An imaging method using fluorescence fingerprints (FFs) for visualizing the distribution of gluten and starch in dough without any staining was developed. Wheat flour dough was mixed up to three stages, i.e., under-mixing, optimum-mixing, and over-mixing, and thin sections of the doughs were prepared with a cryotome. Fluorescence images of the sections were acquired in 63 combinations of excitation and emission wavelengths, thereby constructing the FFs of the constituents at each pixel. The cosine similarity between the FFs of each pixel and the FF of pure gluten were calculated and pixels were arranged in order of cosine similarity. Pixels with higher values of cosine similarity were categorized as "gluten" and the rest as "nongluten". The number of pixels categorized as "gluten" was based on the overall ratio of gluten in the dough. The same process was performed with the FF of pure starch, and all pixels were divided into "starch" and "nonstarch". Colors were assigned to each division, and the distributions of gluten and starch were visualized. Changes in the distributions of gluten and starch were observed at the over-mixing stage, which suggested the breaking up of gluten and the alteration of gluten and starch.Keywords: excitation-emission matrix, imaging, autofluorescence, cosine similarity *To whom correspondence should be addressed. E-mail: sugiyama@affrc.go.jp
IntroductionBread palatability is determined by many factors, one of which is bread texture (Babin et al., 2005). Bread texture largely depends on the condition of the dough from which bread is made, particularly the rheological properties of the dough. In wheat flour dough, the microscopic spatial arrangement of gluten and starch greatly affects dough rheology, and many advanced microscopic techniques have been applied to observe the microstructure of dough (Peighambardoust et al., 2010).Specific chemical compounds in the dough are mainly visualized by combining light microscopy and staining. Many approaches to light microscopy have been used to study dough, such as bright field (Auger et al., 2008;Kuktaite et al., 2005), confocal scanning laser (Durrenberger et al., 2001;Peressini et al., 2008;Renzetti and Arendt, 2009) and epifluorescence light (Peighambardoust et al., 2010) microscopies. Stains are used to differentiate between constituents of interest such as gluten and starch; however, they may introduce sample alteration or artifacts. The results could also vary depending on the selected stains and staining conditions such as concentration, solvent and staining time.Specific constituents have been visualized without staining by hyperspectral or multispectral imaging, a method that integrates conventional imaging with spectroscopy (Gowen et al., 2007). Hyperspectral or multispectral images are composed of multiple wavebands for each spatial position studied, and light absorption/fluorescence/reflectance at these wavebands give intrinsic information on the constituent at that position.Recently, fluorescence fingerprints (FFs) have been used ...