Journal of Chemometrics最新文献

The validation principles on Quantitative Structure Activity Relationship issued by Organization for Economic and Co-operation and Development describe three criteria of model assessment: goodness of fit, robustness and prediction. In the case of robustness, two ways are possible as internal validation: bootstrap and cross-validation. We compared these validation metrics by checking their sample size dependence, rank correlations to other metrics and uncertainty. We used modeling methods from multivariate linear regression to artificial neural network on 14 open access datasets. We found that the metrics provide similar sample size dependence and correlation to other validation parameters. The individual uncertainty originating from the calculation recipes of the metrics is much smaller for both ways than the part caused by the selection of the training set or the training/test split. We concluded that the metrics of the two techniques are interchangeable, but the interpretation of cross-validation parameters is easier according to their similar range to goodness-of-fit and prediction metrics. Furthermore, the variance originating from the random elements of the calculation of cross-validation metrics is slightly smaller than those of bootstrap ones, if equal calculation load is applied.

When using near-infrared (NIR) techniques for analysis, model construction and maintenance updates are essential. When model construction is performed in machine learning, the sample set is usually divided into the calibration set and the validation set. The representativeness of the calibration set and the reasonable distribution of the validation set affects the accuracy of the established model. In addition, when maintaining and updating models, selecting the most informative updated sample not only improves the model prediction accuracy but also reduces sample preparation. In this paper, the spectral information entropy (SIE) is proposed to be used as a similarity criterion for dividing the sample set and use this criterion to select updated samples. The Kennard–Stone (KS) and the sample set portioning based on joint x–y distance (SPXY) methods were used for comparison to verify the superiority of the proposed method. The results showed that the model built after dividing the sample set using the SIE method has good prediction effect compared with KS and SPXY method. When predicting soluble solid content (SSC) and hardness, the prediction determination coefficient ($�R_P^2$) was improved by more than 15%, and the root mean square error (RMSE) of prediction was reduced by 50%. In terms of model updating, selecting a small number of updated samples using the SIE method can improve the correlation coefficient ($��R_P^{}�$) to more than 80%, and updated models' prediction accuracy is higher than that of KS and SPXY method. It is confirmed that the SIE method can make the NIR analysis technique more reliable.

Higher order data are commonly encountered in the domain of chemometrics, often generated by advanced analytical instruments. Due to the multilinear nature of the data, higher order data require different regression approaches compared with traditional two-way data for predictive modelling. The main aim is usually to extract the rich multilinear information, which is often lost if the data are simply analysed in unfolded form. A common algorithm for multilinear predictive modelling is N-way partial least squares (NPLS). However, a limitation of NPLS is that it inherently does not handle outlying observations, which can be detrimental to the model. This work presents a new robust multilinear predictive modelling approach based on iterative down-weighting of the outlier observations in both predictor and response space. A key benefit of the method is that it only requires a single extra parameter to tune. In this work, the algorithm is described, and the method is demonstrated on three real multilinear data sets. In all cases, the presented method outperformed the traditional NPLS modelling regarding the root mean squared error of prediction.

AlSawy NS, ElKady EF, Mostafa EA. A novel eco-friendly methods for simultaneous determination of aspirin, clopidogrel, and atorvastatin or rosuvastatin in their fixed-dose combination using chemometric techniques and artificial neural networks. Journal of Chemometrics. 2023;37(5):e3474. doi:10.1002/cem.3474

The first letter ‘A’ in the article title was mistakenly added. The updated article title is below.

“Novel eco-friendly methods for simultaneous determination of aspirin, clopidogrel, and atorvastatin or rosuvastatin in their fixed-dose combination using chemometric techniques and artificial neural networks”

We apologize for this error.

The state of the art related to parameter correlation in two-parameter models has been reviewed in this paper. The apparent contradictions between the different authors regarding the ability of D-optimality to simultaneously reduce the correlation and the area of the confidence ellipse in two-parameter models were analyzed. Two main approaches were found: (1) those who consider that the optimality criteria simultaneously control the precision and correlation of the parameter estimators and (2) those that consider a combination of criteria to achieve the same objective. An analytical criterion combining in its structure both the optimality of the precision of the estimators of the parameters and the reduction of the correlation between their estimators is provided. The criterion was tested both in a simple linear regression model, considering all possible design spaces, and in a nonlinear model with strong correlation of the estimators of the parameters (Michaelis–Menten) to show its performance. This criterion showed a superior behavior to all the strategies and criteria to control at the same time the precision and the correlation.

In this research, response surface methodology was employed in order to find the most efficient conditions to produce peanut extracts with the highest antioxidative potential (by DPPH, ABTS, FRAP, and ORAC tests) from ultrasound-assisted (UAE) and microwave-assisted (MAE) extractions. The optimal conditions for UAE were 75 ml of 30%v/v ethanol as extraction solvent, extraction temperature of 65°C, and 180 min extraction time. Optimal conditions for MAE were 15 ml of 30%v/v ethanol as extraction solvent, extraction temperature of 80°C, and 90 s extraction time. The generated models presented a reliable antioxidative response by showing non-difference between experimental and predicted values within a 95% confidence level. Characteristics such as antioxidant activities, TPC, resveratrol, and TFC were used to compare the effectiveness of the obtained extracts and the SLE extract from our previous work. The UAE extract displayed higher extraction efficiency than that of MAE in terms of higher levels of all responses (p < 0.05). Slightly lower antioxidant values of MAE might be due to the limited solvent volume used, which was five times less than that of UAE. Moreover, the high efficacy of MAE was confirmed by its extraction time, which was 120 times shorter than that of UAE. Ultimately, the correlation analysis implied that the antioxidant activity of peanut extracts was contributed by TPC rather than resveratrol or TFC, and the revealed TPC in this study was higher than previous reports.

With a goal of timely and adaptively exploiting the inconsistency inherited in the monitored samples of current interest, a novel dynamic process monitoring method based on just-in-time latent autoregressive residual generation (JITLAR²G) model is proposed. Different from the mainstream dynamic modeling and monitoring methods which usually train a signature generating mechanism and then repeatedly apply it for online monitored samples, the proposed JITLAR²G-based approach provides a JITLAR²G model for the online monitored samples after data augmentation, so that the corresponding inconsistency within the given consecutive samples could be timely and adaptively uncovered. Instead of expressing the time-serial relationship that generally accepted by the normal samples in the given dataset, solving the objective function designed for JITLAR²G in a just-in-time manner can adaptively and correspondingly seek but only one projecting vector as well as coefficient vector to generate residual, which points to the potential inconsistency inherited in the monitored samples, for the sole purpose of fault detection. As demonstrated through comparisons, the proposed JITLAR²G model can consistently guarantee its effectiveness, in terms of reducing both false alarm rate and missed alarm rate, for dynamic process monitoring, the salient performance achieved by the proposed JITLAR²G-based method in contrast to the counterparts can be always confirmed.

The present bottleneck in biosimilar bioprocess development has become evaluation of analytical results, due to recent advances in analytics, such as automated sample preparation and development of high-throughput methods. Currently automated chromatogram integration and annotation is only efficient for simple chromatograms. In an ever more competitive field of biosimilars, this represents a serious drawback because chromatographic analytical methods that provide some of the most valuable physicochemical quality attributes of the product also require careful chromatogram integration and annotation. This work focuses on the glycan mapping analytical method as utilized in development of monoclonal antibody biosimilars, evaluating more than 2000 chromatograms spanning the life cycle of multiple biosimilar development projects. It proposes a modified workflow by implementing automatic machine learning algorithms to determine the proportion of specific relevant glycan species in a sample directly from the chromatogram. Data preparation and analysis is performed using a pipeline approach. Pipeline is a modular design of data processing where signal “travels” through various active modules in a series. Each module performs a specific function or transformation on the signal and propagates the transformed signal to the next module. The pipeline is designed in a way that modules can be independently improved and exchanged. Module functions currently implemented are chromatogram resampling by spline interpolation, baseline removal by asymmetric least squares, peak alignment using parametric time warping, and quantification of the relative proportion of a glycan species using partial least squares regression. Hyper-parameters of the pipeline are then optimized using the Nelder–Mead method. The approach stands out for its ability to accommodate a broad landscape of samples, covering multiple different proteins in different stages of biosimilar development, analyzed using different adaptations of the glycan map analytical method. The pipeline presents an intuitive, flexible, and creatively simple method design capable of providing reliable results for a wide range of glycan species essential for biosimilar development. It enables transparent, faster, and less subjective evaluation of analytic raw data (from sample to result). Furthermore, our automated approach maintained an accuracy comparable with manual integration thus demonstrating its readiness for implementation in the conservative and highly regulated environment. The presented methodology reduces the cost and time of biosimilar development and should be applicable for any chromatogram-based analytical method.

In this study, we developed a simple technique for effective parameter estimation and prediction of the quantitative structure activity relationship studies using a two-step procedure. The first step is to choose the important molecular descriptors using the random forest regression, and the second step is to optimally predict the biological activity of the selected chemical compounds using the following estimators: ridge regression, jackknife ridge, Liu regression, jackknife Liu, Kibria–Lukman, and jackknife Kibria–Lukman. We conducted a simulation study and a real-life analysis with a quantitative structure–activity relationship (QSAR) data with 2540 descriptors after preprocessing. The optimal prediction is determined using the cross-validation error. The estimator with minimum cross-validation error is considered best. It is obvious that performing jackknife estimation after random forest selection is preferred.

This work endeavored to develop a high performance liquid chromatography (HPLC) method for simultaneous quantification of three important biotherapeutic molecules namely, L-tryptophan, serotonin, and melatonin, present in low amounts in agro-commodities. In the first approach, using pure chemical standards of the same in a mixture, chromatogram separation parameters such as peak sharpness, peak shouldering, and peak separation were judged by a human panel and biasness-cum-decision uncertainties were averted using fuzzy logic analysis. In the second approach, the separation parameters such as peak resolution and separation factors were evaluated to obtain a well-resolved chromatogram. The parameters of the said separation process were successfully modeled by dimensionless numbers of mass transfer of the analytes in the column; and post-fitting of mass transfer equations, high R² was obtained suggesting successful development of the chromatographic process. Euclidean distances between the values of each chromatographic separation parameter and their respective ideal values; the defuzzified scores; findings of mass transfer study; and separation factors concomitantly established that the mobile phase of composition 1% acetic acid in HPLC water (A)–HPLC grade methanol (B) in gradient elution conditions (90% A–10% B) at a flow rate of 1 mL/min could simultaneously quantify the three molecules (μg) in button mushrooms with good resolution. The presence of the said biomolecules in the extract of button mushrooms was also confirmed by electrospray ionization (ESI)-mass spectrometer (MS). An artificial neural network model (high R²) was also developed for chromatographic users, which would allow accurate prediction of the chromatographic parameters by varying mobile phase composition and flow rates.