A rapid analytical method of human whole blood viscosity with low, medium, and high shear rates [WBV(L), WBV(M), and WBV(H), respectively] was developed using visible and near-infrared (Vis-NIR) spectroscopy combined with a moving-window partial least squares (MW-PLS) method. Two groups of peripheral blood samples were collected for modeling and validation. Separate analytical models were established for male and female groups to avoid interference in different gender groups and improve the homogeneity and prediction accuracy. Modeling was performed for multiple divisions of calibration and prediction sets to avoid over-fitting and achieve parameter stability. The joint analysis models for three indicators were selected through comprehensive evaluation of MW-PLS. The selected joint analysis models were 812–1278 nm for males and 670–1146 nm for females. The root-mean-square errors (SEP) and the correlation coefficients of prediction (RP) for all validation samples were 0.54 mPa•s and 0.91 for WBV(L), 0.25 mPa•s and 0.92 for WBV(M), and 0.22 mPa•s and 0.90 for WBV(H). Results indicated high prediction accuracy, with prediction values similar to the clinically measured values. Overall, the findings confirmed the feasibility of whole blood viscosity quantification based on Vis-NIR spectroscopy with MW-PLS. The proposed rapid and simple technique is a promising tool for surveillance, control, and treatment of cardio-cerebrovascular diseases in large populations.

•Optimal partner wavelength combination is proposed.•Successfully applied for near-infrared spectroscopic analysis•High efficient approach for extracting information wavelengthsUsing binary linear regression, the optimal partner wavelength of each wavelength is selected in an initial wavelength screening region. On the basis of strategy above, a novel approach for selecting appropriate wavelengths combination, named optimal partner wavelength combination (OPWC) coupled with partial least squares (PLS), is proposed, and was successfully applied for reagent-free near-infrared spectroscopic analysis of organic matter in soil. Moving window PLS (MW-PLS), successive projections algorithm (SPA) and Monte Carlo uninformative variable elimination (MC-UVE), which are well-performed wavelength selection methods, were also conducted for comparison.The OPWC-PLS, MW-PLS, SPA-PLS and MC-UVE-PLS methods selected 14, 210, 63, 199 wavelengths, respectively. The root-mean-square error and correlation coefficients for leave-one-out cross validation were 0.165 g kg− 1 and 0.967 for OPWC-PLS, 0.163 g kg− 1 and 0.968 for MW-PLS, 0.198 g kg− 1 and 0.953 for SPA-PLS, and 0.190 g kg− 1 and 0.956 for MC-UVE-PLS, respectively. The results indicate that OPWC-PLS and MW-PLS methods were almost the same, and were obvious better than SPA-PLS and MC-UVE-PLS methods. But the OPWC only contained 14 wavelengths, which is a high efficient approach for extracting information wavelengths and mitigating redundant wavelengths. OPWC can be also provided valuable reference for designing small dedicated spectrometers with a high signal-to-noise ratio.OPWC can be programmed determined, which has small amount of calculation and high operating speed, and it is a deterministic search technique whose results are reproducible. We believe that OPWC has such applicability and can be applied to other fields of spectroscopic analysis.

•Equidistant combination partial least squares was proposed.•Based on various divisions of calibration and prediction sets to achieve stability•Joint analyses of the clinical indicators for hyperlipidemia via NIR spectrum•High prediction accuracy and low model complexity•Provided valuable references for designing a small, dedicated spectrometerThe development of a small, dedicated near-infrared (NIR) spectrometer has promising potential applications, such as for joint analyses of total cholesterol (TC) and triglyceride (TG) in human serum for preventing and treating hyperlipidemia of a large population. The appropriate wavelength selection is a key technology for developing such a spectrometer. For this reason, a novel wavelength selection method, named the equidistant combination partial least squares (EC-PLS), was applied to the wavelength selection for the NIR analyses of TC and TG in human serum. A rigorous process based on the various divisions of calibration and prediction sets was performed to achieve modeling optimization with stability. By applying EC-PLS, a model set was developed, which consists of various models that were equivalent to the optimal model. The joint analyses model of the two indicators was further selected with only 50 wavelengths. The random validation samples excluded from the modeling process were used to validate the selected model. The root-mean-square errors, correlation coefficients and ratio of performance to deviation for the prediction were 0.197 mmol L− 1, 0.985 and 5.6 for TC, and 0.101 mmol L− 1, 0.992 and 8.0 for TG, respectively. The sensitivity and specificity for hyperlipidemia were 96.2% and 98.0%. These findings indicate high prediction accuracy and low model complexity. The proposed wavelength selection provided valuable references for the designing of a small, dedicated spectrometer for hyperlipidemia. The methodological framework and optimization algorithm are universal, such that they can be applied to other fields.

•We developed a simple method to determine glycated hemoglobin in blood.•The method was based on near-infrared spectroscopy with stable wavelength selection capability.•The sensitivity and specificity values for diabetes were above 93.5%.A novel near-infrared-spectroscopy-based quantification method for glycated hemoglobin (HbA1c), a major clinical diagnosis indicator of diabetes, was developed on the basis of simultaneous determination of hemoglobin (Hb) and absolute HbA1c content (Hb•HbA1c) in human hemolysate samples. Equidistant combination partial least squares (EC-PLS) method was proposed to perform wavelengths selection. Competitive adaptive reweighted sampling PLS (CARS-PLS) and Monte Carlo uninformative variable elimination PLS (MC-UVE-PLS) methods were also conducted for comparison. A randomness and stability dependent rigorous process of calibration, prediction, and validation was performed to produce objective and stable models. The search range covered the unsaturated region (780–1880 nm, 2090–2330 nm). For Hb and Hb•HbA1c, only 6 and 14 wavelengths were selected with EC-PLS, 23 and 30 wavelengths were selected with CARS-PLS, and 100 and 120 wavelengths were selected with MC-UVE-PLS, respectively.The predicted values of relative percentage HbA1c were calculated from the predicted Hb and Hb•HbA1c values. The sensitivity and specificity for diabetes were 93.5% and 97.1% with EC-PLS, 91.3% and 94.1% with CARS-PLS, and 89.1% and 76.5% with MC-UVE-PLS, respectively. In three methods, EC-PLS not only employed the least wavelengths but also produced the best quantification accuracy for HbA1c. EC-PLS also achieved the highest classification accuracy for negative and positive samples for diabetes.The results confirm the feasibility of HbA1c quantification based on the simultaneous analysis of Hb and Hb•HbA1c with NIR spectroscopy. This technique is rapid and simple when compared with conventional methods, and is a promising tool for screening diabetes in large populations.

A wavelength selection method for spectroscopic analysis, named correlation coefficient optimization coupled with partial least-squares (CCO-PLS), is proposed, and was successfully employed for reagent-free ATR-FTIR spectroscopic analysis of albumin (ALB) and globulin (GLB) in human serum. By varying the upper bound of correlation coefficient between absorbance and analyte's content, the CCO-PLS method achieved multi-band selection. Two PLS-based methods, which used a waveband having positive peaks of the first loading vector (FLV) and a combination of positive peaks of the correlation coefficient spectrum, were also conducted for comparison. Based on the leave-one-out cross-validation for CCO-PLS, appropriate waveband combinations for ALB and GLB were selected, the root-mean-square errors of prediction for validation samples were 1.36 and 1.35 (g L−1) for ALB and GLB, respectively, which were better than the two comparison methods. The CCO-PLS method provided a new approach for multi-band selection to achieve high analytical accuracy for molecular absorption bands that were composed of several spaced wavebands.

The Savitzky–Golay (SG) method and moving-window waveband screening are applied to a coupling model of principal component (PCA) and linear discriminant analyses (LDA). An SG-pretreatment-based method (MW-PCA-LDA) for spectral pattern recognition is proposed, which is successfully employed for the non-destructive recognition of transgenic sugarcane leaves using visible (Vis) and near-infrared (NIR) diffuse reflectance spectroscopy. A Kennard–Stone-algorithm-based process of calibration, prediction and validation in consideration of uniformity and representative was performed to produce objective models. A total of 456 samples of sugarcane leaves in the elongation stage were collected from a planted field. These samples were composed of 306 transgenic samples containing both Bacillus thuringiensis (Bt) and bialaphos resistance (Bar) genes, and 150 non-transgenic samples. According to the spectral recognition effects, two parallel optimal SG modes were selected. The one of the 1st order derivative, 3rd degree polynomial and 25 smoothing points was taken as an example to pretreat the diffuse reflectance spectra. Based on the MW-PCA-LDA method, the optimal waveband was 768 nm to 822 nm, the optimal PC combination was PC1–PC3 and the corresponding validation recognition rates of transgenic and non-transgenic samples achieved 99.1% and 98.0%, respectively. The results show that Vis-NIR spectroscopy combined with SG pretreatment and the MW-PCA-LDA method can be used for accurate recognition of transgenic sugarcane leaves and provides a quick and convenient means of screening transgenic sugarcane breeding for large-scale agricultural production.

A rapid determination method for the preliminary thalassaemia screening indicators—haemoglobin (Hb), mean corpuscular Hb (MCH) and mean corpuscular volume (MCV)—based on near-infrared (NIR) spectroscopy was developed. Wavelength selections were accomplished using the improved moving window partial least squares (MWPLS) method for stability. Each model was established using an approach based on randomness, similarity and stability to obtain objective, stable and practical models. The optimal wavebands screened using MWPLS were 1538 nm to 1756 nm for Hb and 968 nm to 1028 nm for MCH and MCV, which were within the NIR overtone region. The validation root mean square error, validation correlation coefficients and relative validation root mean square error of prediction were 3.9 g L−1, 0.972 and 2.9% for Hb; 1.41 pg, 0.958 and 4.8% for MCH; and 4.09 fL, 0.929 and 4.9% for MCV, respectively. The three models achieved high validation accuracy, and the Hb, MCH and MCV prediction values were close to the clinical measured values. Based on the cut-off values (MCH = 27.0 pg, MCV = 80.0 fL) for the routine method, the sensitivity and specificity both reached 100% based on the spectral-predicted MCH and MCV values in the validation set. The spectral prediction was highly accurate for determining negative and positive samples in preliminary thalassaemia screening. The proposed method is rapid and simple compared with the conventional method. This technique is a promising tool for preliminary thalassaemia screening in population prevention and control programmes. This study provides valuable references for designing specialised spectrometers.

Near-infrared (NIR) spectroscopy was successfully applied to the rapid chemical-free determination of amino acid nitrogen (AAN) in tuber mustard. Moving window partial least squares, combined with Savitzky–Golay smoothing, was used for the waveband selection. Based on the various divisions in the calibration and prediction sets, an effective modeling approach with good stability was proposed. The results confirmed that the long-wave NIR region contains sufficient information for the quantification of AAN in tuber mustard. An appropriate waveband (5874–4258 cm−1) located in the combination region of NIR was selected. The validation root mean square error of prediction and the validation correlation coefficients of prediction were 0.380 mg/g and 0.962, respectively, both indicating good validation effect and stability. The results provided a reliable NIR model and can serve as valuable references for designing spectroscopic instruments for quality evaluation of tuber mustard.Graphical abstractHighlights► Rapid chemical-free determination of amino acid nitrogen (AAN) in tuber mustard. ► A new effective modeling approach with stability was proposed. ► Long-wave NIR region contains sufficient information for AAN in tuber mustard. ► Easier and more effect waveband located in the combinations region of NIR. ► Provided references for designing instruments for quality evaluation of tuber mustard.

Near-infrared (NIR) spectroscopy combined with the moving window partial least-squares (MWPLS) method and Savitzky–Golay (SG) smoothing was successfully applied for the rapid no-reagent determination of chemical oxygen demand (COD) in sugar refinery wastewater. An appropriate waveband with stability was selected through a rigorous modeling process. Among 141 samples, 60 were randomly selected as the validation set. The remaining 81 samples were divided into the calibration set (50 samples) and the prediction set (31 samples) for a total of 20 times with certain similarities. The results showed that the short-wave NIR region (780 nm to 1100 nm) can be used as the information waveband of COD in sugar refinery wastewater, and the optimal SG smoothing mode was 5th order derivative, 5th degree polynomial, and 7 smoothing points. The waveband selection was performed in the SG smoothing spectra of the short-wave NIR region by the MWPLS method. The optimized waveband was 808 nm to 964 nm, the validation root mean square error of prediction (V-SEP) and validation correlation coefficients of prediction (V-RP) were 25.0 mg L−1 and 0.966, respectively, which had good prediction precision and stability. The results provide a reliable NIR analysis model and valuable references for designing small, specialized NIR instruments.

Savitzky–Golay (SG) smoothing and moving window partial least square (MWPLS) methods were applied to the model optimization and the waveband selection for near-infrared (NIR) spectroscopy analysis of soil organic matter. The optimal single wavelength prediction bias (OSWPB) was used to evaluate the similarity of calibration set and prediction set, and a new division method for calibration set and prediction set was proposed. SG smoothing modes were expanded to 540 kinds. The specific computer algorithm platforms for optimization of SG smoothing mode combined with PLS factor and for MWPLS method with changeable parameters were built up. The optimal waveband for soil organic matter was 1926–2032 nm, the optimal smoothing mode was the 2nd order derivative, 6th degree polynomial, 45 smoothing points, the PLS factor, RMSEP and RP were 8, 0.260 (%) and 0.877 respectively. The prediction effect was obviously better than that in the whole spectral collecting region. To get stable results, all the optimization processes were based on the average prediction effect on 50 different divisions of calibration set and prediction set.

•Equidistant combination partial least squares was proposed.•Independent and joint NIR analyses of corn.•Good prediction ability and low model complexity.Development of small, dedicated, reagentless, and low-cost spectrometer has broad application prospects in large-scale agriculture. An appropriate wavelength selection method is a key, albeit difficult, technical aspect. A novel wavelength selection method, named equidistant combination partial least squares (EC-PLS), was applied for wavenumber selection for near-infrared analysis of crude protein, moisture, and crude fat in corn. Based on the EC-PLS, a model set that includes various models equivalent to the optimal model was proposed to select independent and joint-analyses models. The independent analysis models for crude protein, moisture, and crude fat contained only 16, 12, and 22 wavenumbers, whereas the joint-analyses model for the three indicators contained only 27 wavenumbers.Random validation samples excluded from the modeling process were used to validate the four selected models. For the independent analysis models, the validation root mean square errors (V_SEP), validation correlation coefficients (V_RP), and relative validation root mean square errors (V_RSEP) of prediction were 0.271%, 0.946, and 2.8% for crude protein, 0.275%, 0.936, and 2.6% for moisture, and 0.183%, 0.924, and 4.5% for crude fat, respectively. For the joint-analyses model, the V_SEP, V_RP, and V_RSEP were 0.302%, 0.934, and 3.2% for crude protein, 0.280%, 0.935, and 2.7% for moisture, and 0.228%, 0.910, and 5.6% for crude fat, respectively. The results indicated good validation effects and low complexity. Thus, the established models were simple and efficient.The proposed wavenumber selection method provided also valuable reference for designing small dedicated spectrometer for corn. Moreover, the methodological framework and optimization algorithm are universal, such that they can be applied to other fields.Download high-res image (215KB)Download full-size image

•Repetition rate priority combination method was proposed.•Applied for NIR analysis of human serum albumin.•Effective to remove the redundant wavelengths.•High prediction performance.For the rapid measurement of an analyte in complex samples using near-infrared (NIR) spectroscopy, appropriate wavelength selection is an important and albeit difficult aspect, which is essential for improving prediction performance. Based on equidistant combination partial least squares (EC-PLS), an equivalence model set was proposed. A wavelength selection method, called repetition rate priority combination PLS (RRPC-PLS), was further proposed and applied for NIR analysis of human serum albumin. The competitive adaptive reweighted sampling combined PLS (CARS-PLS) and EC-PLS, which are well-performed wavelength selection methods, were also conducted for comparison. Based on the various divisions of calibration and prediction sets, the modeling was performed to achieve parameter stability. The posterior sample group excluded in modeling was used to validate and achieve an objective evaluation. Using CARS-PLS, EC-PLS and RRPC-PLS methods, the selected optimal models included 32, 36 and 24 wavelengths, respectively. A simpler and high performance model with 15 wavelengths was also selected with RRPC-PLS method. The root-mean-square errors and correlation coefficients for validation were 0.505 g L−1 and 0.997 for the optimal RRPC-PLS model and 0.530 g L−1 and 0.997 for the RRPC-PLS model (N=15), respectively. The validation effects were superior to the previous two methods in two aspects of prediction performance and model complexity. The prediction values were close to the measured values with high precision. The results showed that RRPC-PLS is the good improvement on EC-PLS, which can be more effective to enhance the prediction performance and remove the redundant wavelengths.

Near-infrared (NIR) spectroscopy combined with the moving window partial least-squares (MWPLS) method and Savitzky–Golay (SG) smoothing was successfully applied for the rapid no-reagent determination of chemical oxygen demand (COD) in sugar refinery wastewater. An appropriate waveband with stability was selected through a rigorous modeling process. Among 141 samples, 60 were randomly selected as the validation set. The remaining 81 samples were divided into the calibration set (50 samples) and the prediction set (31 samples) for a total of 20 times with certain similarities. The results showed that the short-wave NIR region (780 nm to 1100 nm) can be used as the information waveband of COD in sugar refinery wastewater, and the optimal SG smoothing mode was 5th order derivative, 5th degree polynomial, and 7 smoothing points. The waveband selection was performed in the SG smoothing spectra of the short-wave NIR region by the MWPLS method. The optimized waveband was 808 nm to 964 nm, the validation root mean square error of prediction (V-SEP) and validation correlation coefficients of prediction (V-RP) were 25.0 mg L−1 and 0.966, respectively, which had good prediction precision and stability. The results provide a reliable NIR analysis model and valuable references for designing small, specialized NIR instruments.

A rapid determination method for the preliminary thalassaemia screening indicators—haemoglobin (Hb), mean corpuscular Hb (MCH) and mean corpuscular volume (MCV)—based on near-infrared (NIR) spectroscopy was developed. Wavelength selections were accomplished using the improved moving window partial least squares (MWPLS) method for stability. Each model was established using an approach based on randomness, similarity and stability to obtain objective, stable and practical models. The optimal wavebands screened using MWPLS were 1538 nm to 1756 nm for Hb and 968 nm to 1028 nm for MCH and MCV, which were within the NIR overtone region. The validation root mean square error, validation correlation coefficients and relative validation root mean square error of prediction were 3.9 g L−1, 0.972 and 2.9% for Hb; 1.41 pg, 0.958 and 4.8% for MCH; and 4.09 fL, 0.929 and 4.9% for MCV, respectively. The three models achieved high validation accuracy, and the Hb, MCH and MCV prediction values were close to the clinical measured values. Based on the cut-off values (MCH = 27.0 pg, MCV = 80.0 fL) for the routine method, the sensitivity and specificity both reached 100% based on the spectral-predicted MCH and MCV values in the validation set. The spectral prediction was highly accurate for determining negative and positive samples in preliminary thalassaemia screening. The proposed method is rapid and simple compared with the conventional method. This technique is a promising tool for preliminary thalassaemia screening in population prevention and control programmes. This study provides valuable references for designing specialised spectrometers.

The Savitzky–Golay (SG) method and moving-window waveband screening are applied to a coupling model of principal component (PCA) and linear discriminant analyses (LDA). An SG-pretreatment-based method (MW-PCA-LDA) for spectral pattern recognition is proposed, which is successfully employed for the non-destructive recognition of transgenic sugarcane leaves using visible (Vis) and near-infrared (NIR) diffuse reflectance spectroscopy. A Kennard–Stone-algorithm-based process of calibration, prediction and validation in consideration of uniformity and representative was performed to produce objective models. A total of 456 samples of sugarcane leaves in the elongation stage were collected from a planted field. These samples were composed of 306 transgenic samples containing both Bacillus thuringiensis (Bt) and bialaphos resistance (Bar) genes, and 150 non-transgenic samples. According to the spectral recognition effects, two parallel optimal SG modes were selected. The one of the 1st order derivative, 3rd degree polynomial and 25 smoothing points was taken as an example to pretreat the diffuse reflectance spectra. Based on the MW-PCA-LDA method, the optimal waveband was 768 nm to 822 nm, the optimal PC combination was PC1–PC3 and the corresponding validation recognition rates of transgenic and non-transgenic samples achieved 99.1% and 98.0%, respectively. The results show that Vis-NIR spectroscopy combined with SG pretreatment and the MW-PCA-LDA method can be used for accurate recognition of transgenic sugarcane leaves and provides a quick and convenient means of screening transgenic sugarcane breeding for large-scale agricultural production.

A wavelength selection method for spectroscopic analysis, named correlation coefficient optimization coupled with partial least-squares (CCO-PLS), is proposed, and was successfully employed for reagent-free ATR-FTIR spectroscopic analysis of albumin (ALB) and globulin (GLB) in human serum. By varying the upper bound of correlation coefficient between absorbance and analyte's content, the CCO-PLS method achieved multi-band selection. Two PLS-based methods, which used a waveband having positive peaks of the first loading vector (FLV) and a combination of positive peaks of the correlation coefficient spectrum, were also conducted for comparison. Based on the leave-one-out cross-validation for CCO-PLS, appropriate waveband combinations for ALB and GLB were selected, the root-mean-square errors of prediction for validation samples were 1.36 and 1.35 (g L−1) for ALB and GLB, respectively, which were better than the two comparison methods. The CCO-PLS method provided a new approach for multi-band selection to achieve high analytical accuracy for molecular absorption bands that were composed of several spaced wavebands.