Application of reflectance spectroscopy to identify maize genotypes and aflatoxin levels in single kernels

Spectroscopy is a rapid, non-destructive, and low-cost analytical technique that has the potential to complement more resource-intensive analytical methods. We explored the use of spectral methods to differentiate maize genotypes and assess aflatoxin (AF) contamination in maize kernels. We compared the performance of two instruments: a research-grade ultraviolet-visible-near infrared (UV-Vis-NIR) spectrometer that measures reflectance from 304 -1,085 nm, and a miniaturised NIR spectrometer that measures reflectance from 740-1,070 nm. Both systems were used to predict AF levels in maize kernels from a single genotype and across 10 genotypes, and to predict genotype for the latter. A partial least square discriminant analysis model was trained on 70% of the kernels and tested on the remaining 30%. The classification accuracy for 10 maize genotypes was 71-72% using the UV-Vis-NIR instrument on 1,170 kernels, and 65-66% using the NIR device on 740 kernels. The classification accuracy for 247 AF-contaminated kernels of a single genotype using the UV-Vis-NIR instrument was 71, 82, and 92% for AF thresholds of 20, 100, and 1000 μg/kg, respectively. Using the same spectrometer on 872 kernels from 10 genotypes, AF classification accuracy was 67, 90, and 95% in validation sets for AF thresholds of 20, 100, and 1000 μg/kg, respectively. The UV-Vis-NIR instrument and the NIR device had similar classification accuracies for AF thresholds of 100 and 1000 μg/kg, whereas the NIR device had higher accuracy for the AF threshold of 20 μg/kg. Reflectance spectroscopy outperformed visual sorting and the bright greenish yellow fluorescence test in identifying AF levels. Applying spectral analysis to estimate mycotoxin levels and to identify maize genotypes could contribute to regional toxin surveillance and action efforts. Further, using AF-associated spectral features for grain sorting can reduce AF exposure.

Geographical Differentiation of Hom Mali Rice Cultivated in Different Regions of Thailand Using FTIR-ATR and NIR Spectroscopy

Although Hom Mali rice is considered the highest quality rice in Thailand, it is susceptible to adulteration and substitution. There is a need for rapid, low-cost and efficient analytical techniques for monitoring the authenticity and geographical origin of Thai Hom Mali rice. In this study, two infrared spectroscopy techniques, Fourier-transform infrared spectroscopy with attenuated total reflection (FTIR-ATR) and near-infrared (NIR) spectroscopy, were applied and compared for the differentiation of Thai Hom Mali rice from two geographical regions over two production years. The Orthogonal Projections to Latent Structures Discriminant Analysis (OPLS-DA) model, built using spectral data from the benchtop FTIR-ATR, achieved 96.97% and 100% correct classification of the test dataset for each of the production years, respectively. The OPLS-DA model, built using spectral data from the portable handheld NIR, achieved 84.85% and 86.96% correct classification of the test dataset for each of the production years, respectively. Direct NIR analysis of the polished rice grains (i.e., no sample preparation) was determined as reliable for analysis of ground rice samples. FTIR-ATR and NIR spectroscopic analysis both have significant potential as screening tools for the rapid detection of fraud issues related to the geographical origin of Thai Hom Mali rice.

Application of Visible/Infrared Spectroscopy and Hyperspectral Imaging With Machine Learning Techniques for Identifying Food Varieties and Geographical Origins

Food quality and safety are strongly related to human health. Food quality varies with variety and geographical origin, and food fraud is becoming a threat to domestic and global markets. Visible/infrared spectroscopy and hyperspectral imaging techniques, as rapid and non-destructive analytical methods, have been widely utilized to trace food varieties and geographical origins. In this review, we outline recent research progress on identifying food varieties and geographical origins using visible/infrared spectroscopy and hyperspectral imaging with the help of machine learning techniques. The applications of visible, near-infrared, and mid-infrared spectroscopy as well as hyperspectral imaging techniques on crop food, beverage, fruits, nuts, meat, oil, and some other kinds of food are reviewed. Furthermore, existing challenges and prospects are discussed. In general, the existing machine learning techniques contribute to satisfactory classification results. Follow-up researches of food varieties and geographical origins traceability and development of real-time detection equipment are still in demand.

Spectral-Based Screening Approach Evaluating Two Specific Maize Lines With Divergent Resistance to Invasion by Aflatoxigenic Fungi

In an effort to control aflatoxin contamination in food and/or feed grains, a segment of research has focused on host resistance to eliminate aflatoxin from susceptible crops, including maize. To this end, screening tools are key to identifying resistant maize genotypes. The traditional field screening techniques, the kernel screening laboratory assay (KSA), and analytical methods (e.g., ELISA) used for evaluating corn lines for resistance to fungal invasion, all ultimately require sample destruction. A technological advancement on the basic BGYF presumptive screening test, fluorescence hyperspectral imaging offers an option for non-destructive and rapid image-based screening. The present study aimed to differentiate fluorescence spectral signatures of representative resistant and susceptible corn hybrids infected by a toxigenic (SRRC-AF13) and an atoxigenic (SRRC-AF36) strain of Aspergillus flavus, at several time points (5, 7, 10, and 14 days), in order to evaluate fluorescence hyperspectral imaging as a viable technique for early, non-invasive aflatoxin screening in resistant and susceptible corn lines. The study utilized the KSA to promote fungal growth and aflatoxin production in corn kernels inoculated under laboratory conditions and to provide actual aflatoxin values to relate with the imaging data. Each time point consisted of 78 kernels divided into four groups (30-susceptible, 30-resistant, 9-susceptible control, and 9-resistant control), per inoculum. On specified days, kernels were removed from the incubator and dried at 60°C to terminate fungal growth. Dry kernels were imaged with a VNIR hyperspectral sensor (image spectral range of 400–1000 nm), under UV excitation centered at 365 nm. Following imaging, kernels were submitted for the chemical AflaTest assay (VICAM). Fluorescence emissions were compared for all samples over 14 days. Analysis of strain differences separating the fluorescence emission peaks of resistant from the susceptible strain indicated that the emission peaks of the resistant strain and the susceptible strains differed significantly (p < 0.01) from each other, and there was a significant difference in fluorescence intensity between the treated and control kernels of both strains. These results indicate a viable role of fluorescence hyperspectral imaging for non-invasive screening of maize lines with divergent resistance to invasion by aflatoxigenic fungi.

Cultivar Classification of Single Sweet Corn Seed Using Fourier Transform Near-Infrared Spectroscopy Combined with Discriminant Analysis

Seed purity is a key indicator of crop seed quality. The conventional methods for cultivar identification are time-consuming, expensive, and destructive. Fourier transform near-infrared (FT-NIR) spectroscopy combined with discriminant analyses, was studied as a rapid and nondestructive technique to classify the cultivars of sweet corn seeds. Spectra with a range of 1000–2500 nm collected from 760 seeds of two cultivars were used for the discriminant analyses. Thereafter, 126 feature wavelengths were identified from 1557 wavelengths using a genetic algorithm (GA) to build simplified classification models. Four classification algorithms, namely K-nearest neighbor (KNN), soft independent method of class analogy (SIMCA), partial least-squares discriminant analysis (PLS-DA), and support vector machine discriminant analysis (SVM-DA) were tested on full-range wavelengths and feature wavelengths, respectively. With the full-range wavelengths, all four algorithms achieved a high classification accuracy range from 97.56% to 99.59%, and the SVM-DA worked better than other models. From the feature wavelengths, no significant decline in accuracies was observed in most of the models and a high accuracy of 99.19% was still obtained by the PLS-DA model. This study demonstrated that using the FT-NIR technique with discriminant analyses could be a feasible way to classify sweet corn seed cultivars and the proper classification model could be embedded in seed sorting machinery to select high-purity seeds.

Evaluation of a miniaturized NIR spectrometer for cultivar identification: The case of barley, chickpea and sorghum in Ethiopia

Crop cultivar identification is fundamental for agricultural research, industry and policies. This paper investigates the feasibility of using visible/near infrared hyperspectral data collected with a miniaturized NIR spectrometer to identify cultivars of barley, chickpea and sorghum in the context of Ethiopia. A total of 2650 grains of barley, chickpea and sorghum cultivars were scanned using the SCIO, a recently released miniaturized NIR spectrometer. The effects of data preprocessing techniques and choosing a machine learning algorithm on distinguishing cultivars are further evaluated. Predictive multiclass models of 24 barley cultivars, 19 chickpea cultivars and 10 sorghum cultivars delivered an accuracy of 89%, 96% and 87% on hold-out sample. The Support Vector Machine (SVM) and Partial least squares discriminant analysis (PLS-DA) algorithms consistently outperformed other algorithms. Several cultivars, believed to be widely adopted in Ethiopia, were identified with perfect accuracy. These results advance the discussion on cultivar identification survey methods by demonstrating that miniaturized NIR spectrometers represent a low-cost, rapid and viable tool. We further discuss the potential utility of the method for adoption surveys, field-scale agronomic studies, socio-economic impact assessments and value chain quality control. Finally, we provide a free tool for R to easily carry out crop cultivar identification and measure uncertainty based on spectral data.

Multi-platform metabolomics analyses of a broad collection of fragrant and non-fragrant rice varieties reveals the high complexity of grain quality characteristics

The quality of rice in terms not only of its nutritional value but also in terms of its aroma and flavour is becoming increasingly important in modern rice breeding where global targets are focused on both yield stability and grain quality. In the present paper we have exploited advanced, multi-platform metabolomics approaches to determine the biochemical differences in 31 rice varieties from a diverse range of genetic backgrounds and origin. All were grown under the specific local conditions for which they have been bred and all aspects of varietal identification and sample purity have been guaranteed by local experts from each country. Metabolomics analyses using 6 platforms have revealed the extent of biochemical differences (and similarities) between the chosen rice genotypes. Comparison of fragrant rice varieties showed a difference in the metabolic profiles of jasmine and basmati varieties. However with no consistent separation of the germplasm class. Storage of grains had a significant effect on the metabolome of both basmati and jasmine rice varieties but changes were different for the two rice types. This shows how metabolic changes may help prove a causal relationship with developing good quality in basmati rice or incurring quality loss in jasmine rice in aged grains. Such metabolomics approaches are leading to hypotheses on the potential links between grain quality attributes, biochemical composition and genotype in the context of breeding for improvement. With this knowledge we shall establish a stronger, evidence-based foundation upon which to build targeted strategies to support breeders in their quest for improved rice varieties.

Review of methods for the detection and quantification of adulteration of rice: Basmati as a case study

Rice is a staple and widely grown crop endowed with rich genetic diversity. As it is difficult to differentiate seeds of various rice varieties based on visual observation accurately, the harvested seeds and subsequent processed products are highly prone to adulteration with look-alike and low quality seeds by the dishonest traders. To protect the interests of importing countries and consumers, several methods have been employed over the last few decades for unambiguous discrimination of cultivars, accurate quantification of the adulterants, and for determination of cultivated geographical area. With recent advances in biotechnology, DNA based techniques evolved rapidly and proved successful over conventional non-DNA based methods to purge the problem of adulteration at commercial level. In the current review, we made an attempt to summarize the existing methods of adulteration detection and quantification in a comprehensive manner by providing Basmati as a case study to enable the traders to arrive at a quick resolution in choosing the apt method to eliminate the adulteration practice in the global rice industry.

Rice seed cultivar identification using near-infrared hyperspectral imaging and multivariate data analysis

A near-infrared (NIR) hyperspectral imaging system was developed in this study. NIR hyperspectral imaging combined with multivariate data analysis was applied to identify rice seed cultivars. Spectral data was exacted from hyperspectral images. Along with Partial Least Squares Discriminant Analysis (PLS-DA), Soft Independent Modeling of Class Analogy (SIMCA), K-Nearest Neighbor Algorithm (KNN) and Support Vector Machine (SVM), a novel machine learning algorithm called Random Forest (RF) was applied in this study. Spectra from 1,039 nm to 1,612 nm were used as full spectra to build classification models. PLS-DA and KNN models obtained over 80% classification accuracy, and SIMCA, SVM and RF models obtained 100% classification accuracy in both the calibration and prediction set. Twelve optimal wavelengths were selected by weighted regression coefficients of the PLS-DA model. Based on optimal wavelengths, PLS-DA, KNN, SVM and RF models were built. All optimal wavelengths-based models (except PLS-DA) produced classification rates over 80%. The performances of full spectra-based models were better than optimal wavelengths-based models. The overall results indicated that hyperspectral imaging could be used for rice seed cultivar identification, and RF is an effective classification technique.