Optimal estimation of intensity of noisy peaks by matched filtering with application to chromatography : Part 2. Program description and simulation experiments

The (digital) matched filter described in Part 1 is applied to chromatographic data files. Practical aspects concerning the use of this filter are discussed. A FORTRAN program for interactive application of the matched filter is described. This program was used for simulation experiments in order to assess the model errors. Results indicate only a weak dependence on the signal-to-noise ratio. Selection of a 1/f noise model is recommended when the type of the signal noise is unknown. The relation between the Gamma and Guassian functions used is shown.     

Improved method for the determination of kynurenic acid in rat plasma by column-switching HPLC with post-column fluorescence detection

Kynurenic acid (KYNA), an endogenous antagonist of ionotropic glutamate and α7 nicotinic receptors, was fluorometrically determined by column-switching high-performance liquid chromatography (HPLC) with fluorescence detection. The HPLC system consists of two octadecyl silica (ODS) columns, both of which are connected with an anion-exchange column (trapping column). Following sample injection onto the HPLC column, KYNA was separated on the first ODS column with a mobile phase of H₂O/acetonitrile (95/5) containing 0.1% acetic acid. The peak fraction of KYNA was trapped on the anion-exchange column by changing the position of a six-port valve and then introduced into the second ODS column. Subsequently, KYNA was detected fluorometrically as a fluorescence complex formed with zinc ion which was pumped constantly. Instrumental limit of detection was approximately 0.16 nM, which corresponded to 8.0 fmol (per 50 μl injection, signal to noise ratio 3), and the limit of quantification was 0.53 nM (signal to noise ratio 10). Intra- and inter-day relative standard deviations were 1.1-3.9% (n = 3) and 3.0-5.3% (n = 3), respectively. The peak of KYNA in rat plasma was clearly detected by the proposed column-switching HPLC system after a facile pretreatment procedure. Intra- and inter-day relative mean errors were −1.6-1.4% (n = 3) and −2.4 to −0.4% (n = 3), respectively, with a satisfactory precision (within 5.0%). A calibration curve for the determination of KYNA showed a good linearity (r² > 0.999) in the range of 25-200 nM. The KYNA concentrations in the plasma of male Sprague-Dawley rats (8-week-old) were 44 ± 5.5 nM (mean ± S.E., n = 5). In ketamine-treated rats, which are animal models of schizophrenia, the plasma KYNA concentrations were significantly increased compared with those in the control rats (p < 0.05).     

Two-dimensional liquid chromatography coupled with mass spectrometry for the analysis of Lobelia chinensis Lour. using an ESI/APCI multimode ion source

A comprehensive approach for the separation and identification of components in a traditional Chinese medicine Lobelia chinensis Lour. was developed using 2D-HPLC coupled with an online photodiode array (PDA) detector and a mass spectrometer. The extract of L. chinensis Lour. was separated on a CN column in the first-dimensional HPLC, and then each of the collected fractions was further separated on an ODS column followed by an online PDA detector. After separation in the two different chromatographic modes, the eluents were delivered to a quadrupole mass spectrometer equipped with a multimode ion source of an ESI and an atmospheric pressure chemical ionization (ESI/APCI). At least 536 components in L. chinensis Lour. extract were detected and 6 of them were identified as apigenin 7-O-rutinoside, luteolin, lobetyolinin, lobetyolin, diosmin, and linarin, respectively, according to their UV spectrum and mass spectrum. The results demonstrated the powerful resolution, high peak capacity, as well as the identification capability of the 2D-HPLC combined with PDA and ESI/APCI-MS for the analyses of complex samples.     

Application of static signal-to-noise theory to the detection and integration of dynamic signals

Static, time-independent noise terms are combined to estimate the noise and signal-to-noise ratio (S/N) of dynamic, time-dependent methods. Simplified S/N equations for shot-noise-limited and flicker-noise-limited analyte signals with and without background signal noises are given. The intuitive aspects of peak integration are justified by these equations. For an ideal Gaussian peak, the equations are used to derive the chromatographic limit of detection (LOD). A comparison of the chromatographic detection limit with the detection limit for a static system using the same amount of sample predicts poorer chromatographic detection limits because of peak broadening. A possible determinate error in area measurement is shown to result when integration limits are chosen on the basis of the static detection limit.     

Computation of distribution of minimum resolution for log-normal distribution of chromatographic peak heights

General equations are derived for the distribution of minimum resolution between two chromatographic peaks, when peak heights in a multi-component chromatogram follow a continuous statistical distribution. The derivation draws on published theory by relating the area under the distribution of minimum resolution to the area under the distribution of the ratio of peak heights, which in turn is derived from the peak-height distribution. Two procedures are proposed for the equations' numerical solution. The procedures are applied to the log-normal distribution, which recently was reported to describe the distribution of component concentrations in three complex natural mixtures. For published statistical parameters of these mixtures, the distribution of minimum resolution is similar to that for the commonly assumed exponential distribution of peak heights used in statistical-overlap theory. However, these two distributions of minimum resolution can differ markedly, depending on the scale parameter of the log-normal distribution. Theory for the computation of the distribution of minimum resolution is extended to other cases of interest. With the log-normal distribution of peak heights as an example, the distribution of minimum resolution is computed when small peaks are lost due to noise or detection limits, and when the height of at least one peak is less than an upper limit. The distribution of minimum resolution shifts slightly to lower resolution values in the first case and to markedly larger resolution values in the second one. The theory and numerical procedure are confirmed by Monte Carlo simulation.     

Comprehensive two-dimensional high performance liquid chromatography system with immobilized liposome chromatography column and monolithic column for separation of the traditional Chinese medicine Schisandra chinensis

A comprehensive two-dimensional (2D) separation is one that employs two separation dimensions (columns) and draws on all of the available resolving power from each of the dimensions of separate the components in a sample. In this study, a comprehensive 2D chromatography approach was developed for the separation and identification of membrane permeable compounds in a famous traditional Chinese medicine of Schisandra chinensis. The first dimensional column was the immobilized liposome chromatography (ILC) column, which mimics the biological membranes and can be used to study drug-membrane interactions in liquid chromatography. Using an automatic ten-port switching valve equipped with two sample loops, the section of the first-dimension was introduced in the second-dimension consist of a silica monolithic column. More than 40 components in Schisandra chinensis were resolved by using the developed separation system and among them 14 compounds were identified interacting with the ILC column based on their retention action, UV and mass data. With this comprehensive 2D-HPLC system, the three-dimensional chromatographic fingerprints of Schisandra chinensis were preliminarily established and processed by using principal component analysis and hierarchical clustering analysis. The obtained information can distinguish the unacceptable samples of the quality control. The result demonstrated that the 2D biochromatography system has been demonstrated to have more advantages of finding strong binding bioactive components, providing an enhanced peak capacity, good sensitivity and powerful resolution biological fingerprinting analysis of complex TCMs, which was a useful means to control the quality of and to clarify the membrane permeability of the compounds in Schisandra chinensis.     

Mobile phase selection for the combined use of liquid chromatography–inductively coupled plasma mass spectrometry and electrospray ionisation mass spectrometry

Four different organic solvents: dimethylformamide, 1,4-dioxane, n-propanol and ethanol were evaluated as alternative organic modifiers to acetonitrile for liquid chromatography (LC) separations. The aim was to establish common sets of chromatographic conditions that could be applied for LC hyphenation to inductively coupled plasma mass spectrometry (ICPMS) as well as to electrospray ionization MS (ESIMS). The approach was to evaluate candidate solvents that, compared to acetonitrile, potentially could give improved analytical performance (low solvent vapor loading, maximized analyte sensitivity and minimized carbon depositions on instrumental parts) in ICPMS analysis while retaining chromatographic and ESIMS performances. The study showed that dimethylformamide, 1,4-dioxane, n-propanol and ethanol all can be advantageous chromatographic modifiers for LC–ICPMS analysis, giving superior performance compared to acetonitrile. For the combined use of LC–ICPMS and LC–ESIMS with a common set of chromatographic conditions, n-propanol gave the best overall performance. The ¹⁹⁵Pt⁺ signal in ICPMS was continuously monitored during a 0–60% organic solvent gradient and at 25% of organic modifier, 100% of the signal obtained at the gradient start was preserved for n-propanol compared to only 35% of the signal when using acetonitrile. Platinum detection limits were 5–8 times lower using n-propanol compared with acetonitrile. Signal-to-noise ratio in continuous ESIMS signal measurements was 100, 90 and 110 for a 100 μg/ml solution of leucine–enkephaline using acetonitrile, ethanol and n-propanol, respectively. Chromatographic efficiency in reversed phase separations was preserved for n-propanol compared to acetonitrile for the analysis of the whole protein cytochrome C and the peptide bacitracin on a column with particle and pore sizes of 5 μm and 300 Å, but slightly deteriorated for the separation of the peptides leucine–enkephaline and bacitracin on a 3 μm and 90 Å column as the peak width at half height for both peptides increased by a factor of two. The performance on the smaller dimensioned column could however be improved by running the separations at 40 °C.     

Probabilistic peak detection for first-order chromatographic data

We present a novel algorithm for probabilistic peak detection in first-order chromatographic data. Unlike conventional methods that deliver a binary answer pertaining to the expected presence or absence of a chromatographic peak, our method calculates the probability of a point being affected by such a peak. The algorithm makes use of chromatographic information (i.e. the expected width of a single peak and the standard deviation of baseline noise). As prior information of the existence of a peak in a chromatographic run, we make use of the statistical overlap theory. We formulate an exhaustive set of mutually exclusive hypotheses concerning presence or absence of different peak configurations. These models are evaluated by fitting a segment of chromatographic data by least-squares. The evaluation of these competing hypotheses can be performed as a Bayesian inferential task. We outline the potential advantages of adopting this approach for peak detection and provide several examples of both improved performance and increased flexibility afforded by our approach.     

Linear Modulated Stochastic Resonance Algorithm Applied to Determination of Thidiazuron Residue in Water Using Stochastic Resonance

Abstract

The newly proposed linear modulated stochastic resonance algorithm (LSRA) was used to amplify and detect the weak chromatographic peaks of thidiazuron. The output chromatographic peak is often distorted when using the traditional stochastic resonance algorithm because of the existence of strong noise. In LSRA, the distortion of the output peak can be corrected by introducing a linear force into the nonlinear system. A two‐step optimization method was proposed to give attention to both the signal‐to‐noise ratio and the peak shape of output signal. The weak chromatographic peaks of thidiazuron can be amplified significantly and the distortion of the output peaks can be corrected using LSRA. The algorithm was used to detect thidiazuron residue in water with solid phase extraction‐high performance liquid chromatography. The limit of detection and limit of quantification were improved to 2.5 ng/l and 10 ng/l, respectively.     

Stochastic resonance algorithm applied to quantitative analysis for weak liquid chromatographic signal of pyrene in water samples

The phenomenon of stochastic resonance (SR), which was discovered in recent years, rendered an entirely new way for the detection of weak signals, and it has been widely studied in many different science fields. This phenomenon is manifest in nonlinear systems whereby a weak signal can be amplified when the noise, signal and nonlinear system attain the proper cooperation. The introduced algorithm was employed to detect pyrene in drinking water samples with solid-phase extraction–liquid chromatography. The weak chromatographic peak of the analyte was amplified significantly, and the profiles of the peaks were also satisfactory. The limit of detection and the limit of quantification were improved from 0.022?ng?mL^?1 and 0.08?ng?mL^?1 to 0.004?ng?mL^?1 and 0.01?ng?mL^?1, respectively. The results showed an excellent quantitative relationship between concentrations and chromatographic responses. It is expected that the SR will be an effective tool to detect weak chromatographic peaks quantitatively in trace analysis.     

New algorithms for processing and peak detection in liquid chromatography/mass spectrometry data

Two new algorithms for automated processing of liquid chromatography/mass spectrometry (LC/MS) data are presented. These algorithms were developed from an analysis of the noise and artifact distribution in such data. The noise distribution was analyzed by preparing histograms of the signal intensity in LC/MS data. These histograms are well fit by a sum of two normal distributions in the log scale. One new algorithm, median filtering, provides increased performance compared to averaging adjacent scans in removing noise that is not normally distributed in the linear scale. Another new algorithm, vectorized peak detection, provides increased robustness with respect to variation in the noise and artifact distribution compared to methods based on determining an intensity threshold for the entire dataset. Vectorized peak detection also permits the incorporation of existing algorithms for peak detection in ion chromatograms and/or mass spectra. The application of these methods to LC/MS spectra of complex biological samples is described.     

Development of an analytical methodology to quantify melamine in milk using micellar liquid chromatography and validation according to EU Regulation 2002/654/EC

Melamine is a toxic triazine, illegally used as an additive in milk to apparently increase the amount of protein. A chromatographic procedure using a C18 column and a micellar mobile phase of sodium dodecyl sulphate (0.05 M) and propanol (7.5%), buffered at pH 3, and a detection set by absorbance at 210 nm, was reported for the resolution and quantification of melamine in liquid and powdered milk samples. In this work, samples were diluted with a SDS solution and were directly injected, thus avoiding long extraction and experimental procedures. Melamine was eluted in nearly 9.3 min without overlapping the protein band or other endogeneous compounds. The optimal mobile phase composition was taken using a chemometrical approach that considers the retention factor, efficiency and peak shape. Validation was performed following the European Commission's indications (European Decision 2002/657/EC), and the main analytical parameters studied were: linearity (0.02-100 ppm; r² = 0.999), limit of detection (5 ppb), intra- and inter-day precision (R.S.D. <7.6% and <9.7%, respectively) and robustness (R.S.D. <7.4% for retention time and <5.0% for area). Sensitivity was adequate to detect melamine under the safety limits proposed by the US FDA. Finally, recoveries for several milk samples were found in the 85-109% range.     

Trace LC/MS quantitative analysis of polypeptide biomarkers: impact of 1-D and 2-D chromatography on matrix effects and sensitivity

We investigated the impact of one dimension (single reverse phase (RP) column) and two dimension (two different RP columns) chromatographic methods on SIM (MS) and multiple reaction monitoring (MRM; MS/MS) performance from human plasma. We find that MRM analysis is clearly preferable for 1-D applications; however, implementation of SIM detection in conjunction with 2-D separation technique resulted in an over 60-fold increase in analyte peak area and improved S/N compared to MRM for our analyte, human C-peptide. Implementation of a 2-D RP-RP technique with SIM detection is capable of eliminating matrix effects and greatly increases signal response and data quality. For two large peptides in complex biological samples, we found that a 2-D approach performed better than high quality sample preparation together with 1-D chromatography and MRM, even on a high-end mass spectrometer.     

Application of principal component analysis to the extraction of pure mass spectra in chemical analysis by gas chromatography/mass spectrometry

In this work, we investigated the possibility of the purification of mass spectra of coeluted substances containing at least one characteristic value of m/z using principal component analysis. Simulated and real gas chromatographic/mass spectrometric data were analyzed. For the simulated data, the resolution of the chromatographic peaks varied from 0.075 to 0.75, and the ratio of component concentrations ranged from 0.01 to 100. Noise-free simulated data and simulated data with randomly distributed noise were considered. It was shown that the method gave satisfactory results even when the chromatographic resolution (R _s ) was less than 0.05.     

Development of a green chromatographic method for determination of colorants in food samples

A green chromatographic analytical method for determination of Tartrazine, Brilliant Blue and Sunset Yellow in food samples is proposed. The method is based on the modification of a C18 column with a 0.25% (v/v) Triton X-100 aqueous solution at pH 7 and in the usage of the same surfactant solution as mobile phase without the presence of any organic solvent modifier. After the separation process on the chromatographic column, the colorants are detected at 430, 630 and 480 nm, respectively. The chromatographic procedure yielded precise results and is able to run one sample in only 8 min, consuming 15.0 mg of Triton X-100 and 38.8 mg of phosphate. When the flow rate of the mobile phase is 1 ml min⁻¹ the retention times are 2.1, 3.6 and 7.0 min for Tartrazine, Brilliant Blue and Sunset Yellow, respectively; and all peak resolutions are ca. 2. The analytical curves present the following linear equations: area = 7.44 10⁵ + 2.71 10⁵ [Tartrazine] (R = 0.998, n = 7); area = 1.09 10⁵ + 3.75 10⁵ [Brilliant] (R = 0.9995, n = 7) and area = −7.34 10⁴ + 2.33 10⁵ [Sunset] (R = 0.998), n = 7) and, the limits of detection for Tartrazine, Brilliant Blue and Sunset Yellow were estimated as 0.125, 0.080 and 0.143 mg l⁻¹. When the proposed method is applied to food samples analysis, precise results are obtained (R.S.D. < 5%, n = 3) and in agreement with those obtained by using the classical spectrophotometric method. The traditional usage of organic solvent as mobile phase in HPLC is not used here, which permits to classify the present method as green.     

Dynamics of ions of intact proteins in the Orbitrap mass analyzer

While allowing analysis of intact proteins without a theoretical upper mass limit, the Orbitrap mass analyzer demonstrates reduced resolving power as ion mass increases even at a constant mass-to-charge ratio. It is shown that this effect comes from the effects of ion scattering on background gas molecules. The main mechanisms causing decay of acquired transient appear to be fragmentation as well as accelerated dephasing of ion packets. Isotopic resolution of proteins including bovine serum albumin (MW 66.4 kDa) and transferrin (MW 78 kDa) has also been demonstrated. As a part of this study, detection of individual multiply-charged ions of myoglobin (MW 16.9 kDa) has been demonstrated. Quantized distribution of signal intensities for myoglobin ions well above the noise threshold was observed, with high mass accuracy and resolution of recorded individual ions used as an independent confirmation of correct assignment of signal to ions rather than to noise. The latter also allowed us to benchmark the sensitivity of image-current detection and explore in detail factors responsible for signal decay.     

High performance liquid chromatography analysis of aliphatic thiols in alimentary supplements and pharmaceuticals using menadione as a new useful derivatization reagent

The use of menadione (MD) as a pre-column reagent for high performance liquid chromatography (HPLC) analysis of aliphatic thiols is proposed. The reaction was carried out for 5 min at room temperature and pH 8.5. The developed method was applied to the N-acetylcysteine (NAC) analysis of alimentary supplements and pharmaceutical formulations. The effect of the complex matrix was evaluated by the study of the thiol derivatization reaction both in standard and in placebo solutions. The yield of NAC-MD adduct was found to be quantitative at a reagent to thiol molar ratio of about 4 in comparison with an authentic specimen of synthesized NAC adduct, which was characterized by ¹H NMR, IR and UV. The routine chromatographic separations were performed on a Synergi MAX-RP column using a mobile phase consisting of methanol/triethylammonium (TEA) phosphate buffer (pH 3; 0.05 mol L⁻¹) 70:30 (v/v) at a flow-rate of 0.4 mL min⁻¹. UV-diode array detection was used setting the wavelength at λ = 260 nm. The validation parameters such as linearity, sensitivity, accuracy, precision, selectivity and ruggedness were found to be highly satisfactory. Similar linear responses were observed by standard and placebo solutions (determination coefficient: 0.9996). Limit of detection was about 0.019 μg g⁻¹. Intra-day precision (relative standard deviation, R.S.D.) was ≤0.81% for NAC to internal standard (IS) peak area ratio, ≤0.28% and ≤0.32%, respectively, for NAC and IS retention times (t_R), without significant differences between intra- and inter-day data. NAC recovery studies gave good results (100.12%) with R.S.D. = 1.05%.     

复杂色谱信号自动解析中的化学计量学方法

色谱及其联用技术日趋完善,并向自动化、高通量和快速的方向发展。化学计量学利用"数学分离"手段,可以实现色谱信号的自动化解析,已成为现代色谱分析中非常活跃的研究领域。但以往的化学计量学方法并不能完全有效地实现复杂色谱信号自动化解析。为此,自动化色谱解析算法成为科研工作者关心的重点,众多新型的自动化解析算法被提出。针对复杂一维色谱数据以及联用仪器得到的二维和更高维数据的自动化分析,化学计量学研究主要集中在自动色谱峰识别、背景以及基线漂移校正、色谱谱峰漂移校正以及重叠色谱峰的解析。该文对近十年来发展的复杂体系色谱信号自动化解析中化学计量学方法的原理与应用进行了总结与评述,比较了各类方法的优势与不足。在此基础上,针对当前色谱自动化分析过程中的难题对未来该领域的研究方向进行了展望。     

Extension of the two-dimensional mass channel cluster plot method to fast separations utilizing low thermal mass gas chromatography with time-of-flight mass spectrometry

Implementation of a data reduction and visualization method for use with high-speed gas chromatography and time-of-flight mass spectrometry (GC-TOFMS) is reported. The method, called the “2D m/z cluster method” facilitates analyte detection, deconvolution, and identification, by accurately measuring peak widths and retention times using a fast TOFMS sampling frequency (500 Hz). Characteristics and requirements for high speed GC are taken into consideration: fast separations with narrow peak widths and high peak capacity, rapid data collection rate, and effective peak deconvolution. Transitioning from standard GC (10–60+ minute separations) to fast GC (1–10 min separations) required consideration of how to properly analyze the data. This report validates use of the 2D m/z cluster method with newly developed GC technology that produces ultra-fast separations (∼1 min) with narrow analyte peak widths. Low thermal mass gas chromatography (LTM-GC) operated at a heating rate of 250 °C/min coupled to a LECO Pegasus III TOFMS analyzed a 115 component test mixture in 120 s with peak widths-at-base, w_b (4σ), of 350 ms (average) to produce a separation with a high peak capacity, n_c ∼ 340 (at unit resolution R_s = 1). The 2D m/z cluster method is shown to separate overlapped analytes to a limiting R_s ∼ 0.03, so the effective peak capacity was increased nearly 30-fold to n_c ∼10,000 in the 120 s separation. The method, when coupled with LTM-GC-TOFMS, is demonstrated to provide unambiguous peak rank (i.e. the number of analytes per overlapped peak in the total ion current (TIC)), by visualizing locations of pure and chromatographically overlapped m/z. Hence, peak deconvolution and identification using MCR-ALS (multivariate curve resolution – alternating least squares) is demonstrated.     

Simultaneous amperometric determination of lignin peroxidase and manganese peroxidase activities in compost bioremediation using artificial neural networks

High-performance liquid chromatography (HPLC) and fluorescence derivatization were applied for a nanogram-level N-nitrosodimethylamine (NDMA) analysis of water samples. For the analysis of N-nitrosodimethylamine, samples were first denitrosated by a mixed solution of hydrobromic acid and acetic acid to produce dimethylamine, which was derivatized with dansyl chloride for HPLC fluorescence detection. Fluorescence detection was optimized with excitation and emission wavelengths of 340 and 530 nm, respectively. pH adjustment after denitrosation was necessary to maximize fluorescence intensity with pHs in the range of 9-12. A dansyl chloride concentration of 500 mg l⁻¹ was found to be optimal for measuring a fluorescence signal. An instrumental detection limit of 0.1 ng of NDMA was possible with fluorescence derivatization. The NDMA in water samples was extracted by continuous solid-phase extraction using Ambersorb 572. Although the determination of NDMA was variable at lower concentrations (less than 200 ng l⁻¹), it was observed that the NDMA detection limit with this method could be lowered to a concentration of 10 ng l⁻¹. Another benefit of this method can be found in its selectivity for NDMA. Unlike gas chromatographic (GC) methods, this method generates a distinct peak for NDMA without interference even in the complex matrix of wastewater effluents. The HPLC with fluorescence derivatization method may be applicable for determining NDMA in water and wastewater samples for various research purposes and for screening environmental samples.