Characterization of Edible Seaweed Harvested on the Galician Coast (Northwestern Spain) Using Pattern Recognition Techniques and Major and Trace Element Data

• Published in: Journal of agricultural and food chemistry Vol. 58; no. 3; pp. 1986 - 1992
• Main Authors: Romarís-Hortas, Vanessa, García-Sartal, Cristina, Barciela-Alonso, María Carmen, Moreda-Piñeiro, Antonio, Bermejo-Barrera, Pilar
• Format: Journal Article
• Language: English
• Published: Washington, DC American Chemical Society 10.02.2010
• Subjects: accuracy; Animal, plant, fungal and microbial proteins, edible seaweeds and food yeasts; automatic detection; Biological and medical sciences; coasts; edible seaweed; food analysis; Food Chemistry/Biochemistry; food composition; Food industries; Fundamental and applied biological sciences. Psychology; mineral content; pattern recognition techniques; Pigmentation; prediction; Seaweed - chemistry; Spain; trace elements; Trace Elements - analysis; Spain; Europe; Seaweed; soft independent modeling of class analogy; trace elements; atomic spectrometry; cluster analysis; linear discriminant analysis; principal component analysis; Cluster analysis; Discriminant analysis; Algae; Harvesting; Pattern recognition; Modeling; Edible alga; Characterization; Trace element; Spectrometry; Principal component analysis
• ISSN: 0021-8561; 1520-5118
• DOI: 10.1021/jf903677y

Major and trace elements in North Atlantic seaweed originating from Galicia (northwestern Spain) were determined by using inductively coupled plasma−optical emission spectrometry (ICP-OES) (Ba, Ca, Cu, K, Mg, Mn, Na, Sr, and Zn), inductively coupled plasma−mass spectrometry (ICP-MS) (Br and I) and hydride generation−atomic fluorescence spectrometry (HG-AFS) (As). Pattern recognition techniques were then used to classify the edible seaweed according to their type (red, brown, and green seaweed) and also their variety (Wakame, Fucus, Sea Spaghetti, Kombu, Dulse, Nori, and Sea Lettuce). Principal component analysis (PCA) and cluster analysis (CA) were used as exploratory techniques, and linear discriminant analysis (LDA) and soft independent modeling of class analogy (SIMCA) were used as classification procedures. In total, t12 elements were determined in a range of 35 edible seaweed samples (20 brown seaweed, 10 red seaweed, 4 green seaweed, and 1 canned seaweed). Natural groupings of the samples (brown, red, and green types) were observed using PCA and CA (squared Euclidean distance between objects and Ward method as clustering procedure). The application of LDA gave correct assignation percentages of 100% for brown, red, and green types at a significance level of 5%. However, a satisfactory classification (recognition and prediction) using SIMCA was obtained only for red seaweed (100% of cases correctly classified), whereas percentages of 89 and 80% were obtained for brown seaweed for recognition (training set) and prediction (testing set), respectively.