Detection of Scab-Damaged Hard Red Spring Wheat Kernels by Near-Infrared Reflectance

Cereal Chem. 81(5):643–649 Scab (Fusarium head blight) is a fungal disease that has become increasingly prevalent in North American wheat during the past 15 years. It is of concern to growers, processors, and the consumers because of depressed yields, poor flour quality, and the potential for elevated concentrations of the mycotoxin, deoxynivalenol (DON). Both wheat breeder and wheat inspector must currently deal with the assessment of scab in harvested wheat by manual human inspection. The study described herein examined the accuracy of a semi-automated wheat scab inspection system that is based on near-infrared (NIR) reflectance (1,000– 1,700 nm) of individual kernels. Using statistical classification techniques such as linear discriminant analysis and nonparametric (k-nearest-neighbor) classification, upper limits of accuracy for NIR-based classification schemes of ≈88% (cross-validation) and 97% (test) were determined. An exhaustive search of the most suitable wavelength pairs for the spectral difference, [log(1/R)λ1 – log(1/R)λ2], revealed that the slope of the lowwavelength side of a broad carbohydrate absorption band (centered at ≈1,200 nm) was very effective at discriminating between healthy and scab-damaged kernels with test set accuracies of 95%. The achieved accuracy levels demonstrate the potential for the use of NIR spectroscopy in commercial sorting and inspection operations for wheat scab. Fusarium head blight, also known as scab, is a fungal disease that affects wheat and other cereals. Primarily caused by F. graminearum, scab occurs during the flowering and early kernel development stages of the plant under conditions of rain and elevated humidities. Depending on the year, hard red spring wheats of the Northern Great Plains of the United States and western prairie provinces of Canada are susceptible to scab (such as in 1993), as are the soft red winter wheat growing regions in the eastern United States, particularly in 1996 and 2003. Infection with Fusarium results in depressed yields caused by kernels that are either too light in mass for mechanized harvesting (Bai and Shaner 1994) or low in test weight (Cunfer 1987; Dexter et al 1996). Scab infection may also have a deleterious effect on flour color, ash content, and baking performance (Dexter et al 1996), with these and other quality issues recently reviewed by Dexter and Nowicki (2003). Also considered as a component in a category collectively called damaged in federal official inspection, wheat lots >2% damaged kernels by weight are progressively downgraded from the U.S. No. 1 grade (USDA-GIPSA 1997). Grade is reduced (to No. 5) with increased levels of damaged kernels until the weight concentration reaches 15%, whereupon a lot is assigned the grade of sample grade. Often, domestic processors and overseas buyers have even more stringent criteria for percentages of scab-damaged kernels. Beyond that of physical damage to the kernel, F. graminearum can produce the metabolite, deoxynivalenol (DON), or vomitoxin, a tricothecene mycotoxin that is a health concern. The U.S. Food and Drug Administration (FDA) has established an advisory level for DON in finished wheat products (e.g., flour, semolina) intended for human consumption at 1 ppm, and at higher levels (5–10 ppm) for livestock and poultry feeds. Teich et al (1987) measured a positive correlation between the levels of DON and scab damage. Therefore, as a means to improve the quality and safety of wheat in the United States, the USDA is currently seeking new methods for grain inspection, inclusive of those that address kernel damage (e.g., scab) that are rapid and objectively based. Numerous procedures are available for measurement of DON level concentration in wheat meal and flour, including those based on thin-layer chromatography (Truckness et al 1984; Fernandez et al 1994), liquid chromatography (Chang et al 1984; Trenholm et al 1985; Rajaklyä et al 1987), gas chromatography (GC) (Terhune et al 1984; Ware et al 1984; Scott et al 1986; Tacke et al 1996), GC with mass spectrometry (Scott et al 1981), and enzyme-linked immunosorbent assay (ELISA) (Casale et al 1988; Usleber et al 1991, 1993; Abramson et al 1998). However, these procedures are not adaptable to rapid (<1 min/sample) testing for inspection or process control. Rapid inspection or sorting methods for grain are typically based on kernel density (Tkachuk et al 1991) or optical properties (Ruan et al 1998). Preliminary research by Williams (1997) demonstrated moderate success with the NIR modeling of DON in bulk samples. Recent work by Pettersson and Åberg (2003) on scab-damaged wheat prepared by dilution series has also indicated the potential of DON level by whole grain NIR transmittance. At the single kernel level, Dowell et al (1999) examined the ability of NIR reflectance to measure DON concentration. Using a similar approach to that described herein on a limited set of officially inspected wheat, we demonstrated the ability to distinguish normal, molddamaged, and scab-damaged categories (Delwiche 2003). Based on a more diverse collection of hard red spring wheat samples, the research described herein was directed at identifying regions within the near-infrared (NIR) region of 1,000–1,700 nm that can be used, with or without kernel mass, to determine the extent of scab-damage within wheat samples through single kernel inspection. Specifically, the objective of this study was to identify and characterize a pair of wavelengths for an NIR reflectance system that can form the basis of a commercial sorting device or inspection instrument. MATERIALS AND METHODS Wheat. Hard red spring wheat samples were obtained from two distinct breeders’ trials. The first trial consisted of the 37 lines that were the basis for the Hard Red Spring Uniform Regional Nursery (HRS-URN) trial set for the 2002 harvest. Of the 37 lines, five were commercial releases (cultivars Marquis, Chris, 2375, Verde, and Keene); the remaining were breeders’ advanced lines. All lines were grown in field plots under conventional growing practices at two geographical locations, Crookston and St. Paul, MN. The second trial consisted of a breeder’s advanced 1 USDA/ARS, Beltsville Agricultural Research Center, Instrumentation and Sensing Laboratory, Building 303, BARC-East, Beltsville, MD 20705-2350. Names are necessary to report factually on available data; however, the USDA neither guarantees nor warrants the standard of the product, and the use of the name by the USDA implies no approval of the product to the exclusion of others that may also be suitable. 2 Corresponding author. E-mail: [email protected] 3 USDA/ARS, Hard Red Spring and Durum Wheat Quality Laboratory, Fargo, ND. Publication no. C-2004-0805-02R. This article is in the public domain and not copyrightable. It may be freely reprinted with customary crediting of the source. American Association of Cereal Chemists, Inc., 2004.