Modelling the vigour of maize seeds submitted to artificial accelerated ageing based on ATR-FTIR data and chemometric tools (PCA, HCA and PLS-DA)

Abstract

The main goals of this research were to use ATR-FTIR spectroscopy associated with multivariate analyses to identify biochemical changes in high and low vigour seed tissues (embryo and endosperm) in response to accelerated ageing and to create a model to predict seed vigour based on spectroscopic data. High-vigour seeds undergo minimal changes in biochemical composition during stress by accelerated ageing while low-vigour seeds are more sensitive to stress and this lower tolerance is associated with reduced lipid and protein content and increased amino acids, carbohydrates and phosphorus compounds in the embryo. High-vigour seeds show an increase in peaks associated with amino acids and phosphorous compounds in the endosperm after 24 h of stress while low-vigour seeds present these high-intensity peaks only after 72 h in the embryo. The results of this research provide the theoretical basis for the genetic improvement of maize cultivars that aim at higher physiological seed quality.

Machine Learning for Seed Quality Classification: An Advanced Approach Using Merger Data from FT-NIR Spectroscopy and X-ray Imaging

“…The spectral bands that had greater contribution to classify the germination capacity of U. brizantha seeds were 1221,1902,2029,2037,2045,2230,2259,2289,2309, 2320, 2351 nm. The chemical compounds related to these wavelengths are amino acids, carbohydrates (cellulose, hemicellulose, pectic polysaccharides, pyranose compounds, starch, and sucrose) and nucleic acids [11,13,25,29,30]. Using the NIR data, the best algorithm to classify the seed germination capacity was PLS-DA, reaching an accuracy of 82%.…”

“…Meanwhile, methods based on spectrometry and X-ray imaging techniques have been successfully used to collect data on complex traits related to seed quality. For instance, Fourier transform near-infrared (FT-NIR) spectroscopy has proved great potential in detecting seed compounds by acquisition of a large number of spectral details [6][7][8][9][10][11][12]. FT-NIR spectroscopy is based on the absorption of electromagnetic radiation at wavelengths ranging from 780 to 2500 nm [13].…”

“…Robust algorithms can capture linear and non-linear relationships, and they can achieve high classification accuracy. Several algorithms have been proven to be effective for solving problems in many fields of research, such as linear discriminant analysis (LDA), partial least squares discriminant analysis (PLS-DA), random forest (RF), naive Bayes (NB), support vector machine with linear (SVM-l) and radial basis (SVM-r) kernel, and artificial neural network (ANN) [6,9,11,19,20]. However, different algorithms operate differently and they can have different performances [20].…”

Grade Diagnosis of Human Glioma Based on Fingerprint and Artificial Neural Network

“…Multifactorial statistical analysis methods related to FTIR have been widely used for identifying changes in lipids, proteins, nucleic acids, and carbohydrates, such as principal component analysis (PCA) [15][16][17] and partial least squares (PLS) [18,19] combined with discriminant analysis (DA), hierarchical cluster analysis (HCA) [20,21], support vector machines (SVMs) [22,23] and random forest (RF) [24]. Smith et al [25] used the supervised machine learning algorithm of RF as a classifier to separate patients into cancer and noncancer categories based upon the intensities of wavenumbers presented in their spectra and finally achieved a sensitivity and specificity up to 92.8% and 91.5%, respectively.…”

Use of non-destructive instrumental techniques to evaluate effects of gamma irradiation, quality and sensory attributes in orange juice

“…At the same time, with the advancement of technology in food analysis, more data and information are produced, which makes it essential to apply chemometric tools or data science. This combination allows several applications in food analysis, including the extraction of important information from the spectra by principal component analysis (PCA) (Andrade, Medeiros Coelho, & Uarrota, 2020), development of predictive and classification models, by the method of partial least squares regression (PLS) (Croce et al, 2020) and method of discriminant analysis by partial least squares (PLS-DA), respectively. Thus, new applications of nondestructive methods combined with data science, may represent alternatives for juice processing companies to measure the impact of fruit quality at the end and thus adopt corrective measures to improve juice production processes throughout the year.…”