Determination of geographical origin Turkish hazelnuts according to fatty acid composition

Abstract This study focuses on detecting geographical origin of round type hazelnut (Corylus avellena L.) in Turkey using fatty acid (FA) composition. The samples were collected from Western, Central, and Eastern Black Sea regions between 2015 and 2016. FA profiles were determined by gas chromatography (GC), and most abundant fatty acid was oleic acid (C18:1) followed by linoleic acid (C18:2), palmitic acid (C16:0), and stearic acid (C18:1). The effect of geographical origin on the fatty acid profile of hazelnut oils was statistically analyzed by one‐way ANOVA and linear discriminant analysis (LDA). The results showed that the Central Black Sea region had high content of total saturated fatty acids (%8.45), total monounsaturated fatty acids (%83.54), low content of total polyunsaturated fatty acids (%7.85), and Eastern Black Sea region had high content of linoleic (%9.10) and linolenic acid (%0.096). Six fatty acids (C16:1, C18:1, C18:2, C18:3, C20:0, and C20:1) identified by LDA provide 86.2% of correct predictions.

analyzing using one-way ANOVA and linear discrimination analysis (LDA) and to estimate diversity of the hazelnut related to with fatty acid composition and to compare this properties to appropriate other countries.

| Hazelnut samples
The hazelnut samples were randomly selected from local hazelnut traders. The samples were harvested in September of 2015 and 2016, and the latitude of the collection orchard is 41° 07′ 07″ N, and 39° 92′ 27″ E., which represents the Eastern, Western, and Central Black Sea regions of Turkey. Details about origin of hazelnut including regions are given in Table 1. At least six hazelnut bags from each region and a total of 87 samples were collected into plastic bags of one kg which were kept at −20°C to prevent quality loses until analysis.

| Oil extraction
Fifty grams of each sample of hazelnut was manually cracked and shelled, and kernels were grinded finely using a coffee grinder (Waring America, model HGTWTS3). The crude oil was extracted with petro-

| Gas chromatography
The analysis was performed at Agilent 6890N gas chromatography system (GC) (Agilent, Waldbronn, Germany) equipped with flame ionization detector (FID) and JB Scientific DB-23 column (60 m × 0.25 mm × 0.25 μm) (ISO 12966-2:2011(ISO 12966-2: , 2011. Temperature of front inlet and detector were adjusted to 250°C and 260°C, and the flow rate of the carrier gas (nitrogen) was set at 1.4 ml/min. Injection volume was 0.2 μl with a split ratio of 1:40. The column temperature program was performed in four stages. At the first stage, it was held at 130°C for 1 min. At the second stage, the column temperature was increased to 170°C with a rate of 6.5°C/min. At the third stage, the temperature was increased to 215°C with a rate of 2.75°C/ min. At the fourth stage, the temperature was held isothermally for 15 min, and finally increased to 230°C with a rate of 15°C/min and held isothermal for 10 min.

| Statistical analysis
One-way ANOVA and linear discriminant analysis (LDA) were performed using a statistical package program (SPSS ver. 17.0, 2008; International Business Machines Corp., New Orchard Road, Armonk, New York) for p < 0.05 significance level.

| RESULTS AND DISCUSSIONS
The predominant fatty acid was oleic acid ranging between 82.35% and 83.14%, followed by linoleic acid ranging from 7.75% to 9.10%, palmitic acid 5.18% to 5.65%, and stearic acid from 2.44% to 2.59% as shown in Table 2. Myristic acid, margaric acid, and heptadecanoic acid were present in all samples in trace quantities but not reported. The region of the Central Black Sea had highest oleic acid (83.14%) and lowest linoleic acid values (7.75%) competed to Western and Eastern Black Sea regions. Taş and Gökmen (2015) has been reported that Giresun (which is part of Eastern Black Sea) round hazelnuts had 80.1% oleic acid, 10.91% linoleic acid, 5.7% palmitic acid, and 2.4% stearic acid, respectively. Moreover, Bacchetta et al. (2013) examined that major fatty acid of European hazelnuts was oleic acid (80.63%) followed by 10.57% of linoleic acid, 5.95% of palmitic acid, and 2.48% of stearic acid. These findings were slightly in agreement with our results, and the differences may be due to harvesting season, growing conditions, and locations. On the other hand, Parcerisa et al. (1998) Salisbury, and Nicolas (2000) reported that when the amount of oleic acid increased in oily seed plants, the amount of linoleic acid and linolenic acid decreased and similarly when the high air temperatures and spring precipitation raised the amount of oleic acid increased. Parcerisa et al. (1993) also reported that the fruit chemical composition is strongly influenced by environmental and growing conditions.
Total monounsaturated fatty acids (MUFAs) are the major group of hazelnut oil due to oleic acid, while total polyunsaturated fatty acids (PUFAs) and total saturated fatty acids (SFAs) were in lower amounts. These findings were disagreement with our SFAs, MUFAs, and PUFAs values as well. These differences may be due to geographical origin, climate, varieties, fertilizing, and environmental conditions. Cristofori, Ferramondo, Bertazza, and Bignami (2008) reported that the oil of the fruits grown in cold regions contains more unsaturated fatty acids than the fat of the fruits grown in hot and dry regions.
Discriminant analysis performed discrimination among the three regions of Black Sea in Turkey. Discriminant function 1 (explaining 69.3% of the variability) was positively significantly related to palmitoleic acid, linolenic acid, and arachidic acid and negatively related to oleic acid, linoleic acid, and eicosenoic acid. Discriminant function 2 (explaining 30.7% of the variability) was negatively related to oleic acid, linoleic acid, and arachidic acid and positively related to palmitoleic acid, linolenic acid, and eicosenoic acid (see Table 3).
The percentages of predictive probability in the differentiation of origin concerning hazelnut samples for three regions are shown in  (2015), fatty acid profiles could serve as markers of geographical origin of cream products. Diraman, Saygi, and Hisil (2010) found that their prediction model was correctly classified by 84.6% of the samples from five different regions.
D'Archivio, Giannitto, Incani, and Nisi (2014) reported that they predicted correctly classified more than 80% of the 27 Italian saffron spices using mineral composition.

| CONCLUSION
The compositions of fatty acid in hazelnut samples for Black Sea regions in Turkey were investigated in order to detection of geographical origin by ANOVA and linear discriminant analysis, and significant differences were maintained on the fatty acid compositions (C16:0, C16:1, C18:2, C18:3, C20:0, SFAs, and PUFAs