Low-rank latent matrix-factor prediction modeling for generalized high-dimensional matrix-variate regression.

Motivated by diagnosing the COVID-19 disease using two-dimensional (2D) image biomarkers from computed tomography (CT) scans, we propose a novel latent matrix-factor regression model to predict responses that may come from an exponential distribution family, where covariates include high-dimensional matrix-variate biomarkers. A latent generalized matrix regression (LaGMaR) is formulated, where the latent predictor is a low-dimensional matrix factor score extracted from the low-rank signal of the matrix variate through a cutting-edge matrix factor model. Unlike the general spirit of penalizing vectorization plus the necessity of tuning parameters in the literature, instead, our prediction modeling in LaGMaR conducts dimension reduction that respects the geometric characteristic of intrinsic 2D structure of the matrix covariate and thus avoids iteration. This greatly relieves the computation burden, and meanwhile maintains structural information so that the latent matrix factor feature can perfectly replace the intractable matrix-variate owing to high-dimensionality. The estimation procedure of LaGMaR is subtly derived by transforming the bilinear form matrix factor model onto a high-dimensional vector factor model, so that the method of principle components can be applied. We establish bilinear-form consistency of the estimated matrix coefficient of the latent predictor and consistency of prediction. The proposed approach can be implemented conveniently. Through simulation experiments, the prediction capability of LaGMaR is shown to outperform some existing penalized methods under diverse scenarios of generalized matrix regressions. Through the application to a real COVID-19 dataset, the proposed approach is shown to predict efficiently the COVID-19.

Spatio-temporal variation and source changes of potentially toxic elements in soil on a typical plain of the Yangtze River Delta, China (2002-2012).

The spatio-temporal variation and temporal changes in the sources of Cr, Pb, Cd, Hg, and As in soil on the Hangzhou-Jiaxing-Huzhou (H-J-H) Plain were analysed based on 4,359 soil samples collected in 2002 and 2012. Geostatistical and spatial analysis methods were used to explore the spatio-temporal variation in the pollution levels and 'pollution hotspots' for potentially toxic elements (PTEs), and the positive matrix factor model was used to quantitatively appoint and analyse temporal changes in PTE sources. The results indicated that the PTE content in most parts of the survey area were at a safe level in both 2002 and 2012, but a clearly upward trend was detected for Cr, Pb, and Cd. Moreover the pollution index for Cr, Pb, Cd, and the Nemerow composite pollution index increased in the west but decreased in the east of the H-J-H Plain from 2002 to 2012. The pollution index for Hg and As presented the opposite spatial pattern. It is obvious that there have been changes in the spatial pattern of pollution hotspots for PTEs on the H-J-H Plain from 2002 to 2012. Four sources of PTEs in soil were quantitatively appointed. In 2002, 2012, the dominant sources of Cr, Cd, Hg, and As were soil parent materials, industrial activities, atmospheric deposition and agricultural inputs, respectively. The dominant source of Pb in the soil changed from traffic emissions to soil parent materials, indicating the benefit of banning the use of leaded gasoline in China. This study highlights the importance of monitoring soil environmental quality and highlights the significance of spatio-temporal variation in PTEs in suburban zones or transitional areas undergoing rapid industrialization and urbanization, like the H-J-H Plain.

Fast DLLME-GC-MS Method for Determination of Pesticides in Carmelite Drops and Evaluation of Matrix Effects in Related Medicinal Products.

The production of nutraceuticals is a growing trend, as many consumers consider them an important part of the modern active lifestyle. Others rely on the use of nutraceuticals instead of prescribing pharmaceuticals to improve their health more naturally. One of the major concerns in the nutraceutical industry is the potential presence of contaminants. Even low concentrations of contaminant residues can be harmful, so analytical methods that are sensitive at ultratrace levels are needed. Dispersive liquid-liquid microextraction method combined with fast gas chromatography and mass spectrometry was developed for the inspection of pesticide residues in Carmelite drops. The most suitable recoveries are presented when the alcohol content is fixed at 20% in Carmelite drops. The method was validated; the linearity, limits of detection/quantification, the method accuracy and precision were obtained. The complex nutraceutical matrix causes significant complications in quantitative analysis; therefore, the main target of the work was placed on studying the effects of the matrix on the correct expression of the resulting concentration of contaminants in different types of samples. An in-depth study of matrix factors was carried out, and its relationship with the content of potential interferents from the medicinal products as well as other components added during the drops' production was discussed. Related medicinal plant-derived nutraceuticals were tested, the method was applied for real-life samples, and positive findings are herein reported.

How does the order of sample analysis influence the matrix effect during LC-MS bioanalysis?

Mass spectrometry coupled with liquid chromatography is a valuable tool for drug development and personalised drug therapy. The matrix effect is caused by enhancing or suppressing the analyte signal intensity by the interfering compounds of biological fluids. The matrix effect may influence the reliability of the quantitative results. Thus, its evaluation is a critical part of bioanalytical method validation. Identified factors affecting the matrix effect are the physicochemical properties of the analyte, type of biological material, analytical conditions, the ion source construction and calculation method. The order of analysis of test samples (pure solutions and post-extraction spiked samples) is another factor possibly affecting quantifying the matrix effect variability between sources. Our primary goal was to find which experimental design - interleaved or set of blocks - is more sensitive to detect matrix effect variability. Additionally, to better understand the reason of variability, we evaluated the influence of chromatographic elution and the type of plasma (normal, lipemic or hemolyzed), co-elution, and carry-over of phospholipids. We used chemometric methods: Principal Component Analysis and Partial Least-Squares Discriminant Analysis. Although a comparable (but statistically different) matrix effect (%RSDMF) is observed using the interleaved and block schemes, for some compounds, the order of the samples strongly influences the results. The interleaved scheme was generally more sensitive in detecting the matrix effect than the block scheme. Thus, reporting the order of samples is needed to ensure the repeatability of experiments. Chemometrics suggests that lipemic samples analyzed in isocratic conditions are most prone to the matrix effect. Different compositions of matrix lots of the same type - especially lipemic - may influence method reliability. Thus, evaluating more than one source of lipemic and hemolyzed plasma is recommended.

Source Apportionment of Heavy Metals Based on Multiple Approaches for a Proposed Subway Line in the Southeast Industrial District of Beijing, China.

Apportioning the sources of heavy metals (HMs) in soil is of great importance for pollution control. A total of 64 soil samples from 13 sample points at depths of 0-21 m were collected along a proposed subway line in the southeast industrial district of Beijing. The concentrations, distribution characteristics, and sources of eight HMs were investigated. The results showed that the concentrations of Hg, Cd, Cu, Pb, As, and Zn in the topsoil (0-2 m) exceeded the Beijing soil background values. Three sources were identified and their respective contribution rates calculated for each of the HMs using multiple approaches, including correlation analysis (CA), top enrichment factor (TEF), principal component analysis (PCA), and positive matrix factor (PMF) methods. As (63.11%), Cr (61.67%), and Ni (70.80%) mainly originated from natural sources; Hg (97.0%) was dominated by fossil fuel combustion and atmospheric deposition sources; and Zn (72.80%), Pb (69.75%), Cu (65.36%) and Cd (53.08%) were related to traffic sources. Multiple approaches were demonstrated to be effective for HM source apportionment in soil, whilst the results using PMF were clearer and more complete. This work could provide evidence for the selection of reasonable methods to deal with soils excavated during subway construction, avoiding the over-remediation of the soils with heavy metals coming from natural sources.

Pollution and ecological risk assessment, and source identification of heavy metals in sediment from the Beibu Gulf, South China Sea.

The Beibu Gulf is an important passageway between China and the Association of Southeast Asian Nations, where there has been an increase in pollution of heavy metals (HMs). High concentrations of Pb, Cr, Cd, Cu, Zn, As, and Hg in surface sediment were found in Qinzhou Bay, Fangchenggang, and other coastal areas. Stochastic geo-accumulation analyses identified the pollution to be "uncontaminated"; however, it had an 18% probability of deterioration. The Cd, Hg, and As pollution were relatively serious. Principal component analysis, positive matrix factor model, and mercury isotopes demonstrated that the HMs could mainly be attributed to industrial sources including petrochemical, coal-fired, metal and metalloid processing, leather tanning, and human activities: anthropogenic sources accounted for approximately 70% of all the contaminations. This study demonstrates the contribution of terrigenous input to HMs even at a low level and provides basic data for the coordinated development of land and marine resources.

Synchronous degradation of aqueous benzotriazole and bromate reduction in catalytic ozonation: Effect of matrix factor, degradation mechanism and application strategy in water treatment.

Ozone-based technologies are used for micro-pollutants removal in wastewater treatment. However, the generation of the toxic by-product bromate (BrO3-) is of a great concern. LaCoO3 (LCO) catalytic ozonation has been used to overcome this significant drawback in the sole ozonation, achieving better BrO3- elimination efficiency. However, a key challenge is how to enhance micro-pollutant (benzotriazole, BZA) degradation efficiency and to eliminate formed BrO3- synchronously under various water qualities in drinking water or wastewater treatment. Therefore, the objective of this study is to propose a practical strategy of BZA removal and BrO3- reduction synchronously in water or wastewater treatment. In this study, important factors influencing BZA removal and BrO3- reduction were investigated, including [catalyst], [BZA], initial pH solution, [NH4+-N] and [(bi)carbonate alkalinity]. Based on the performance and mechanism of these effects, a practical strategy for BZA degradation and BrO3- elimination with and without Br- in the influent was developed. Additionally, the density functional theory (DFT) calculation successfully predicted the attack site on BZA by molecular ozone and formed hydroxyl radical (HO·) during LCO catalytic ozonation. Fukui indexes of f+ and f0 were calculated to forecast direct ozone molecule and HO· attack, respectively. Combination of DFT calculation with intermediates that identified through liquid chromatography-quadrupole time-of-flight mass spectrometry (LC-Q-TOF-MS), BZA degradation pathway was established more accurately. Additionally, four new intermediates were identified in this study. Overall, this study proposes a useful strategy for synchronous micro-pollutants degradation and BrO3- elimination, while also suggesting the feasibility of LCO catalytic ozonation for water or wastewater purification.

[Characteristics and Sources of 12 Trace Amount Elements in PM2.5 During a Period of Heavy Pollution in Huanggang, Central China].

This study provides data and theoretical support for the prevention and control of air pollution in central China. From January 13, 2018, to January 25, 2018, 48 PM2.5 samples were collected in Huanggang by particulate matter samplers during a severe pollution episode. Twelve elements (Li, Be, V, Se, Sr, Mo, Ag, Ba, Tl, Th, Bi, and U) were analyzed by ICP-MS. A positive definite matrix factor (PMF), trajectory model of the cluster analysis, potential source analysis (PSCF), and the concentration weight analysis method (CWT) were applied to analyze the potential pollution sources and determine the transmission channel. The results showed that the internal cause of the pollution was the appearance of static, stable, and high humidity meteorological conditions, while the external cause was the input of pollution. Five types of pollution source were determined:fuel combustion (10.59%), crustal (24.22%), industry (3.16%), coal (47.57%), and traffic (14.45%). Two main types of air flow path were found; short distance transmission accounted for 62.50% whereas long distance transport accounted for 37.50%. The major contributors to the pollution included central and eastern Hubei, northeastern Hunan, southwestern Anhui, and southern Henan. North-south transmission channels were observed for central China. In addition to local pollution, the impact of regional transmission should not be ignored. In the emergency response period for heavy pollution during the autumn and winter, all regions need to control emission reduction measures. Joint prevention and control are the keys to air pollution control.

[PM2.5 Pollution Characteristics During Winter and Summer in the Hohhot-Baotou-Ordos Region, China].

Based on the source apportionment by positive matrix factorization (PMF) model, we analyze the main sources and characteristics of aerosol fine particulate matter (PM2.5) during winter and summer in the Hohhot-Baotou-Ordos region, China. We found that organic (19.9%-44.6%) and crustal compositions (9.7%-46.2%) accounted for a large proportion of aerosol PM2.5 according to the results of mass closure. The results of source apportionment showed that the contribution of sources rank as:secondary inorganic aerosol (26.7%) > coal (26.1%) > motor vehicle (19.1%) > dust (18.1%) during winter, and as:secondary inorganic aerosol (26.7%) > dust (22.3%) > coal (16.6%) > vehicle exhaust (15.1%) > SOC (8.7%) during summer. Findings suggest that the contribution of sources with secondary inorganic aerosol were the largest sources both in winter and summer, and that the Hohhot-Baotou-Ordos region was also affected by coal during the winter and dust during the summer. Corresponding to the source apportionment, analysis of typical heavy pollution episodes in winter and summer showed that the pollution sources during the winter were mainly secondary inorganic aerosol and coal, whereas they were mainly secondary inorganic aerosol during the summer.

A sensitive method for the simultaneous UHPLC-MS/MS analysis of milrinone and dobutamine in blood plasma using NH4F as the eluent additive and ascorbic acid as a stabilizer.

The purpose of this work was to develop and validate an HPLC-MS/MS method suitable for quantifying two important cardiovascular drugs, milrinone and dobutamine, in neonatal and paediatric patients' blood plasma samples. Sufficiently low LLOQ levels were required to obtain adequate pharmacokinetic data for the evaluation of optimal dosing. Since the specifics of the patient group set some restrictions on the available sample volume, the method was designed to use only 20 µL of plasma for the analysis. Analytes were separated chromatographically in a biphenyl column using a conventional water-methanol-formic acid eluent with the addition of ammonium fluoride. The latter provided a significant signal enhancement in positive ion mode detection for both analytes allowing the LLOQ to reach below 1 ng/mL. Matrix matched calibration was linear in the range of 1-300 ng/mL, between-run accuracy remained within 107-115%, and precision within 4.8-7.4% for both analytes over the calibration range (including LLOQ level). Dobutamine degradation in plasma samples was prevented by the usage of ascorbic acid. The method was applied to plasma samples of neonates from two pharmacokinetic/pharmacodynamics studies (n = 38).

Quantitative evaluation of the matrix effect in bioanalytical methods based on LC-MS: A comparison of two approaches.

Liquid chromatography coupled to mass spectrometry (LC-MS) is a powerful tool for studying pharmacokinetics and toxicokinetics. Reliable bioanalysis requires the characterization of the matrix effect, i.e. influence of the endogenous or exogenous compounds on the analyte signal intensity. We have compared two methods for the quantitation of matrix effect. The CVs(%) of internal standard normalized matrix factors recommended by the European Medicines Agency were evaluated against internal standard normalized relative matrix effects derived from Matuszewski et al. (2003). Both methods use post-extraction spiked samples, but matrix factors require also neat solutions. We have tested both approaches using analytes of diverse chemical structures. The study did not reveal relevant differences in the results obtained with both calculation methods. After normalization with the internal standard, the CV(%) of the matrix factor was on average 0.5% higher than the corresponding relative matrix effect. The method adopted by the European Medicines Agency seems to be slightly more conservative in the analyzed datasets. Nine analytes of different structures enabled a general overview of the problem, still, further studies are encouraged to confirm our observations.

Volumetric absorptive microsampling combined with impact-assisted extraction for hematocrit effect free assays.

AIM: Volumetric absorptive microsampling (VAMS) is a recent technology available for sampling and analyzing low blood volume. The present work describes the utilization of VAMS for the quantitation of naproxen and ritonavir in human blood using a novel bead-based impact-assisted extraction (IAE) procedure. RESULTS: Sampling volume accuracy of the VAMS device was independent of the blood hematocrit (HCT) level, however analyte recovery decreased with increasing HCT when extracted using ultrasonication. In contrast, IAE was unaffected by HCT, resulting in quantitative recovery for all levels evaluated. Precision and accuracy batches, as well as matrix effect evaluation, met acceptance criteria. CONCLUSION: The IAE procedure coupled with VAMS is immune to HCT biases affecting sampling volume and recovery.

Determination of tulobuterol in rat plasma using a liquid chromatography-tandem mass spectrometry method and its application to a pharmacokinetic study of tulobuterol patch.

A sensitive and accurate liquid chromatography-tandem mass spectrometry (LC-MS/MS) method has been developed and validated for determination of tulobuterol in rat plasma for the first time. Plasma samples were extracted by liquid-liquid extraction method with methyl tert-butyl ether and the analyte and clenbuterol (IS) were separated on a Venusil MP C18 column (100mm×2.1mm, 3µm) using 0.1% formic acid-water-methanol as mobile phase, with a runtime of 5min. The analyte was detected in multiple reaction monitoring (MRM) mode with positive electrospray ionization. Transitions of m/z 228.2→154.0 for tulobuterol and m/z 277.1→203.0 for the clenbuterol were monitored. The linear range was 0.5-100ng/ml (r=0.9967) for tulobuterol with the lower limit of quantitation of 0.5ng/ml. The intra-day and inter-day precisions were less than 10.3% for the analyte and the accuracy was less than -8.6%. The RSD of matrix effect and recovery yield were within ±15% of nominal concentrations and tulobuterol was stable during stability studies. The validated method has been successfully applied to a pharmacokinetic study of three doses of tulobuterol patch in rats for the first time.

Validation data for the quantification of the Annonaceous acetogenin annonacin in Rat brain by UPLC-MS/MS.

Annonaceous acetogenins (AAGs) are environmental neurotoxins from the fruit pulp of several Annonaceae species, whose consumption was linked to the occurrence of sporadic atypical Parkinsonism with dementia. The quantification of the prototypical AAG annonacin in Rat brain homogenates was performed by UPLC-MS/MS in selected reaction monitoring (SRM) mode, using a triple quadrupole mass analyzer. A natural analog of annonacin was used as an internal standard. Analyzed data set of the analytical validation of this method is presented, including stability of the samples, extraction recovery and matrix effect, supporting the results described in the article "Quantification of the environmental neurotoxin annonacin in Rat brain by UPLC-MS/MS" N. Bonneau, I. Schmitz-Afonso, D. Touboul, A. Brunelle, P. Champy (2016) [1].

Development and validation of two liquid chromatography-tandem mass spectrometry methods for the determination of silibinin and silibinin hemisuccinate in human plasma.

To investigate the pharmacokinetics of silibinin and silibinin hemisuccinate in human plasma, two high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) methods were developed and validated. The methods require a small volume of sample (100µL), and the recovery of the analytes was complete with a good reproducibility (CV% 1.7-9.5), after a simple protein precipitation. Naringenin was used as internal standard. The chromatographic methods provided a good separation of diastereoisomers A and B of both silibinin and silibinin hemisuccinate onto a Chromolith Performance RP18e 100mm×3mm column, with a resolution of peaks from plasma matrix in less than 6min. The methods precision values expressed as CV% were always ≤6.2% and the accuracy was always well within the acceptable 15% range. Quantification was performed on a triple-quadrupole tandem mass spectrometer by Selected Reaction Monitoring (SRM) mode, in a negative ion mode, via electrospray ionization (ESI). The lower limit of quantitation was set at 5.0ng/mL (silibinin) and 25.0ng/mL (silibinin hemisuccinate), and the linearity was validated up to 1000.0 and 12,500.0ng/mL, for silibinin and silibinin hemisuccinate, respectively, with correlation coefficients (R(2)) of 0.991 or better. The methods were suitable for pharmacokinetic studies and were successfully applied to human plasma samples from subjects treated intravenously with Legalon(®) SIL at the dose of 20mg/kg, expressed as silibinin.