Quality of pasta made of cassava, peach palm and golden linseed flours

lhmeller@ufpa.br Abstract The possibility of industrially producing new types of pastas based on other types of flour than wheat has raised interest for enabling the use of widely available materials and providing adequate products to persons with celiac disease. A mixture design was developed with the purpose of preparing a gluten-free dry using cassava starch, peach palm meal, and golden linseed while observing the nutritional value added and its technological properties. The pasta underwent assessments of its physicochemical and rheological characteristics. The drying temperatures studied were 60, 75, and 90 °C and the influence of those temperatures on pasta quality parameters and texture properties assessed. The percentage composition of the flours met the standards mandated by the Brazilian legislation and the pasta was classified as rich in fibers. The texture parameters of the pasta were influenced by the drying process, with the best results obtained at the lowest and highest temperatures


Introduction
The possibility of industrially producing new types of pasta with no wheat flour has raised interest not only for the cost of raw material, but also for allowing widely available materials with better nutritional characteristics to be used (Silva et al., 2009).
Several studies have been conducted to find alternative sources to produce pasta by replacing the wheat (Chillo et al., 2010;Fiorda et al., 2013). Such studies are important as they offer dietary options to gluten-intolerant individuals (Thompson et al., 2005). Flours with gluten must be replaced to allow persons intolerant to that protein to consume a broader range of foods. The production of pasta based on flours with no wheat may provide this consumer group sensorily satisfactory products with adequate nutritional levels (Dnyango et al., 2009).
Quality pasta can be obtained by adding highly nutritious raw materials. Such enrichment mainly aims at increasing the nutritional quality of the pastas regarding the contents of proteins, dietary fibers, vitamins, and minerals (Chillo et al., 2010;Marconi & Carcea, 2001).
The fruit of the peach palm tree has fleshy, thick, and fibrous pulp. The fruits are consumed cooked and can be raw material for the extraction of oil or for flour production for various purposes (Clement, 2000). The potential for using peach palm as raw material to produce flour lies in its high nutritional value and composition since it carries all essential amino acids, has a significant dietary fiber content, and its main fatty acids are oleic and palmitic. On addition, it has high concentration of carotenoids, a nutritionally very relevant component (Yuyama et al., 2003).
Cassava starch has physicochemical characteristics of great industrial interest, such as the high peak viscosity and low setback, forms a clear and high transparent paste, has a light flavor, and resistance to retrogradation. Ot has wide application in several industrial sectors, such as the food industry (Rolland-Sabaté et al., 2013).
Linseed is used to prepare food products for its water holding capacity, total dietary fiber content, and, in particular, as a replacement for fat. Besides its technological role, the use of this fiber as an ingredient favors the development of foods with specific healthful effects (Jenkins et al., 1999;Tarpila et al., 2005;Chen et al., 2006).
On the overall quality assessment of pasta, texture properties are the most important since they are directly related to consumer acceptance. As a whole, desirable pasta has al dente characteristics (moderate firmness and high elasticity), no tacky surface, good integrity, brilliant color, and attractive flavor (Mariotti et al., 2011;Pagani et al., 2007). Those characteristics are directly impacted by drying conditions (Marchylo & Dexter, 2004).
The objective of this study was to verify the viability of using cassava starch, peach palm flour, and golden linseed to produce gluten-free pasta and to evaluate the combined effect of the component blend on technological characteristics.

Raw materials
The experiment was developed at the Laboratory of Physical Measurements (LAMEFO) of the Federal University of Pará (UFPA) in the city of Belém, PA, Brazil. The peach palm (Bactris gasipaes Kunth) fruits used were purchased at the Ver-o-Peso market in Belém. The cassava starch and golden linseed flour were purchased on the local market.

Obtaining peach palm flour
On order to produce peach palm flour, the skins and pits were removed and the mesocarp was cut into cubes approximately 2 cm thick, which were dried in a forced circulation oven at 55 °C for approximately 48 h until they reached 12% moisture. After drying, the fruits were ground in a hammer mill and sieved (60 mesh) to homogenize the granulometry.

Peach palm flour characterization
Dnly the peach palm flour was characterized, whereas the information contained in the packaging was used for the other raw materials. Association of Dfficial Analytical Chemist (1997) methods were used to quantify the contents of moisture (method no. 925.10), ash (method no. 923.03), crude protein (method no. 960.52), total lipids (method no. 920.39C), and total, soluble, and insoluble dietary fiber (enzymatic-gravimetric method no. 985.29). Carbohydrate content was determined by difference. Total carotenoids were determined using the extraction method by Godoy & Rodriguez-Amaya (1994). pH was determined using method no. 981.12 of the ADAC methodology (Association of Dfficial Analytical Chemist, 1997). Water activity (aw) was measured using a water activity meter (Aqualab 4TEv). All analyses were carried out in triplicate.

Mixture design of experiment
Formulations were prepared to produce pastas with different contents of cassava starch (CS), peach palm flour (PPF), and golden linseed flour (GLF). The concentrations of the components varied according to the experimental mixing scheme using lower and upper limits established after preliminary tests. On order to better visualize the effects of the independent variables (CS, PPF, and GLF) in the studied region, the actual concentrations of the components were transformed into another set of components, expressed in pseudocomponents, assuming values from 0 to 1 are presented in Table 1. The pseudocomponents (Xi) were calculated from the original concentrations (ci) by Xi = (ci-Li)/(1-L), where Li is the lowest level for component i and L is the sum of the lowest levels of all components (Barros et al., 2003).
The data obtained for each response were submitted to analysis of variance. Regression models were used to express the relationship between the pseudocomponents in each experiment and the values obtained for optimal cooking time (DCT), mass gain (MG), and loss of solids in water (LS) for the samples of experimental pastas. The graphical representation of the mixture system of each response variable was constructed using triangular diagrams based on the fitted model. The highest desirability formulation was determined using the software Statistica version 7.0 (StatSoft, 2007) to select the pasta with the best technological quality according to the shortest cooking time, highest mass gain, and lowest loss of solids in water.

Pasta production
The pastas were produced in four steps: weighing and mixing dry ingredients, adding water, rolling out and cutting the dough, and drying. The dry ingredients were manually mixed for 5 min followed by water addition and then kneading for 15 min to obtain homogeneous dough. The dough was rolled out in a pasta machine (Marcato, Atlas 150, Otaly) down to 0.4 cm thickness and then cut in the same machine using the dimensions of traditional tagliarini noodles, i.e., 26 cm long and 0.5 cm wide. The pasta then underwent fixed-bed drying at 60, 75, and 90 °C with air velocity of 1.5 m/s, during which mass loss over time was monitored until the product reached 12% moisture. The data obtained during the drying process at the different temperatures were analyzed using the Page, Midilli, Newton, and Henderson and Pabis models in the software Statistica version 7.0 (StatSoft, 2007). The coefficient of determination (R 2 ) and the mean standard error (SE) were used to select the best fit to each equation.

Rheological properties
Pasta texture was analyzed using a TA-XTi texturometer (Stable Micro Systems, Halesmere, England). Firmness and elasticity were determined using AACC method 66-50 (American Association of Cereal Chemists, 2000). Firmness determination used readings of the maximum force required to cut the noodle. The device was operated under the following conditions: force measured in compression, pre-test velocity: 0.50 mm/s, test velocity: 0.17 mm/s, post-test velocity: 10.0 mm/s, distance: 4.5 mm, trigger: base (initial height: 5 mm). Elasticity was determined as the pasta's resistance to break and its extensibility. The test conditions were: force measured in tension, pre-test velocity: 1.0 mm/s, test velocity: 3.0 mm/s, post-test velocity: 10.0 mm/s, distance: 30 mm, trigger: auto -5 g.

Pasta quality analysis
The quality of the dry pasta was assessed after the cooking assay and sensory analysis. The cooking assay was performed according to AACC method 16-50 (American Association of Cereal Chemists, 2000). The quality analyses were performed in the pastas that had the best results in the rheological analysis.
For the sensory analysis, the research project was submitted to the Research Ethics Committee of the Federal University of Pará (protocol no. 1.030.354/15) and was only carried out after approval. Eighty untrained tasters took part in the research and performed the monadic sequential assessment of appearance, texture, flavor, and odor of the pasta using a 9-point structured hedonic scale. The 25 g samples were served alone and with tomato sauce in individual booths at the Sensory Analysis Laboratory. The attributes were judged under natural light (Sereewat et al., 2015).

Pasta chemical analysis
The sample selected in the desirability analysis was characterized according to Association of Dfficial Analytical Chemist (1997) techniques regarding moisture (method no. 925.10), ash (method no. 923.03), crude protein (method no. 960.52), total lipids (method no. 920.39C), and total, soluble, and insoluble dietary fiber by enzymatic digestion (according to enzymatic-gravimetric method no. 985.29). Carbohydrates were determined by difference.

Statistical analysis
The graphical representation of the mixture system of each response variable was made using triangular diagrams. The software Statistica version 7.0 (StatSoft, 2007) was used to create the experimental design, data analysis, and plotting. The data obtained for each response underwent analysis of variance (ANDVA) to assess significance (p < 0.05), coefficient of variation, and adjusted coefficient of determination (R 2 adj) of the mathematical models obtained. The pasta with the best technological quality according to the shortest cooking time, greatest mass gain, and lowest loss of solids in water was selected using the software Statistica version 7.0 (StatSoft, 2007).

Microbiological evaluation
Total and thermotolerant coliforms, molds and yeasts, sulfite-reducing clostridia, and coagulase-positive staphylococcus counts were performed. The techniques described by the American Public Health Association (2001) were followed. The microbiological analyses were performed in accordance with the microbiological standards established by RDC no. 12 of the National Health Surveillance Agency (ANVOSA) of the Ministry of Health (Brasil, 2001).

Chemical characterization of the peach palm flour raw materials
The percentage composition of the raw materials of peach palm flour is presented in Table 2.
The moisture content found for the flours used in this study is below the upper limit of 15% established by the Brazilian legislation (Brasil, 2005). The protein content found was 4.71 and 25.03% for peach palm and linseed flours, respectively. The greatest protein contribution comes from linseed, which adds more nutritional value to the diet. The properties of the protein, particularly its network-forming capacity, are very important for pasta texture (Mariotti et al., 2011).
The total fiber content found in the peach palm flour was 7.65 g. Adding fruits and vegetables to gluten-free extruded products is a possible alternative to provide consumers with celiac disease products enriched with healthy dietary fibers (Stojceska et al., 2007).

Mixture design of experiment
The regression models adjusted for DCT, MG, LS of the pastas prepared based on the mixture design of experiment and dried at 75 °C are presented in Table 3. The quadratic regression model adjusted to the DCT, MG, and LS responses was significant at 5% probability with a regression coefficient of 0.999 for DCT and LS and of 0.77 for MG.
The CS, PPF, and GLF contents were significant for at least one of the responses (p < 0.05). Ot can be seen that DCT was significantly influenced by the three raw materials used, while MG was influenced only by GLF. LF was influenced by PPF and GLF.
Regarding the interactions among the components, DCT was significantly impacted by interactions between CS and PPF and between PPF and GLF. The interaction between CS and PPF was significant for LS, while PPF was the component that most influenced LS. MG, in turn, was not influenced by the interactions among those matrices.
The pasta samples had different DCT values. Figure 1 shows the lowest DCT for assays 1, 3, 4, 5, 6, 7, and 8 with values of 7.20, 10. 64, 5.39, 9.26, 10.13, 9.94, and 9.94 min, respectively, which are in the experimental region delimited by points a, b, and c. A similar value was obtained by Baiano & Conte (2006) for pasta with semolina and wheat flour, whose DCT was 8 min.
The MG parameter is related to the water-absorption capacity of the pasta and depends on its shape. According to Hummel (1966), wheat-based pastas should have minimum weight gain of 100 g/100 g. According to the criteria of that author, only the value of assay 4 was beyond the threshold for that quality parameter. The lowest value obtained was 31.08% (assay 2) and the highest, 113.5% (assay 4). The highest MG values are in the region delimited by points a, b, and c.
The LS values found showed that the pastas prepared in assay 1 (1.90%), 4 (4.84%), and 5 (8.62%) are within the quality range according to the criteria by Hummel (1966). That author stated that losses of solids of up to 6 g/100 g are characteristic of high-quality wheat-based pastas. The lowest LS values are in the region delimited by points a, b, and c.
The values optimized for DTC, MG, and LS were observed in assay 4 of this design, whose formulation had the highest peach palm flour content and the best results among the quality parameters. On face of the results obtained, the pasta from assay 4 was chosen for the rheological, quality, microbiological, and sensory tests.

Chemical composition of the pasta
The moisture content of the dry pasta was 12%, within the range recommended by the current legislation, which sets a maximum limit of 13% moisture (Brasil, 2000). The pasta prepared had 10.33 g of total dietary fiber, with a breakdown of 2.77 g of soluble and 7.59 g of insoluble fibers. That value is higher than the pastas available on the market, thus this product is considered a food rich in dietary fiber (Brasil, 2012). The higher fiber content may have beneficial physiological effects to the organism, which may include reduction in serum cholesterol, glycemia modulation, and laxative properties (Brennan, 2005). After processing, the carotenoid content in the final processing was 1.16 µg/g, which is close to the 1.8 to 2.9 µg/g found by Pereira et al. (1998) in pastas produced with wheat flour and fortified with synthetic β-carotene.

Rheological analysis
The technological quality of pasta during and after cooking is the most important quality parameter for consumers of this product, besides taste and aroma. The experimental results Table 3. Adjusted regression models, regression coefficients (R 2 ) for optimal cooking time (DCT), mass gain (MG), and loss of solids (LS) of the experimental pastas as a function of variables Cassava starch (CS) (x), peach palm flour (PPF) (y), and golden linseed flour (GLF) (z).  of the texture characteristics of the pasta dried at different temperatures obtained in assay 4 of this experimental design are presented in Table 4.
On the statistical analysis, the drying parameters significantly impacted pasta texture (firmness and elasticity). No difference in firmness was found between the samples dried at the lowest temperature (60 °C; 223 gf) and the intermediate temperature (75 °C; 205 gf), which had the best results for this parameter. This behavior was also observed by D' Amico et al. (2015) when assessing the effect of drying (60, 80, and 100 °C) on two types of gluten-free pasta, which justifies that the drying process at high temperatures may change the physicochemical properties of starch and of the protein present in the pasta.
Pasta elasticity significantly differed for the three temperatures studied, yielding values of 47.72, 34.26, and 43.44 gf, respectively. The drying effects at the temperatures studied were less pronounced in the pasta and elasticity was not significantly extended. Gluten-free pasta has lower elasticity (Huang et al., 2001). Aravind et al. (2012) observed a significant reduction in pasta viscosity by adding soluble fibers, which is consistent with a competition between dietary fibers and starch for the available water, inhibiting starch pasting and retrogradation and/or disruption of the starch/protein matrix. On this case, the food fibers probably interfere with the pregelatinized starch and pasta viscosity is directly influenced by the amount of this available starch to structure the pasta, which also decreases the firmness of the starch (Hsieh et al., 1991).
Although high-temperature drying appears to be promising for improving pasta elasticity, it should be stated that the values achieved require further improvement.

Quality characteristics
On order to analyze pasta quality parameters, the positive results obtained in the rheological analysis of the pasta dried at 60 and 75 °C were considered and no significant difference in firmness was found.
The cooking time of the dry pasta was 8 min. For the shortest DCT values (<9.5 min), the pastas with maximum CS (55-80%) and minimum amounts of PPF (10-35%) and GLF (10-35%) can be found in the area delimited by points a, b, and c ( Figure 1a). The same cooking time was obtained by Sereewat et al. (2015) for pastas based on rice flour and defatted soybean. A trend was observed for the pastas with lower CS and higher PPF and GLF contents to result in longer DCT (<10 min). This may occur because fibers and proteins in the PPF and GLF absorbed more water during cooking, thus reducing the amount of available water for starch gelling and requiring longer cooking times.
According to Hummel (1966), high-quality pastas must have minimum increase values of 100%. According to this criterion, the pasta prepared in this study (assay 4) had a satisfactory value (113.5%) for this quality parameter. A trend was found for greater MG with small PPF and GLF amounts and higher amounts of CS in the experimental formulations ( Figure 1b). However, since it is desirable that the pasta doubles in size (100% increase), this trend occurred close to experimental points 3, 5, and 6, i.e., when the contents of GLF were between 10% and 35%, CS contents were between 30% and 55%, and PPF contents were between 10% and 35% (region delimited by points a, b, and c). That value is close to the one reported by Bastos et al. (2016), who found MG of 111.5% for pastas prepared with amaranth flour and potato pulp.
Based on the criteria by Hummel (1966), classifies pasta according to the analysis of loss of solids. High-quality wheat-based pastas are those whose LS is up to 6%. Based on that classification, the pasta prepared in the present study can be considered good given its LS of 5.6% (assay 4). A trend was also observed for lower LS values (≤8%) near experimental points 1, 4, and 5 ( Figure 1c), a region delimited by points a, b, and c, representing formulations with intermediate amounts of GLS (10-33%) and PPF (10-33%) and maximum amounts of CS (55-80%). Ot can also be seen that CS influenced this index as the highest LS values (>10%) were found in the regions with the lowest CS amount.

Microbiological analysis
The microorganism counts of the dry pasta were within the legal limits over 5 months of storage, confirming its good hygienic and sanitary quality according to the current legislation (Brasil, 2001).

Sensory analysis
The dry pasta was approved by 96, 90, 89, and 88% of the tasters regarding appearance, aroma, texture, and flavor, respectively, with a mean score of 7.7 for all attributes assessed. That value is higher than the one reported by Mirhosseini et al. (2015), who obtained an average score (<7) for pasta produced with corn flour and pumpkin meal. The intent of purchase survey showed that 47.5% of tasters "would certainly buy" and 33% "would possibly buy" the dry pasta made with cassava starch and peach palm and golden linseed flours.

Conclusion
Adding cassava starch and flours from peach palm and golden linseed enabled the production of well-structured dry pastas. All formulations or mixtures studied produced firm pastas, which were consistent with different cooking characteristics.
The combination of 55% cassava starch, 35% peach palm flour, and 10% golden linseed flour resulted in pasta with the best cooking and nutritional characteristics. Adding peach palm and linseed flours to that pasta improved its nutritional properties, particularly concerning fiber content, thus it can be characterized as a food rich in fibers. The drying temperatures of 60 and 75 °C positively impacted pasta firmness. The optimal cooking time, mass gain, and loss of solids of the pasta were within the quality thresholds of this type of food product. The pasta prepared complies with the microbiological standards of the current sanitary legislation. Furthermore, the pasta had a good acceptance index and intent of purchase by tasters.
The results showed that the mixture of cassava starch, peach palm flour, and golden linseed flour has potential for commercialization and pasta preparation. That mixture widens the possibilities of using peach palm flour for consumption in finished products. The possibility of developing gluten-free products with peach palm flour may expand the supply of food products to individuals with celiac disease and will contribute to a more diversified diet for this population group.