We applied near-infrared (NIR) spectroscopy with chemometrics for the rapid and reagent-free analysis of serum urea nitrogen (SUN). The modeling is based on the average effect of multiple sample partitions to achieve parameter selection with stability. A multiparameter optimization platform with Norris derivative filter–partial least squares (Norris-PLS) was developed to select the most suitable mode （d = 2, s = 33, g = 15）T. Using equidistant combination PLS (EC-PLS) with four parameters (initial wavelength I, number of wavelengths N, number of wavelength gaps G and latent variables LV), we performed wavelength screening after eliminating highabsorption wavebands. The optimal EC-PLS parameters were I = 1228 nm, N = 26, G = 16 and LV = 12. The root-mean-square error (SEP), correlation coefficient eRP T for prediction and ratio of performance-to-deviation (RPD) for validation were 1.03 mmol·L^-1, 0.992 and 7.6, respectively. We proposed the wavelength step-by-step phase-out PLS (WSP-PLS) to remove redundant wavelengths in the top 100 EC-PLS models with improved prediction performance. The combination of 19 wavelengths was identified as the optimal model for SUN. The SEP, RP and RPD in validation were 1.01 mmol·L^-1, 0.992 and 7.7, respectively. The prediction effect and wavelength complexity were better than those of EC-PLS. Our results showed that NIR spectroscopy combined with the EC-PLS and WSP-PLS methods enabled the high-precision analysis of SUN. WSP-PLS is a secondary optimization method that can further optimize any wavelength model obtained through other continuous or discrete strategies to establish a simple and better model.

The moving-window bis-correlation coe±cients (MW-BiCC) was proposed and employed for the discriminant analysis of transgenic sugarcane leaves and --thalassemia with visible and nearinfrared (Vis–NIR) spectroscopy. The well-performed moving-window principal component analysis linear discriminant analysis (MW-PCA–LDA) was also conducted for comparison. A total of 306 transgenic (positive) and 150 nontransgenic (negative) leave samples of sugarcane were collected and divided to calibration, prediction, and validation. The diffuse reflection spectra were corrected using Savitzky–Golay (SG) smoothing with first-order derivative (d=1), third-degree polynomial (p=3) and 25 smoothing points (m=25). The selected waveband was 736–1054 nm with MW-BiCC, and the positive and negative validation recognition rates (V REC+, V REC-T were 100%, 98.0%, which achieved the same effect as MW-PCA–LDA. Another example, the 93 --thalassemia (positive) and 148 nonthalassemia (negative) of human hemolytic samples were collected. The transmission spectra were corrected using SG smoothing with d=1, p=3 and m=53. Using MW-BiCC, many best wavebands were selected (e.g., 1116–1146, 1794–1848 and 2284–2342nm). The V REC+ and V REC- were both 100%, which achieved the same effect as MW-PCA–LDA. Importantly, the BiCC only required calculating correlation coe±cients between the spectrum of prediction sample and the average spectra of two types of calibration samples. Thus, BiCC was very simple in algorithm, and expected to obtain more applications. The results first confirmed the feasibility of distinguishing --thalassemia and normal control samples by NIR spectroscopy, and provided a promising simple tool for large population thalassemia screening.

Teicoplanin (TCP) is an important lipoglycopeptide antibiotic produced by fermenting Acti-noplanes teichomyceticus. The change in TCP concentration is important to measure in the fermentation process. In this study, a reagent-free and rapid quantification method for TCP in the TCP–Tris–HCl mixture samples was developed using near-infrared (NIR) spectroscopy by focusing our attention on the fermentation process for TCP. The absorbance optimization (AO) partial least squares (PLS) was proposed and integrated with the moving window (MW) PLS, which is called AO–MW–PLS method, to select appropriate wavebands. A model set that includes various wavebands that were equivalent to the optimal AO–MW–PLS waveband was proposed based on statistical considerations. The public region of all equivalent wavebands was just one of the equivalent wavebands. The obtained public regions were 1540–1868 nm for TCP and 1114–1310 nm for Tris. The root-mean-square error and correlation coefficient for leave-one-out cross validation were 0.046 mg mL 1 and 0.9998 mg mL 1 for TCP, and 0.235 mg mL 1 and 0.9986 mg mL 1 for Tris, respectively. All the models achieved highly accurate prediction effects, and the selected wavebands provided valuable references for designing specialized spectrometers. This study provided a valuable reference for further application of the proposed methods to TCP fermentation broth and to other spectroscopic analysis fields.

The selection of stable wavebands for the near-infrared (NIR) spectroscopic analysis of total nitrogen (TN) in soil was accomplished by using an improved moving window partial least squares (MWPLS) method. A new modeling approach was performed based on randomness, similarity and stability, which produced an objective, stable and practical model. Based on the MWPLS method, a search was in the overall scanning region from 400 to 2498 nm, and the optimal waveband was identified to be 1424 to 2282 nm. A model space that includes 41 wavebands that are equivalent to the optimal waveband was then proposed. The public range of the 41 equivalent optimal wavebands was 1590 to 1870 nm, which contained sufficient TN information. The wavebands of 1424 to 2282 nm, 1590 to 1870 nm, and the long-NIR region 1100 to 2498 nm all achieved satisfactory validation effects. However, the public waveband of 1590 to 1870 nm had only a minimum number of wavelengths, which significantly reduced the method complexity. Various equivalent wavebands serve as guidelines for designing spectroscopic instruments. These wavebands could address the restrictions of position and the number of wavelengths in instrument design.

A new strategy for quantitative analysis of a major clinical biochemical indicator called glycated hemoglobin (HbA1c) was proposed. The technique was based on the simultaneous near-infrared (NIR) spectral determination of hemoglobin (Hb) and absolute HbA1c content (Hb · HbA1c) in human hemolysate samples. Wavelength selections were accomplished using the improved moving window partial least square (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 obtained using MWPLS were 958 to 1036nm for Hb and 1492 to 1858 nm for Hb · HbA1c, which were within the NIR overtone region. The validation root mean square error and validation correlation coefficients of prediction (V -SEP, V -R_P) were 3.4 g L^-1 and 0.967 for Hb, respectively, whereas the corresponding values for Hb · HbA1c were 0.63 g L^-1 and 0.913. The corresponding V -SEP and V -R_P were 0.40% and 0.829 for the relative percentage of HbA1c. The experimental results confirm the feasibility for the quantification of HbA1c based on simultaneous NIR spectroscopic analyses of Hb and Hb · HbA1c.