Effect of spices formulations on the physicochemical and sensory properties of Nnam gon, a Cameroonian dish prepared with cucurbitaceae seeds

Abstract Nnam gon, a cake made by steam cooking a mixture of Cucurbitaceae seeds paste and others ingredients specially spices, is a highly prized dish in Central African region. A preliminary investigation conducted as part of this study highlighted that formulations used in the processing of Nnam gon vary according to the spices used. This study was carried out to determine the best formulation for the preparation of this dish. For this purpose, Nnam gon samples were produced from four formulations which differ according to the number of spices used: F0 (no spices); F1 (0.91 g of Allium cepa paste); F2 (0.88 g of A. cepa paste, 0.35 g of Allium sativum paste, 0.41 g of fresh Officinale zingiber paste, 0.41 g of fresh Petroselinum crispum paste, 0.33 g of Monodora myristica, 0.48 g of fresh Celery graveolens paste, 1.19 g of fresh Allium porrum paste, 0.13 g of Allium lepidophyllus powder, and 0.13 g of Piper nigrum), and F3 (0.90 g of A. cepa paste, 0.35 g of A. sativum paste, and 0.42 g of fresh O. zingiber paste). The samples were evaluated for their physicochemical characteristics and sensory profile (Quantitative Descriptive Analysis). The results revealed that proteins (16.56–17.38%), carbohydrates (4.71–5.10%), lipids (23.14–24.25%), ash (4.03–5.92%), and fibers (2.17–2.68%) increased significantly (p < .05) with spices adding. The increase in polyphenols (310.55–592.80 mg/100 g FM) and phytates (2.23–12.49 mg/100 g FM) contents was positively correlated with antioxidant properties of Nnam gon which also increased with spices adding. Significant differences were observed between the samples for all attributes generated (appearance, odor, taste, flavor, texture, and oral texture). Spices adding induced a decrease in hardness, cohesivity, elasticity, and granulous of cake but enhanced oily. Nnam gon produced with spicy formulation (F2 and F3) had higher mean score for general acceptance which was highly correlated (p < .05) with spice odor (r = .99), spice taste (r = .92), and color (r = .84). From this study, it is suggested that spicy Cucurbitaceae paste could improve nutritional value, antioxidant properties, and general acceptance of Nnam gon.

Despite this high nutritional potential, the cultivation and use of cucurbits seeds remain limited in food industry. In fact, the world production of these seeds was estimated at 184 million tons per year (FAO, 2014), but in Cameroon the annual production is around 146,000 tons (FAO, 2010). Cameroon production is still very low compared to other countries such as South Africa (378.776 tons), Egypt (690.000 tons), and China (5.767.700 tons) (FAO, 2014). An average cost of kilogram of dry Cucurbitaceae seeds is 2.29 euros in Cameroon, and that price is one and half times the average price of kilogram of cocoa and half times that of coffee which are the main industrial crop in that country (Minader, 2012). Then, the sector of cucurbit seeds could be an important activity for agriculture in Cameroon. One way to boost the underutilized seeds production could be to increase demand for these seeds by using them as raw materials in industry or by industrializing the production of foods made from these seeds (Ebert, 2014). In this respect, improving the production of Cucurbitaceae cake could be a solution to stimulate the production of cucurbit seeds.
In fact, Cucurbitaceae cake, called Nnam gon in Cameroon, is a highly prized traditional dish formulated with Cucurbitaceae paste mixed with water and other ingredients such as salt, fresh eggs, oil, fish, or meat. Spices (Allium cepa, Allium sativum, Officinale zingiber, Petroselinum crispum, Monodora myristica, Celery graveolens, etc.) are also used to season the paste. The seasoned paste is packed in the leaves of katemfe (Thaumatococcus daniellii) and steam cooked (Ponka et al., 2005). This cake is prepared in all regions of Cameroon and has an important value in traditional societies of this country. It is usually offered at ceremonies like wedding and funeral (Ponka et al., 2005).
Moreover, Nnam gon is a street food whose production is mainly ensured by women, but its consumption in the street is often linked to some digestive disorders that could probably be due to poor conditions of cooking and storage. Few studies have already been done on Cucurbitaceae cake, nevertheless the study of the nutritional value of this food realized by Ponka et al. (2005) noted that it contains around 8.96% of proteins, 13.5% of lipids, 1.86 of fibers, 1.77% of ash, and minerals such as Magnesium (108.9 mg/100 g DW), Iron (2.99 mg/100 g DW), and Zinc (3.29 mg/100 g DW).
The valorization of Cucurbitaceae seeds by industrial production of Nnam gon requires improving the production process, packaging, and storage conditions. In this respect, the preinvestigation conducted as part of this study has identified that the ingredients used for the formulation of Nnam gon vary according to household in Cameroon. Thus, it would be interesting to determine the formulation mostly accepted by the consumers. For that, chemical composition, sensory characteristics, and general acceptability of different formulated cake must be studied. For this purpose, the Quantitative Descriptive Analysis (QDA) methodology which is one of the most used descriptive approaches is recommended for sensory analysis of Nnam gon.
As part of main study aimed on improving the quality of this dish, this study was carried out to determine the chemical composition of Nnam gon produced with different formulations, their sensory characteristics, and overall acceptability.

| Sampling of Cucurbitaceae seeds
Dried samples of Cucurbitaceae seeds (Cucumeropsis mannii) harvested on February 2015 were purchased on March 2015 from local market in Ngaoundere, Cameroon. They were sorted manually to remove foreign matter and immature and damaged seeds. Then, the seeds were soaked for 5 min in order to facilitate dehulling which was manual. The dehulled seeds were milled using an electric grinder. Then, the flour obtained was directly used for production of Nnam gon samples.

| Production of Nnam gon samples
The Cucurbitaceae seeds flour was divided into four subsamples corresponding to the different formulations (F0, F1, F2, and F3). The formulations (Table 1) and Nnam gon production process (Figure 1) are the results of a preliminary investigation prior to this study. As reported in Figure 1, flour and water were introduced in a large plastic bowl and mixed with a wooden spatula during 10 min. Then, salt, spices, and other ingredients were added to the mixture which was remixed for 5 min. About 300 g of the final mixture was packaged in cleaned leaves of katemfe (Thaumatococcus daniellii) which are traditionally used for the packaging of Nnam gon. Packaged samples were filled into aluminum pot (10 L of capacity, with tripod) containing 1.5 L of tap water. Five samples were deposited on the tripod and were not in direct contact with boiling water. Cooking was carried out for 90 min on hearth of three stones fed with wood. The samples of the different formula were cooked separately. At the end of cooking, the samples were removed from the pot and cooled on the bench. The cakes were separately processed into three subsamples. One subsample was used for chemical analysis, another subsample for textural analysis, and the last was used for sensorial analysis.

| Proximate analysis of Nnam gon samples
Moisture, crude protein, crude fat, and crude ash contents were determined according to the AOAC procedures (1990). Protein was calculated as N × 6.25. Total carbohydrate was determined after digestion in concentrated sulfuric acid (Dubois, Gilles, Hamilton, Rebers, & Smith, 1956).

| Analysis of bioactive and antinutritional factors
Polyphenol content was determined using a colorimetric method as described earlier by Mang et al. (2015). Tannin level in the cakes was determined by the colorimetric method of Makkar, Siddhuraju, and Becker (2003). The phytic acid content was evaluated using a colorimetric method as reported by Gao et al. (2007). The oxalate content was determined by a titrimetric method according to AOAC procedures (AOAC 1990).

| Antioxidant properties analysis
The reducing power of cakes was measured according to the method described by Oyaizu (1986). Evaluation of DPPH Free radical scavenging activity was determined following De Ancos, Sgroppo, Plaza, and Cano (2002) with some modifications. Chelating power of cakes was measured according to the method described by Decker and Welch (1990).

| Textural analyses
For the texture evaluation, the freshly cooked samples of cakes were equilibrated at ambient temperature for 5 min and submitted to a compression test as described by Nourian, Ramaswamy, and Kushalappa (2003) using a computer interfaced universal testing machine (Lloyd Model LRX-2500N) equipped with a 500 N load cell.
One measurement was made per cake, three cakes were tested per formulation, and their average values taken to represent the mean texture value of test samples. From the generated texture profile, the hardness was obtained from the peak force of the first compression. Adhesiveness was obtained from the final force of the first compression. Viscoelasticity index was the ratio between the peak force of the first compression and the peak force of the second compression.

| Preselection of panelists
Twenty panelists (10 women and 10 men) drawn from a subuniversity population were preselected for the descriptive sensory evaluation, based on their capacity to detect sensory differences in this kind of product.

| Development of sensory descriptors
The judges previously selected, after four meetings, have developed sensory descriptors used in the definitive tests. For this, Repertory Grid Kelly's Method (Moskovitz, 1983) was used. The four formulated cakes (F0, F1, F2, and F3) were evaluated using a sensory evaluation form constituted of nonstructured 20 cm scales for each sensory descriptor. From a common agreement, the judges chose reference materials to determine the extreme points of the intensity scales and to help in the identification of sensory characteristics of the samples.

| Selection and training of panel
Previously judges were trained following QDA (Stone & Sidel, 1993) during three sessions with a view to select the definitive panel. The final panelists were selected based on their ability to discriminate different samples and the repeatability of their results as reported by Bannwart, Bolini, Toledo, Kohn, and Cantanhede (2008).

| Sensory evaluation
The selected and trained judges participated in the sensory evaluation of Nnam gon. The cake samples were evaluated in individual booths under white light and provided with room temperature, water, and unsalted crackers. One cake from each formulation was used for sensory analysis. The products (5 g each) were presented in glass plate coded with three digit random numbers. All the products were assessed in four random repetitions.

| Statistical analysis
All the results were carried out in triplicate determinations. Analysis of variances was used to determine the effect of formulation on the dependent variables. Duncan's Multiple Range Test was performed to classify samples at the significant level of 5%. Statgraphics 6.0 Program was applied for the statistical analysis. Principal component analysis (PCA) and Pearson analysis were applied using Minitab Statistical Software. Table 2 presents the proximate composition of Nnam gon samples.

| Proximate composition of Nnam gon
The moisture content in these cakes ranged around the average value of 46%. The lack of significant difference observed between moisture content of the different formulated cakes suggested no significant effect of spices on moisture content of Nnam gon. But the water content observed was lower compared to those of Ekomba (61.1%) and Ekwang (77.6%) which are others Cameroonian cakes produced with maize flour and Cocoyam flour, respectively (Ponka, Fokou, Beaucher, Piot, & Gaucheron, 2016).
The ash content of cakes was in the range 4.03-5.92% of DM.
This is significantly higher than ash content of Akara (3.1%), a traditional bean cake prepared in Nigeria (Okeke & Eze, 2006 Lipids are the main components of Nnam gon. The fat content in the samples ranged from 23.14 to 24.25%. These values were higher than fat content of 2.7% in Jeqe (cake made with white flour) consumed in rural KwaZulu-Natal, South Africa (Spearing et al., 2013).
Adding spice had no effect on lipid content of these cakes, it would probably be linked to their low-fat content. But addition of soy oil in the formulation also contributes to increase oil content of the cakes.
After lipids, proteins are the third major constituents of Nnam gon. The highest crude protein content was recorded for the cake from F2 formulation (17.38%). Apart from control cake (F0) which had a lower crude protein (16.56%), no significant difference was observed between crude protein contents of others formulated cakes, but spices induced an increasing in crude protein of cake probably linked to their protein content. These protein contents are close to that of moinmoin (17.71%) (Olapade & Adetuyi, 2007) indicating that Nnam gon is a good source of proteins.  doubled phenolic content of cake. In addition to the foregoing ingredients, adding Garlic and ginger pastes (F3) also slightly increased phenolic content of Nnam gon. We noted that phenolic content of the cake increased with the number of spices introduced in the cake formulation. Studies have shown that spices such as disc, leek, and celery are rich in phenolic compounds (Belewu, Olatunde, & Giwa, 2009;Uhegbu et al., 2011). In this vein, the phenolic content of cakes produced with F2 formulation was significantly higher (p < .05) than that of others formulations, probably due to the several spices of this formulation.

| Antinutrient content of Nnam gon
Comparatively, adding spices in formulation also induces an increase in the oxalate content of Nnam gon which ranged between 1.41 and 3.30 mg/100 g DM. The oxalate content also appears to be related to the number of spices introduced in the formulation, thus, F2 presented a high content of oxalate. In fact, some health problems (corrosive gastroenteritis, shock convulsive symptoms, low plasma calcium, high plasma oxalates, and renal damage) are caused by a consumption of high levels of oxalates (Kelsay, 1985). Specifically, oxalate content increasing would be linked to the addition of garlic (2.6 mg/g), ginger (0.7 mg/g), and onion (0.3 mg/g) which have high oxalate content compared with other spices according to the literature (Nwinuka et al., 2005;Belewu et al., 2009;Oluwatoyin, 2014 (2014) and Nwinuka et al. (2005). Similar increase in tannins levels was observed with addition of spice in Nnam gon formulation. Cake produced from formula 2 had a high amount of tannins (527.44 mg/100 g DM), but low level was observed for control cake (131.59 mg/100 g DM).

| Antioxidant properties of Nnam gon
It has already been established that the consumption of foods rich in antioxidants helps to prevent the diseases related to oxidative stress (Boskou, 2006;Saikat, Raja, Sridhar, Reddy, & Biplab, 2010).
In general, as shown in Figure   was comprised between 0.67 and 5.84 g Vit C/100 g DM. The strong correlations observed between phenolic compounds and DPPH free radical scavenging (r = .99; p < .05), reducing power (r = .99; p < .05) and chelating power (r = .97; p < .05) indicate that these compounds whose antioxidant activity is well established contribute to raising the antioxidant properties of cakes which are enhanced by spices.

| Textural properties of Nnam gon
Texture is the most important quality attributes that affect consumer acceptability of cake (Aboua, Konan, Kossa, Agro, & Kamenan, 1989;Aboubacar, Yacizi, & Hamaker, 2006). According to the results presented in Table 4, textural parameters of cakes varied significantly with the formulations. Irrespective of the probe used, Formulation

| Sensory characteristics of Nnam gon
The candidates used for sensory analysis were selected based on their ability to discriminate samples. As presented in One of the most important characteristics of good food is its odor.
In this respect, general acceptance of Nnam gon was highly correlated with spice odor (r = .99; p < .05). In this study, cake produced with formula 2 and 3 presented the highest means score of spices odor but lower means Cucurbitaceae seeds odor. It could probably be due to the fact that spices odor mask that of Cucurbitaceae seeds odor, hence the negative correlation observed between these attributes (r = −.79; p < .05). However, formula 1 cake had higher mean score of fish and shrimp's odor, whereas musty odor was observed on control cake F0. Taste and Cucurbitaceae seeds flavor were highly represented in control cake, but addition of spices in cake formulation seems to reduce this attributes. As observed in Table 6, cakes produced with formula 2 and 3 presented the higher mean score of taste of spice. The attributes bitter and bitter after taste were not related for all the samples.
Regarding texture, the sample produced with formula 1 presented the higher mean score of hardness. Spices adding appears to induce a decrease in hardness, cohesivity, elasticity, and granulous of cake, but it seems to enhance the oiliness. This previous observation could mean that spices increase oil absorption capacity of Cucurbitaceae paste. Another parameter that could affect the texture of cake is the moisture content; however, it was not correlated with any texture attribute. Moreover, not only the spicy of paste seems to increase juiciness of cake but it also induced a decrease in cake masticability.
In order to study the relations between sensory attribute and  F0) was characterized by the attributes musty odor, Cucurbitaceae seeds taste and flavor, and persistence of Cucurbitaceae seeds taste, whereas F1 formulated cake was highly correlated with dried shrimps flavor and odor. The attributes that characterized cake produced with formula 2 were color, spice odor, and spice taste. However, formula 3 cake was characterized by fresh fish odor, wrapping leaves taste, and odor. It is interesting to note that the samples were also somewhat separated in terms of spice adding. Then, as seen in Figure 4, cakes produced with spicy Cucurbitaceae paste (formula 2 and 3) had higher mean score for general acceptance which was highly correlated with spice odor (r = .99; p < .05), spice taste (r = .92; p < .05), and color (r = .84; p < .05).

| CONCLUSION
This study aimed at evaluating the effect of formulation on proximate composition, and textural and sensory properties of Nnam gon.
The overall differences observed among the samples evaluated are directly related to spices adding in the formulations. Then, the study revealed that proteins, carbohydrate, lipid, ash, and fibers contents of Nnam gon increased significantly with spices adding. In addition, the polyphenols and phytates contents of cake increase and are positively correlated with antioxidant properties of cake which also increases with spices adding. The general acceptance of cake is highly correlated with spice odor, spice taste, and color. Cake produced with spicy formulation had higher mean score for general acceptance. However, several species of Cucurbitaceae are edible. Thus, it would be interesting to include this aspect in order to know the species of Cucurbitaceae seeds that are more accepted by consumers for the production of this dish. In this respect, further investigation could be done to determine the influence of species of Cucurbitaceae seeds on proximate composition, and textural and sensory properties of Nnam gon.