Effect of processing technology on chemical, sensory, and consumers' hedonic rating of seven olive oil varieties

Abstract This study established physicochemical and sensory characteristics of virgin olive oils (VOOs) and linked them to consumers’ liking using external preference mapping. We used five Tunisian and two foreign VOO varieties produced by two processing systems: discontinuous (sp) and continuous three‐phase decanter (3p). The samples were analyzed and evaluated by a panel of 274 consumers. The external preference mapping revealed five VOO clusters with a consumer preference scores rating from 40% to 65%. Consumers highly appreciated the foreign Coratina cultivar's olive oil; the main drivers being richness in polyphenols (markers of bitterness and pungency), mainly the oleuropein aglycone, and volatile compounds (markers of green fruity, green leaves, green apple, cut grassy almond, and bitterness), particularly the trans‐2‐hexenol. The Tunisian Chemlali (3p) oil was second highly preferred (scoring 55%). The positive drivers for olive oil preference (a profile of almond fruity green and low bitterness and pungency) are the richness in hexanal compounds. Arbequina (sp and 3p) and Chemlali (sp) were the least appreciated due to the fact that Arbequina VOO is not in the tradition of Tunisian consumers, whereas Chemchali VOO is a minor variety representing only 2% of olive oil production in Tunisia and consumed mostly in blends. The differentiation between the two processing systems depends on the variety of cultivar; consumers are able to identify the two processing system in the case of Chetoui, Leguim, and Chemchali.

the invention of the centrifugation system (Vaz-Freire et al., 2008).
Currently, the extraction of olive oil using continuously a mechanical system is becoming commonly used; namely the three-phase and the two-phase centrifugation systems (Boudebouz et al., 2021;De Bruno et al., 2020).
Although extraction processing is a crucial factor influencing the chemical composition and the sensorial characteristics of olive oils (El-Riachy et al., 2018), we find only few studies that have investigated the impact of extraction system on the sensory properties of VOO. Since sensory quality plays an important role in consumers' preference, many attempts have been made to clarify the relationships between the sensory attributes in a VOO as perceived by assessors and its volatile and phenol profiles which are responsible for aroma and taste, respectively (Preedy & Watson, 2010).
To be more competitive in international markets, the olive oil industry in Tunisia is tempted to develop a business strategy based on consumers' preferences and orientations. Thus, an essential step consists in identifying the preferences of olive oil consumers and proposing a series of business strategies. The olive oil industry is subject to the advent and impacts of globalization, new commercial structures, technological advances, and consumer demands. Facing stiff competition, firms doing business in international markets need permanent improvement and innovation according to consumer preferences. A key factor for achieving this goal would be to understand consumer behavior and to identify consumer needs and desires (Parras-Rosa et al., 2013).
During the purchase process, consumers establish preferences combining price, quality, country of origin (Dekhili et al., 2011;Mesquita & Andrade, 2014), taste, color, certification, and production method (Chrysochou et al., 2022). Zamuz et al. (2020) studied the effect of sensory and nonsensory factors on purchase intent and consumer choice. In this context, Delgado and Guinard (2011) reported disconnections on consumer behavior of olive oil between experts and ordinary consumers.
Olive oil experts use internal and external quality mapping as an efficient validation tool for the assessment of sensory quality. Such mapping uncovers potential segmentation among the experts and identifies the sensory drivers of their understanding of sensory quality. Understanding the interpretation of extra-virgin olive oil quality is beneficial for producers and traders to commercialize olive oils. It also helps providing a clear methodology for the evaluation and the understanding of consumer preferences and expectancies for virgin olive oil.
Building on authors previous research on consumer preferences (Ben-Hassine et al., 2014), this study investigates the main drivers of consumers' liking and disliking for selected Tunisian and foreign olive oils using external preference mapping techniques. In particular, we test whether consumers are able to differentiate between different VOO cultivars and processing systems (sp and 3p).

| Consumer test
Two hundred and seventy-four consumers were recruited among trainees at the government cooker training center in Tunis-Tunisia.
They were selected according to their consumption frequency and familiarity with EVOOs. Tests were carried out on the total of 14 virgin olive oil (VOO) samples. Each sample was tasted 3 times during eight sessions (30 min per session) organized on 4 weeks in a training center specialized in food agriculture "Brevet Superior Technician." The test was carried out in a specific room under controlled conditions to reduce external influences. The samples were presented at room temperature and served in plastic glasses coded with threedigit numbers. A volume of 20 ml of each sample was served to taste with no obligation to finish the glass. After each test, the order of presentation of the samples was put at random. During the test, each participant evaluated the 14 samples with a 15-min break taken after every 5 samples in order to ovoid fatigue and rinses his mouth using water or a piece of apple between each pair of VOO tastings.
For each product, consumers had to rate their hedonic judgment using the labels "I like the oil" and the other "I don't like the oil," hedonic ratings were then translated into scores ranging from 0 to 9 (Ben-Hassine et al., 2014). The sensory trial was approved by the ethics committee of the National Institute of Nutrition and Food Technology in Tunis-Tunisia. Professional Rancimat, Metrohm SA, Herisau, Switzerland) according to the method described by Tura et al. (2007). This method consists of increasing the oxidation reactions by keeping 3 g of oil at 120°C under a constant air flow of 20 L/hand, then determining the conductivity variation of water (60 ml) due to the increase in oxidative compounds.

| Pigment content
The total chlorophyll and carotenoid compounds (mg/kg) were determined calorimetrically as described by Minguez-Mosquera et al. (1991). Olive oil samples were put into quartz cuvette and absorbance values were taken at 630, 670, and 710 nm against carbon tetrachloride for chlorophyll fraction and at 470 nm for carotenoid fraction.

| Fatty acid composition
The composition of fatty acids was evaluated as the methyl esters of fatty acids (FAME) using a cold saponification according to the method described by IOC (2018). The FAME were prepared by vigorous shaking of an oil solution in hexane (0.1 g in 2 ml) with 0.2 ml of 2 N methanolic potassium hydroxide (KOH) solution and analyzed by GC with a Hewlett-Packard (HP 5890) chromatograph equipped with an FID detector. A fused silica column, HP-Innowax (30 m × 0.25 mm, i.d. 0.25 µm), was used. Nitrogen was employed as a carrier gas, with a flow rate of 1 ml/min. The temperatures of the injector and detector were set at 250 and 270°C respectively.
An injection volume of 1 µl was used. The operating conditions were as follows: oven temperature was held at 180°C for 1 min and then increased by 10°C/min to 220°C, held for 1 min at 220°C, increased again to 240°C at 2°C/min, and finally isotherm at 240°C for 1 min.
Results were expressed as percent of relative area.

| Triacylglycerol composition
Triacylglycerols of olive oils were separated by high-performance liquid chromatography (HPLC) equipped with a reverse phase C18 column (5 mm, 4.60 × 250 mm; Waters Associates). The eluent was monitored by refractive index detector. The mobile phase was acetone:acetonitrile (60:40 v/v) with a flow rate of 1.50 ml/min. All solvents were of HPLC grade. Samples (5 µl) were prepared by dissolving the oil in acetone (9:91 v/v). Peak assignment was carried out by comparison with chromatograms and with the retention times of some pure standards (Ben Hassine et al., 2014).

| Phenolic composition
The polar fraction was extracted by placing 5 g of oil into a 50-ml tube containing 2 ml of hexane, and subsequently, 5 ml of methanol (water 80:20 v/v) were added. The solution was vortexed for 10 min. The emulsion was subjected to centrifugation for 20 min at 5,500g at 4°C to separate the two phases. The alcoholic extract was recovered, and this procedure was repeated three times. Finally, the alcoholic extract was evaporated in cold and reduced pressure conditions and the dried extract was resuspended in 1 ml of 80% methanol as described by (Montedoro et al. 1993).
Results are expressed as mg of hydroxytyrosol/kg of oil. The phenolic identification was performed using the Agilent 1200 Liquid Chromatography System (Agilent Technologies equipped with a standard autosampler and Agilent column Zorbax extended C 18 50 × 2.1 mm, 1.8 μ. The separation was carried out at 30°C with a gradient elution program at a flow rate of 0.4 ml/min. The mobile phases consisted of water plus 0.1% formic acid (A) and acetonitrile (B). The following multistep linear gradient was applied: 0 min, 10% B; 10 min, 25% B; 14 min, 50% B; 20 min, 80% B; 20 min 90% B. The injection volume in the HPLC system was 5 μl. The HPLC system was coupled to an Agilent diode array detector (DAD) (λ detection was 280 and 330 nm) and Agilent 6320 time-of-flight (TOF) mass spectrometer equipped with a dual electrospray interface (ESI) (Agilent Technologies) operating in negative ion mode. Detection was carried out within a mass range of 50-1700 m/z. Accurate mass measurements of each peak from the total ion chromatograms (TICs) were obtained by means of an Isocratic Pump (Agilent G1310B, company) using a dual nebulizer ESI source that introduces a low flow (20 μl/min) of a calibration solution that contains the internal reference masses at m/z 112. 9856, 301.9981, 601.9790, and 1033.9881 in negative ion mode. The accurate mass data of the molecular ions were processed through the software Mass Hunter (Agilent Technologies). The quantification of phenolic compounds was achieved using calibration curves of authentic chemical standards: hydroxytyrosol, oleuropein, pinoresinol, luteolin, and apigenin.

| Volatile compounds analysis
The headspace of 2 ml of olive oil containing into a 5-ml vial was sampled and allowed to equilibrate for 30 min using Supelco SPME devices coated with polydimethylsiloxane (PDMS, 100 µm). After the equilibration time, the fiber was exposed to the headspace for 50 min at room temperature. The fiber was then withdrawn into the needle and subjected to GC-MS analysis. GC-EI/MS analyses were detector. Analytical conditions were as follows: injector and transfer line temperatures were 250 and 240°C, respectively; oven temperature was programmed from 60°C to 240°C at 3°C/min; helium was used as carrier gas at a flow rate of 1 ml/min. The identification of the volatile compounds was based on the comparison of the retention times with those of authentic standards, comparing their linear retention indices (LRI) relative to a series of n-hydrocarbons, and on computer matching against commercial (NIST 98 and Adams) and homemade library mass spectra, built from pure substances, components of known oils, and MS literature data (Adams 1995).
Moreover, the molecular weights of all the substances identified were confirmed by GC-CI/MS, using methanol as the ionizing gas.

| Data analysis
In order to study the effect of the product factor on all physicochemical parameters, one-way analysis of variance (ANOVA, Tukey's honest significant difference multiple comparison) was carried out using the package agricolae (version 1.1) in the R software (ver- to classify the products. Then, a preference mapping was performed using SensoMineR package (version 1.17) and according to the method of Danzart et al. (2004). A response surface is computed per consumer; then according to certain threshold, preference zones are delimited and finally superimposed.

| RE SULTS AND D ISCUSS I ON
3.1 | Influence of cultivar and extraction process on quality indices, pigments, volatile and phenolic compounds, and saponifiable fraction of VOOs 3.1.1 | Free acidity, absorbances in the UV, and peroxide value As shown in Table 1, all the olive oil samples exhibited quality parameters within the range allowed by the regulation EC2568/91 for the extra-virgin olive oil category (free acidity ≤0.8%; peroxide value ≤20 meq O 2 /kg; K270 ≤0.22; K232 ≤2.5) except the variety Leguim.
In fact, this oil variety obtained by press system exceeded the limits established for "extra-virgin olive oil" in free acidity (1.42%). For this reason, it could not be labeled as "extra-virgin" according to the European Union regulations (EEC, 2003).
Acidity values of the studied oil samples ranged from 0.16% for the Coratina oil variety obtained by the three-phase decanter (Coratina 3ph) to 1.42% for Leguim oil one obtained by press system (Leguim sp). It is worth noting that in most cases, the free acidity of oils obtained by the sp system was higher than oil samples extracted by the 3ph decanter. Such a result is consistent with previous study conducted by Ben-Hassine et al. (2013) on two olive varieties "Chemlali" and "Coratina" extracted by super press, dual and triple phase decanter. Besides the extraction process, it is important to mention that free acidity of oils is highly influenced by other factors mainly storage conditions and time (Ghanbari Shendi et al., 2018).
The free acidity of olive oil corresponds to the proportion of fatty acids found in the free state as a result of the lipolytic action of intrinsic or extrinsic lipases. It reflects the degree of stability of the oil and its susceptibility to rancidity (Khlil et al., 2017). As reported in literature, during the SP process, olive oil is extracted with the vegetable water (aqueous phase plus solid wastes) and they remain together until they are separated by decanting, which may favor the hydrolysis of triglycerides, resulting in an increase of free fatty acids level (Torres & Maestri, 2006a, 2006b.

| Oxidative stability
Stability to oxidation is an important property for olive oil, which is improved by synergistic interactions between the lipid composition and intrinsic antioxidants. According to Najafi et al. (2015), the oxidative stability of virgin olive oil is influenced by SFA/UFA ratio and tocopherolic compounds. It is negatively affected by their fatty acid composition and minor components such as tocopherols, phytosterols, vitamin E, phenolic compounds, enzymes, and trace metals (Ghanbari Shendi et al., 2018;Gómez-Alonso et al., 2003).
Phenolic compounds reputed for their antioxidant properties play a key role for the stabilization of unsaturated fatty acids (UFA) (Miho et al., 2020).
Our results showed that cultivar and extraction system have both significant effect on oxidative stability. Among samples, Coratina olive oils obtained by both systems "sp" and "3ph" (13.41 and 15.37h, respectively) were the most stable to oxidation (Table 1). It is also important to note that olive oils obtained by the three phase system were more stable than those obtained by press system in this study.
VOO are known to be more resistant to oxidation than other edible oils, thanks to their content of natural antioxidants and lower unsaturation levels (Torres & Maestri, 2006a, 2006b. A recent study conducted by Jaber Houshia et al. (2019) has shown that the relative phenolic profile highly explained the VOO oxidative stability.
Their preliminary study revealed that it is possible to predict VOO oxidative stability with a regression model based on hydroxytyrosol, aldehydic open forms of oleuropein aglycone, and linoleic acid as explanatory variables.

| Pigment content
In addition to their antioxidant capacities, pigments are responsible for the color of olive oil, which is one of the factors that influence consumers thoughts and is considered as a quality parameter. Chlorophylls are responsible for the greenish color of olive oils, whereas the yellow color is due to carotenes (Psomiadou & Tsimidou, 2001). The pigment profile of olives is mainly affected by the variety (or cultivar) (Aparicio-Ruiz et al., 2009;Lazzerini & Domenici, 2017), the ripening degree (Criado et al., 2007;Ranalli et al., 2005), and the edaphoclimatic and agronomic conditions (Jolayemi et al., 2016). Moreover, the conditions of olive oil production mostly malaxation stage and oil extraction olive oil pigment profile further influence the final content and percentage of pigments in olive oil (Ruiz-Domínguez et al., 2013;Vaz-Freire et al., 2008). Chlorophyll content varied from 2.20 (Zalmati 3ph) to 7.15 ppm (Arbequina 3ph). In the studied samples, β-carotene concentration varied significantly and ranged from 0.69 ppm (Zalmati 3ph) to 5.44 ppm (Leguim sp). The extraction system and cultivar had significant effect on chlorophyll and β-carotene amounts for the majority of our samples. However, the extraction system did not show any significant effect neither on chlorophyll levels in Chemchali and Leguim oils nor on β-carotene levels for Arbequina, Chetoui, and Leguim oils (Table 1).
Regarding triacylglycerols, significant differences were noticed among the analyzed samples (Table 2). A significant effect of the extraction system on some triacylglycerols was reported in literature (Kelebek et al., 2015). Cultivar also showed a significant effect on the triacylglycerol composition of olive oil .
Among the main triacylglycerols (LOO, LOP, OOO), the percentage of OOO was the highest ( These results are in accordance with those reported in literature . and Chetoui, the total phenol content was higher for oils extracted by the press system in contrast to the rest of cultivars (Arbequina, Chemlali, and Chemchali). In press extraction process, the amount of added water is minimal when compared with the continuous sys- were the major compounds (Table 3). Oleuropein and ligstroside aglycones' concentrations varied largely among samples according to the extraction process and cultivar, being the highest levels registered for "Coratina sp" and "Chetoui sp" oils, respectively (Table 3).
Hydroxytyrosol concentrations varied significantly according to the extraction system where the highest ones were registered for threephase decanter in the case of Chemchali, Chemlali, and Coratina VOOs. The same fact was noticed for tyrosol whose concentrations were higher in Arbequina, Coratina, and Chemchali oils from threephase decanter processing than in those with the super press. Like in the case of hydroxytyrosol, pinoresinol concentrations showed significant variation among samples extracted with pressure and centrifugation regardless the cultivar. It is important to mention that oils produced from three-phase processing were richer in these phe-  demonstrated that phenolic compound concentrations significantly decline after a year storage time.
Although a full description of the organoleptic characteristics of the oil is only obtainable through sensory analysis, the qualiquantitative determination of the volatile compounds can provide very useful information on product quality (Angerosa et al., 2004).
The main volatiles contributing to the peculiar aroma VOO are C6 and C5 biogenerated through the lipoxygenase pathway during oil production (Brkić Bubola et al., 2012). The concentration and activity of enzymes involved in this biogenesis are influenced by several agronomical and technological factors, among them are cultivar (Runcio et al., 2008), stage of olive ripeness, and production conditions (Olias et al., 1993).
In our study, several compounds belonging to different chemical classes (carbonyl: aldehydes and ketones, alcohols, esters, hydrocarbons, some acids, and furane derivatives) were detected (Table 4).
The chemical composition of all tested olive oils showed that C 6 compounds (trans-2-hexenal, cis-3-hexenal, hexanol, and cis-3-hexenol) were the most abundant compounds. For the monovarietal olive oils involved in this study, trans-2-hexenal was the major C 6 aldehyde volatile in Coratina olive oils (372.41 ppm). However, olive oils obtained from Leguim variety contained the lowest amount (1.24 ppm).
The C 6 aldehydes hexanal and trans-2-hexenal, as well as hexanol, contribute to the typical green sensory perception. Produced via the LOX pathway from polyunsaturated fatty acids (linolenic and linoleic acids), hexanal and trans-2-hexenal accumulate in virgin olive oils during physical extraction procedures (Iraqi et al., 2005). The latter is derived from cis-3-hexenal, which undergoes isomerization to a more stable compound that can then be further reduced to trans-2-hexen-1-ol (Luna et al., 2006). Hexyl and trans-3-hexenyl acetate were present in the aroma of all our samples, but at different levels. As shown in Table 4, the chemical composition of the volatile fraction of studied olive oil samples was variable, depending on both extraction system and cultivar. Regarding the major volatiles, their concentration varied significantly according to the extraction system. In fact, oils obtained by three-phase centrifugation were the richest in cis-3-hexenol for the majority of cultivars and in hexanol for all cultivars, which emphasizes the perception of green, fruity, astringent, bitter, and pungent (Ben Lawlor et al., 2001;Brkić Bubola et al., 2012). Meanwhile oils produced by pressure process were richer in trans-2-hexenal for the majority of cultivars.

| Influence of the cultivar and extraction process on the sensory attributes
The oils extracted from Chetoui and Arbequina cultivars by continuous processing system had a different aroma profile in comparison to those extracted by press system from Chemlali, Chetoui, and Leguim cultivars (Table 1).
The study of the effect of extraction system on our olive oils sensorial profile revealed a significant difference between oils obtained by three-phase system and those extracted by press system. Oils obtained by centrifugation system were characterized by higher sensory scores for bitterness, fruitiness, and pungency than those obtained by press, except concerning Zalmati, Chemchali, and Arbequina VOOs.  was reported that its intensity was higher in two-phase than in threephase decanter extracted oils (Clodoveo, 2012). In our study, the ANOVA test showed that the sensory profile of the present VOOs was more influenced by the extraction processing rather than the cultivar.  and high antioxidant activities (Guerfel et al., 2009). VOO with a high intensity of bitterness, astringency, or pungency are hardly marketable in emergent markets. Consequently, blends should be made between such oils and nonbitter VOO. It is worthy to notice that the sensory attributes of EVOO are mainly correlated to the content of minor components such as phenolic and volatile compounds (Dabbou et al., 2010).

| Principal component analysis (PCA)
According to the Principal component analysis (PCA) (Figure 1a), Samples of group 4 could not be well separated within PC1 and PC2.

| Preference mapping
To explore the consumer preferences toward the VOOs under study according to processing system and cultivar, a preference mapping approach based on PCA was used. External preference mapping is a very useful statistical technique which covers and measures the positive and negative drivers of olive oil sensory and chemical quality as perceived by consumer. On the map, the distance between each VOO product and the optimum (70%) was shown ( Figure 2). Five main groups of VOOs could be observed on the map with a good segmentation according to the consumer preference score and chemical parameters. The first group is composed by the foreign olive oil Coratina (3p and sp) which was the most preferred one among the tested samples with 65% of consumer preference being the nearest to the optimum product preference (70%). The second group is also highly appreciated by consumers and presented the Tunisian olive oils Chemlali 3p, Leguim 3p, and Chetoui sp that attracted 50%-55% of the consumer appreciation (Figure 2). The third group composed of Chetoui 3p, Chemchali (sp and 3p), Zalmati 3p, Zalmati sp, and Leguim sp olive oils that attracted 45%-50% of consumer appreciation. The least appreciated group and the farthest from the optimum product is composed of Arbequina (sp and 3p) and Chemlali sp and was appreciated by 40%-45% of consumer choice. It is noteworthy to say that consumers could not distinguish between the two extractions systems (3p and sp).
Based on the PCA, the positive drivers of consumer choice for Coratina olive oil cultivar were its richness in total polyphenols (markers of bitterness and pungency), Ac-pin, LA, major TAGs (OOO, AOL, SOO), β-carotene, C18:1 (marker of freshness), and volatile compound cis-3-hexenol, 3-hexenyl acetate, markers of green fruity, green leaves, green apple, cut grassy almond, and bitter perception.
The positive drivers for the olive oil Chemlali (3p) (52% of consumer appreciation) known by a profile of almond fruity green and low bitterness and pungency were its richness in hexanal related to the sensory perception of green fruity, green leaves, floral bitter, cut grassy, ripe apple, trans-2-hexenol, 1-pentanol, TAG (LnLo, LOO, LLO, LLL, LnLP, LnOO, LOP, SOO), and C16:1. The Chetoui cultivar attracted consumers for its richness in H-Tyr, DFLA (related to the sensory perception of astringency, bitterness, and pungency), and in LOOO and LLL.
This study allowed us to distinguish that the foreign introduced variety Arbequina produced by sp or 3p is not in the tradition for Tunisian consumers. Regarding Chemchali sp VOO, the nonappreciation can be explained by the fact that this is a minor variety of olive oil as it represents only 2% of olive oil Tunisian production. It is not available in the market as a monovariable olive oil; it is mainly consumed in blends.
Preference mapping has been used extensively to describe the characteristics that contribute to consumer's acceptance or rejection by the identification of both positive and negative drivers of liking (Delgado & Guinard, 2011). Concerning the sensorial attributes, researches revealed that the most valued attributes from the consumer viewpoints are the "ripe fruity" and "sweet" (Valli et al., 2014), high intensity for color, odor, taste, and flavor, and pungent and floral series (Zamuz et al., 2020). However, the bitter (Zamuz et al., 2020) and pungent positive sensorial attributes are rejected by the consumer (Delgado & Guinard, 2011). This rejection could be related to the unfamiliarity of consumers (Valli et al., 2014) with these attributes and their lack of knowledge concerning the relation among bitter and pungent attributes and nutritional and healthy properties of olive oils (Zamuz et al., 2020). In Tunisia, researches conducted by Mtimet et al. (2013) concluded that Tunisian consumers have a good knowledge of the characteristics of olive oil.

| CON CLUS ION
The consumer preference evaluation for the studied Tunisian and foreign VOOs showed that liking was related to chemical and sensory profiles and varied among consumers as revealed by the preference mapping. The study identifies the key liking drivers, such as polyphenols, oleic acid, and good aroma compounds namely cis-3hexenol and 3-hexenyl acetate. Among studied samples, Coratina VOOs were the most appreciated followed by Chemlali. In general, consumers appreciated the fruity attribute and, in part, the pungent sensation, whereas they recognize bitterness as a negative attribute.
Processing systems are found to be the main drivers of consumers' choice in favor of the discontinuous system (sp). We believe that these findings ought to have been supplemented with an evaluation of foreign consumers' preference in order to accommodate their need according to their chemical and sensory olive oil profile, since Tunisian olive oils are widely exported around the world.
Furthermore, it could be useful if we explore preference evaluation according to consumer characteristics (gender, age, etc.). This remains a subject for future investigation.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

ACK N OWLED G EM ENTS
Required.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data are available on request by the first author Dr. Kaouther Ben-Hassine (kaoutheragro@yahoo.fr).