse polarities in the data treatment we have extracted more information from the samples, which is a better approach for CE metabolomics. The traditional problems associated with variability in electrophoretic peak migration times for analytes were countered by using a dynamic programming algorithm for the electropherograms alignment. Principal component analyses of these aligned electropherograms and partial least square discriminant analysis (PLS-DA) data are shown to provide a valuable means of rapid and sample classification. This approach may become an important tool for the identification of biomarkers, diagnosis and disease surveillance.     

Fabrication and performance of a three-dimensionally adjustable device for the amperometric detection of microchip capillary electrophoresis

A microchip CE-amperometric detection (AD) system has been fabricated by integrating a two-dimensionally adjustable CE microchip and an AD cell containing a one-dimensionally adjustable disk detection electrode in a Plexiglas holder. It facilitates the precise 3-D alignment between the channel outlet and the detection electrode without a complicated 3-D manipulator. The performance of this unique system was demonstrated by separating five aromatic amines (1,4-phenyldiamine, aniline, 2-methylaniline, 4-chloroaniline, and 1-naphthylamine) of environmental concern. Factors influencing their separation and detection processes were examined and optimized. The five analytes have been well separated within 140 s in a 74 cm long separation channel at a separation voltage of +2500 V using a 10 mM phosphate buffer (pH 3.5). Highly linear response is obtained for the five analytes over the range 20-200 microM with the detection limits ranging from 0.46 to 1.44 microM, respectively. The present system demonstrated long-term stability and reproducibility with RSDs of less than 5% for the peak current (n = 9). The new approach for the microchannel-electrode alignment should find a wide range of applications in CE, flowing injection analysis, and other microfluidic analysis systems.     

Automated detection of differently expressed fragments in mRNA differential display

We present a novel method for the automated detection of fragments showing dissimilar expression in mRNA differential display. The analysis is based on aligning the numerical electrophoretic lane data in respect of a given distance function defined on a set of fragments, or signal peaks in general. We presume that significant dissimilarities between peaks result in extreme score values computed for aligned peak pairs. Whereas in sequence comparison, an overall sequence similarity score is conventionally used, the current method defines a special dissimilarity score for searching the peak pairs showing the largest relative differences between the lanes. The output of the analysis is a highly reduced list of peak pairs, along with a set of associated features extracted from the lanes. Only the peaks of this list need to be visually confirmed instead of the vast amount of peaks in the original electrophoretic results. The results obtained by the algorithm correlate well with results of visual evaluation of the same electropherograms. The current algorithm may be applied to the study of complex expression patterns in multiple lanes and, in general, to automated recognition of variously defined patterns of quantitative electrophoretic data.     

Improvements in pentosan polysulfate sodium quality assurance using fingerprint electropherograms

Complex samples from polymer production, plant extracts or biotechnology mixtures can be characterized by fingerprints. Currently, the standard approach for sample characterization employs near-infrared (NIR) spectroscopy fingerprinting. Up to now, however, fingerprints obtained by chromatography or electrophoresis could only be visually evaluated. This type of inspection is very labor-intensive and difficult to validate. In order to transfer the use of fingerprints from spectroscopy to electrophoresis, spectra-like properties must be obtained through a complete alignment of the electropherograms. This has been achieved by interpolation and wavelet filtering of the baseline signal in the present work. The resulting data have been classified by several algorithms. The methods under survey include self-organizing maps (SOMs), artificial neural networks (ANNs), soft independent modeling of class analogy (SIMCA) and k-nearest neighbors (KNNs). In order to test the performance of this combined approach in practice, it was applied to the quality assurance of pentosan polysulfate (PPS). A recently developed capillary electrophoresis (CE) method using indirect UV detection was employed in these studies [1]. All algorithms were well capable of classifying the examined PPS test batches. Even minor variations in the PPS composition, not perceptible by visual inspection, could be automatically detected. The whole method has been validated by classifying various (n = 400) unknown PPS quality assurance samples, which have been correctly identified without exception.     

Quantitative and qualitative precision improvements by effective mobility-scale data transformation in capillary electrophoresis analysis

By transforming the time-based x-axis of electropherograms in capillary zone electrophoresis (CZE) into the corresponding effective mobility-scale, we propose a simple and robust data representation for a better qualitative and quantitative capillary electrophoresis (CE) analysis. The time scale of the raw electrophoretic data (detection signal versus time) is transformed into an effective electrophoretic mobility scale (mu eff-scale) with account of the electroosmotic flow (EOF) peak or of an internal standard of known effective mobility. With the new scaling (detection signals versus effective mobility), the obtained electropherograms are more representative of the velocity-based electrophoretic separation and the comparison of complete electropherograms is directly possible. This is of importance when tracking peaks in real samples where alteration in EOF stability can occur or when comparing electrophoretic runs from different experimental setups (independence in column length and voltage). Beside the qualitative possibilities, a quantitative improvement is achieved in the mu eff-scale with significant better peak area reproducibility and equal to more precision in quantitative analysis than with the primary time-scale integration.     

Two-dimensional wavelet analysis based classification of gas chromatogram differential mobility spectrometry signals

This study introduces two-dimensional (2-D) wavelet analysis to the classification of gas chromatogram differential mobility spectrometry (GC/DMS) data which are composed of retention time, compensation voltage, and corresponding intensities. One reported method to process such large data sets is to convert 2-D signals to 1-D signals by summing intensities either across retention time or compensation voltage, but it can lose important signal information in one data dimension. A 2-D wavelet analysis approach keeps the 2-D structure of original signals, while significantly reducing data size. We applied this feature extraction method to 2-D GC/DMS signals measured from control and disordered fruit and then employed two typical classification algorithms to testify the effects of the resultant features on chemical pattern recognition. Yielding a 93.3% accuracy of separating data from control and disordered fruit samples, 2-D wavelet analysis not only proves its feasibility to extract feature from original 2-D signals but also shows its superiority over the conventional feature extraction methods including converting 2-D to 1-D and selecting distinguishable pixels from training set. Furthermore, this process does not require coupling with specific pattern recognition methods, which may help ensure wide applications of this method to 2-D spectrometry data.     

Multi‐scale Gaussian/Haar wavelet strategies coupled with sub‐window factor analysis for an accurate alignment in nontargeted metabolic profiling to enhance herbal origin discrimination capability

Metabolic dataset can provide an overview of different herbal origin, which is conducted by some statistical procedures. Such results often deviate to a certain degree, due to peaks shifts in chromatographic signals. In order to solve this problem, an improved algorithm of combining sub‐window factor analysis with the mass spectrum information is proposed. The algorithm uses a peak detection approach derived either from multi‐scale Gaussian function or Haar wavelet to locate the peaks with different application scope; the candidate drift points at each peak are estimated by Fast Fourier transform cross correlation; Specifically, the best drift points at each candidate peaks are confirmed by sub‐window factor analysis and mass spectrum information in nontargeted metabolic profiling. Finally, the peak regions were aligned against a reference chromatogram, and the non‐peak regions were used linear interpolation. The chromatographic signals of 30 Bupleurum samples were aligned as an illustration of this algorithm, and they could be well distinguished using some statistical procedures. The result demonstrates that the presented method is stronger than other mass‐spectra based algorithms, when facing the alignment of some co‐eluted peaks.     

On-column silver substrate synthesis and surface-enhanced Raman detection in capillary electrophoresis

A new, simple, and efficient approach for on-column surface-enhanced Raman scattering (SERS) detection in capillary electrophoresis (CE) is reported. A ∼50-μm SERS substrate spot was prepared by laser-induced growth of silver particles in the 100-μm inner diameter CE capillary window or in a flow cell consisting of a 250-μm inner diameter fused silica capillary connector. For this purpose, the Raman laser was focused by a 20× objective into the detection window filled with a 0.5 mM silver nitrate and 10 mM citrate buffer solution. During the CE runs, the silver substrate spot was formed in a few seconds after the analyte injection, hence the analytes adsorbed sequentially to the silver surface when the detection window was reached, followed by desorption from the silver surface and continuing the electrophoretic migration to the capillary end. Thus, beyond migration time, valuable molecular specific information was delivered by the SERS spectra. Accurate separations and high-intensity SERS spectra are shown by CE-SERS time-dependent 3D electropherograms for the analytes rhodamine 6G, 4-(2-pyridylazo)resorcinol (PAR), PAR complex with Cu(II) and methylene blue at 0.25–25 ppm concentrations, by using 1.4–3.6 mW HeNe laser power and an acquisition time of 5 s for each spectrum. Before and after each analyte passes the detection window, clean background spectra were recorded and no memory effects perturbed the SERS detection. The silver substrate is characterized by a fast preparation rate, good reproducibility, a preparation success rate of over 95% and no mentionable influence on the electrophoretic migration time, the CE-SERS and CE-UV electropherograms being in good agreement. The successful coupling of CE and on-column SERS detection opens new perspectives for monitoring CE separations.     

Simplified 2-D CE-MS mapping: analysis of proteolytic digests

It has been demonstrated that CE-MS is a very useful hyphenated technique for proteomic studies. However, the huge amount of data stored in a single CE-MS run makes it necessary to account with procedures able to extract all the relevant information made available by CE-MS. In this work, we present a new and easy approach capable of generating a simplified 2-D map from CE-MS raw data. This new approach provides the automatic detection and characterization of the most abundant ions from the CE-MS data including their mass-to-charge (m/z) values, ion intensities and analysis times. It is demonstrated that visualization of CE-MS data in this simplified 2-D format allows: (i) an easy and simultaneous visual inspection of large datasets, (ii) an immediate perception of relevant differences in closely related samples, (iii) a rapid monitoring of data quality levels in different samples, and (iv) a fast discrimination between comigrating polypeptides and ESI-MS fragmentation ions. The strategy proposed in this work does not rely on an excellent mass accuracy for peak detection and filtering, since MS values obtained from an IT analyzer are used. Moreover, the methodology developed works directly with the CE-MS raw data, without interference by the user, giving simultaneously a simplified 2-D map and a much easier and more complete data evaluation. Besides, this procedure can easily be implemented in any CE-MS laboratory. The usefulness of this approach is validated by studying the very similar trypsin digests from bovine, rabbit and horse cytochrome c. It is demonstrated that this simplified 2-D approach allows specific markers for each species to be obtained in a fast and simple way.     

基于有限脉冲响应滤波器的实时小波算法及其在色谱信号解析中的应用

对色谱信号进行正确、实时地解析是利用色谱仪器系统进行过程检测的关键技术。小波分析方法能对色谱信号进行有效解析,但实时性一直是制约该方法在色谱信号处理中广泛应用的一个瓶颈。多分辨率分析与重构算法(MALLAT)是小波分析的一种快速算法,但是该算法的实时性仍然需要进一步提高。针对这一问题,以MALLAT算法原理为基础,利用在分解重组过程中两组滤波器系数之间的关系提出了一种基于有限脉冲响应滤波器的实时小波分析算法。仿真结果表明,与经典的MALLAT小波快速算法相比,该算法在保持信号有效分解与重构的基础上,运行耗时明显缩短,实时性得到较大提高。     

Confirmation of peak assignments in capillary electrophoresis using immunoprecipitation. Application to D-aspartate measurements in neurons

Capillary electrophoresis (CE) with laser-induced fluorescence (LIF) detection is a powerful tool for analysis of samples ranging from tissue extracts to single cells. However, accurate peak identification in electropherograms is challenging when complex biological samples are analyzed, as often matching a migration time between an analyte and corresponding standard may be insufficient to confirm the peak's identity. A method which combines single-step immunoprecipitation and CE-LIF analysis for investigation of the chiral amino acids in single cells and small tissue samples is demonstrated. D-Aspartate (D-Asp) has been reported in the central nervous system of the invertebrate neurobiological model Aplysia californica. In order to confirm the identity of D-Asp signal in the complex electropherograms of nerve tissue extracts and individual neurons, anti-D-Asp serum, preincubated with L-Asp conjugate, is added to the sample. This selectively binds the free D-Asp, creating an antibody-antigen complex with a migration time similar to that of antibody alone, but not that of D-Asp. The complete disappearance of the putative D-Asp peak confirms its identity and validates that there are no other detectable analytes co-migrating with D-Asp in the electropherogram.     

Thermal marks as a signal processing aid for a portable capillary electropherograph

The interpretation of raw signals in capillary CE can be challenging if there are unknown peaks, or the signal is corrupt due to baseline fluctuations, EOF velocity drift, etc. Signal processing could be required before results can be interpreted. A suite of signal processing algorithms has been developed for CE data analysis, specifically for use in field experiments for the detection of nerve agents using portable CE instruments. Everything from baseline correction and electropherogram alignment to peak matching and identification is included in these programs. Baseline correction is achieved by interpolating a new baseline according to points found using all local extremes, by applying an appropriate outliers test. Irreproducible migration times are corrected by compensating for EOF drift, measured with the aid of thermal marks. Thermal marks are small disturbances in the capillary created by punctual heating that move with the velocity of EOF. Peaks in the sample electropherogram are identified using a fuzzy matching algorithm, by comparing peaks from the sample electropherogram to peaks from a reference electropherogram.     

A simple light-emitted diode-induced fluorescence detector using optical fibers and a charged coupled device for direct and indirect capillary electrophoresis methods

We constructed a simple fluorescence detector for both direct and indirect CE methods using a blue light-emitted diode (470 nm) as excitation source, a bifurcated optical fiber as a waveguide, and a CCD camera as a detector. The connection of all the components is fairly easy even for nonexperts and the use of a CCD camera improves the applicability of this detector compared to the others using PMTs because it permits the recording of 2-D electropherograms or phosphorescence measurements. This detector provides a compact, low cost, and rapid system for the determination of native fluorescence compounds which have high quantum yields by CE with direct fluorescence detection, showing an LOD of 2.6 x 10(-6) M for fluorescein; the determination of fluorescence derivative compounds by CE with direct fluorescence detection, showing an LOD of 1.6 x 10(-7) M for FITC-labeled 1,6-diaminohexane; and nonfluorescence compounds by CE with indirect fluorescence detection with an LOD of 2.7 x 10(-6) M for gallic acid.     

Identification of herb Acanthopanax senticosus (Rupr. Et Maxim.) harms by capillary electrophoresis with electrochemical detection.

In this paper, a rapid, high efficient, sensitive and inexpensive approach based on a combination of simple ultrasonic extract and capillary electrophoresis (CE) separation with electrochemical detection (ED), is described to identify herbs by comparing their CE-ED profiles (namely, CE-ED electropherograms). The proposed method takes advantage of ultra-small sample volume, low consumption of organic solvent, simple sample pretreatment and easy cleanup procedure. It was applied to analyze the CE-ED profiles of stems of herb Acanthopanax senticosus (Rupr. Et Maxim.) Harms from different sources and different parts (roots, rhizomes, stems and leaves) of this herb. By comparing peak number, peak height and peak height ratio, we found that the CE-ED profiles showed big differences for the herbs from the different sources and the different parts of this herb. In addition, the distribution of bioactive compounds (isofraxidin, rutin and chlorogenic acid) in the different parts of this herb and their content variations affected by the source were studied with the CE-ED method. Based on their own unique CE-ED profiles, these herbs from the different sources and the different parts of this herb could be easily distinguished. Therefore, the proposed approach could be used as a rapid, high efficient and sensitive method for the identification of herbal medicines.     

Noise normalisation in capillary electrophoresis using a diode array detector

When using capillary electrophoresis with a diode array detector, the wavelength at maximum absorbance is often chosen to quantify a given analyte. However, the background noise for every wavelength should be taken into account as it is by maximising the signal to noise ratio that the lowest limit of detection will be obtained. Here, we proposed an algorithm allowing to correct an electropherogram from its background absorption and to estimate the background noise. Applying it to all the electropherograms obtained in each wavelength channel allows obtaining the background noise as a function of the wavelength, which can be used to calculate the signal to noise ratio. This not only allows selecting the best wavelength to maximise the limit of detection of a given analyte, but also to generate a noise normalised base peak electropherogram (nn-BPE). It is shown that the noise normalised base peak electropherograms substantially improve the peaks visualisation. The algorithm is part of a graphic user interface that runs under MatLab environment; it does not require any programming knowledge and is freely available.     

Fast and simultaneous detection of heavy metals using a simple and reliable microchip-electrochemistry route: An alternative approach to food analysis

This paper reports, for the first, the fast and simultaneous detection of prominent heavy metals, including: lead, cadmium and copper using microchip CE with electrochemical detection. The direct amperometric detection mode for microchip CE was successfully applied to these heavy metal ions. The influences of separation voltage, detection potential, as well as the concentration and pH value of the running buffer on the response of the detector were carefully assayed and optimized. The results clearly show that reliable analysis for lead, cadmium, and copper by the degree of electrophoretic separation occurs in less than 3min using a MES buffer (pH 7.0, 25mM) and l-histidine, with 1.2kV separation voltage and -0.8V detection potential. The detection limits for Pb(2+), Cd(2+), and Cu(2+) were 1.74, 0.73 and 0.13microM (S/N=3). The %R.S.D. of each peak current was <6% and migration times <2% for prolonged operation. To demonstrate the potential and future role of microchip CE, analytical possibilities and a new route in the raw sample analysis were presented. The results obtained allow the proposed microchip CE-ED acts as an alternative approach for metal analysis in foods.