The cereals used were produced in 2020 and obtained from “SEGALCO” Company S.A.S. Popayán, Colombia. “SEGALCO” Company provided the quinoa and a polishing machine (500-T, Mavimar, Colombia) was used with a processing capacity of 60 Kg/h, the operative system functioned through rock or stone discs, connected to a central shaft, which rotate inside a cylinder with 1 mm perforations, processing a polished grain, and quinoa hyper-protein flour (HPQF).

The multigrain snack pellets were made according to the formulations shown in Table 1, composed of a cereal blend (CB) corn flour, rice flour, cornstarch and hyper-protein quinoa flour (HPQF), in a single screw extruder. The mixtures were wetted to 27% by spraying the calculated volume of water and mixing in a laboratory blender. The samples were packed in polyethylene bags and stored in the refrigerator overnight at 4°C to equilibrate the moisture. Before extrusion cooking, the samples underwent room temperature, then subjected to the extrusion-cooking process extruder temperature profile: 140°C, screw rotation speed was 200 rpm, nozzle diameter was 3 mm. The snacks were dried at 60°C for 17 h.

Quantification of protein (A.O.A.C 2005, Method 955.39), fat (A.O.A.C 2005, Method 920.85), fiber (A.O.A. C 2005, Method 945.38), and moisture (A.O.A.C 2005, Method 925.10), contained in the samples. Carbohydrates contents were calculated by difference as 100—(proteins + fat + water + ash). Values were expressed on a dry weight basis.

Extrudate density (ED) was measured according to the method described by Li et al. (2019), with samples (n = 10), using chia seeds, and calculated using Equation (1):

PE and VE are the weight and equivalent volume of the extrudate; ρs and Ws represent the density and weight of the chia seeds. ED measurements were performed fifteen times.

The expansion index (EI) was determined by dividing the radial diameter of the extrudate, measured with a caliper, by the extruder die bore diameter (3 mm). Expansion index measurements for each extrusion series (n=15), and a mean value was recorded.

The mechanical properties of the extrudates were determined using a texture analyzer (Shimadzu EZ TEST SM, model 500N-168, Japan). Twenty-five extrudates from each treatment were equilibrated to a moisture content of 8.0 ± 0.5% (wb) using a humidity chamber (30% RH, 94 h). Samples were compressed perpendicular to the extrusion direction to 50% of their original diameter with a 5 mm compression punch at a test speed of 1.0 mm/s. The area under the curve (S; N.mm) and the number of peaks (n) above 1.5 N were obtained from the force-strain curves and used to calculate the spatial frequency of ruptures (Nsr), average crushing force (Fcr), and work of crispness (Wc), following the method of Karkle et al. (2012), with slight modifications.

where d = probe travel distance (mm).

The method described by Wani and Kumar (2016) was used to determine water absorption index (WAI) and water solubility index (WSI). The following were used 10 mL of distilled water was transferred into a 15 mL falcon tube. The extrudates were ground and sieved through a 100 μm sieve. 0.5 g of ground and sieved extrudate was added to the water in the falcon tube. The mixture was shaken in chambered orbital shaker (MaxQ 4450, Thermo Scientific, Germany), for 60 min. The tubes were centrifuged (Z 326, Hermle, Germany) at 3,000 × g for 10 min at 25°C. After centrifugation, the supernatant was at once transferred to a plate and dried in an air oven (ED1 115, Binder, Germany) at 90°C overnight. The supernatant cooled for 30 min to room temperature and weighed. WSI was calculated by Equation (5) and WAI was calculated by Equation (6). All experiments were performed in triplicate.

The color of the extrudate (n = 3) was determined using a Konica Minolta Spectrophotometer CM-5, controlled by SpectraMagic NX software, with D65 illuminant and 10° observer angle. The samples were analyzed in triplicate, taking 5 g of sample. Luminosity (L) and chromaticity parameters a^* (-a ^* = green and + a ^* = redness) and b ^* (-b ^* = blue and + b ^* = yellow) were measured. The total color change (E) was calculated as Equation (7).

The rheological properties of each dispersion were determined using a rheometer (TA INSTRUMENTS, AR 1500, New Castel, USA), equipped with a starch pasting cell used to evaluate the pasting properties of extrudates and non-extruded blends. Three gram (12% moisture basis) of ground and sieved sample (100 μm) and 22 ml of distilled water were mixed to obtain a paste in an aluminum pot. The shear rate was kept constant at 16.75 s-1, throughout the heating range (25 °C−95°C−25°C) while the heating rate was 10°C/min. Peak viscosity (PV), viscosity drop (BD), and set back (SB) were recorded in duplicate (Pasqualone et al., 2021). Using the Savistky-Golay function, the data was smoothed in GraphPad Prism version 6.

A one-way factorial design with four levels of factor quantity of high protein quinoa flour and n = 3 independent replications was used in each treatment. The experimental data was analyzed by analysis of variance (ANOVA) and Tukey's test, defining significance as p < 0.05. All statistical analyses were carried out with GraphPad Prism 5.0 software.

Table 2 shows the main composition of rice flour, maize flour, hyper-protein flour, whole quinoa seed, and snacks. The protein content reported for whole quinoa seed was comparable to data presented by the literature (Abugoch, 2009). The protein content of the HPQF is 53% higher than the whole quinoa, this value corroborates that the polishing process with the abrasion mill allows the removal of the pericarp which decreases the volume of dietary fiber (and saponins) but not the protein or fat content (D'Amico et al., 2019). The moisture content of the snacks is <9.11%, this percentage is a quality index that indicates that the product will be less susceptible to the effect of microbes, therefore, it will have a long stability (Sobowale et al., 2021).

The aim of adding HPQF was increasing the protein content of the snacks, nut crops have been shown to increase the protein content of snacks, as has been established with quinoa-based extrudates (Zapana et al., 2020). Table 2 shows the significant increase (p < 0.05) of protein, fiber and lipids with the increase of the proportion of quinoa flour while the protein values were as expected according to the proportion of HPQF included in each mixture since during the extrusion process, the effect on protein content is minimal (Gulati et al., 2020). The lipid and fiber content of the snacks increased with the addition of quinoa flour. Comparing the raw materials with sanck, a decrease in lipids was observed, due to the formation of complexes between amylose and lipids during extrusion processing, which resulted in a reduction of lipid extractability (Jabeen et al., 2021).

Table 3 shows the effect of HPQF addition on the EI and ED of the extrudates. Adding HPQF, the expansion index of the extrudates decreased, while the influence of adding 25 and 30% was not significantly marked, adding 37% HPQF decreased expanding the extrudates, decreasing from 2.31 to 1.22 compared to snacks without quinoa. The formulations prepared with quinoa contain a higher proportion of protein and fiber; the latter breaks the cell walls and prevents forming air bubbles and expanding to its maximum potential (Jozinović et al., 2016). Likewise, the lipid content influences the expansion properties, in the 37% mixture, 6.72% lipids are present, which with humidities above 21% and temperature profiles above 128°C, favor forming a starch-lipid complex, resulting in a snack with undesirable physical properties, with high density and hardness (Pilli et al., 2012). A high fat content provides a lubrication effect in the barrel at high temperatures, which reduces its expansion (Chadha et al., 2021). Shah et al. (2016) report similar studies on extrudates from grains.

The ED is a very important attribute from the point of view of commercial production of extruded products because most are filled by weight and not by volume (Tas and Shah, 2021).

The results for ED agree with those for expansion index, i.e., extrudates with added quinoa flour have higher ED, proportional to the percentage of quinoa, compared to those with corn. This observation is consistent with Tas and Shah (2021) in that ED and EI are dependent attributes with an inverse relationship. As seen in Table 3, the ED increased with quinoa composition, which has a higher fiber content compared to corn flour because the fiber undergoes minimal size reduction and remains stable during the extrusion process, thus limiting expansion and improving ED, furthermore during extrusion, a higher feed moisture content above 20% reduces the melt viscosity, leading to a drop in the vapor pressure gradient and thus causing a laminar flow of the melt restricting expansion and thus increasing the ED (Jabeen et al., 2021). Similar results were reported by Wani and Kumar (2016) on snacks based on fenugreek, oat and pea.

Figure 1 shows the textural properties of the extrudates. Quinoa content affected the average crushing strength Fcr (p < 0.05), the 30 and 37% blends presenting higher quinoa contents lead to higher values of Fcr, likewise these blends lead to extrudates that require higher work Wc and are harder, these extrudates presented the lowest expansion, which agrees with the results by Jozinović et al. (2016), who concluded that textural properties are influenced by the expansion degree, i.e., extrudates with higher EI had lower toughness and higher fracturability. In a study on the mechanical and microstructural properties of extrudates from corn starch with high amylose content and soy protein concentrate, Zhu et al. (2010) observed that hardness increased with the combination of decreased expansion and protein agglomeration produced by high temperature and shear.

WAI decreased slightly and WSI increased proportionally to the addition of quinoa flour to CB, but after the extrusion process, both parameters increased significantly.

WAI is the measure of the volume occupied by the starch after swelling by excess water, so it can be used as an index of starch gelatinization (Tas and Shah, 2021). As shown in Table 4, the formulations affected the WAI (p < 0.05) of the extrudates, although there are no major differences, the formulations with quinoa present lower values regarding the extrudates with rice and corn, because there is an obstacle in the ability of starch to combine with water because of the greater interaction between the hydrophilic groups of starch and quinoa protein.

As shown in Table 4, the WAI values of extrudates are higher than raw flours because gelatinized starch absorbs water better than native starch granules at room temperature. WAI decreases when dextrinization or starch melting prevails over gelatinization (Sarawong et al., 2014).

WSI shows the degree of molecular degradation and conversion of starch and measures the soluble components released from starch after extrusion (Ye et al., 2018). As shown in Table 4, formulations affect WSI values (p < 0.05), presenting higher percentages in mixtures added with 30 and 37%. This behavior occurs because the formulations present a higher proportion of fat and generate a plasticizing effect, while the high lipid content reduces starch degradation (Tas and Shah, 2021).

Color change is a good measure of the effect of extrusion cooking on extrudates, especially L^* is the best parameter to estimate browning kinetics (Alam et al., 2016). Table 4 shows the color parameters of raw flours, extrudates added with quinoa, and those with only rice and corn.

Compared to the raw formulations, all extrudates had lower L^* (lightness) values, a^* (redness), and b^* (yellowness) values only in the extrudates containing corn and rice. The a^* and b^* values of the extrudates decreased, while the L^* value increased.

As shown in Table 4, the brightness increment in the snack without quinoa suggests an increase in air bubble nucleation sites, resulting in a higher number of cells in the snack (Li et al., 2019). In formulations with quinoa, especially snacks added with 37%, when subjected to excessive dry heat and high-shear forces, dextrinization occurs in the starch, which decomposes into dextrin, altering the color of the extruded products in a yellow to brown color. This confirms the WAI results, where lower values associate with higher dextrinization.

The total color change (ΔE) that shows the color change compared to the raw cereal mixture without quinoa ranged from 30.15, for extruded corn to 38.02 for the cereal and HPQF mixture. According to studies by Jozinović et al. (2016), in raw and extruded cereal mixtures, most of the population perceives a total color difference higher than 3, if the color difference is higher than 6, it is clear the colors are in different color groups. Therefore, adding HPQF to the cereal mixture results in a significant color change both before and after extrusion, as shown in Figure 2.

Table 5 shows the pasting properties of non-extruded and extruded samples. Adding HPQF and after the extrusion process, the cold and hot viscosity values decreased proportionally to the quinoa flour content. There is documentation that the molecular degradation that occurs during extrusion influences lower viscosity values for extruded starch; this decrease in viscosity suggests a molecular size of starch reduction during this process (Jozinović et al., 2016). In addition, matrix components such as fat, protein, and fiber could affect starch gelatinization by interacting (e.g., coast blinds, or complexes) with the granule and contribute to reduced melt viscosity (Ondo et al., 2013). This is consistent with the WAI results, as the fat content along with the high pressure and temperature increases in pressure inside the extruder barrel and suggests a higher level of starch degradation.

Because of starch retrogradation during cooling, after heating and mixing, the viscosity of all samples decreased and after cooling, the viscosity of all samples increased. The viscosity values of the non-extruded samples were higher considering all extruded samples, as shown in Figure 3. This suggests the result of starch damage and the lower tendency of the molecules to tightly pack during cooling. Therefore, smaller polymers have a higher number of hydroxyl groups and a higher affinity to form hydrogen bonds with water molecules; during cooling, re-association between amylose molecules results in a gel formation and leads to an increase in the cold viscosity of the extrudates (Jozinović et al., 2016).

In this study, there was a greater viscosity reduction for extrudates without quinoa followed by quinoa flours and extrudates added with quinoa. The difference between the first and final viscosity of the starch extrudates in the pasting profile reflects the starch transformation during the extrusion, which can indirectly quantify its degradation degree (Fallahi et al., 2016).

The fact extruded mixture viscosities are unaffected by extrusion temperature; starches behave the same at the temperature profile employed; quinoa starches present in the extrudates gel faster and at lower temperatures compared to corn starch may be due to the more organized crystalline structure of corn starch (Fallahi et al., 2016).

Although there was no significant difference between the initial and final extrudates' viscosity, the lower values are directly associated with lower expansion index values. In this study, extrudates with 37% quinoa presented the lowest ER value according to the lowest viscosity value, also reported by Fallahi et al. (2016).

“Breakdown” values of non-extruded samples decreased after adding quinoa flour, showing the better stability of non-extruded samples during mixing at extrusion temperature, as quinoa starch of various genotypes has a lower amylose content (7–27%), its size can vary between (0.5–3 μm in diameter) and these granule characteristics make quinoa starch a successful candidate for stabilizing food grade emulsions (Wang and Zhu, 2016). In their results, Jozinović et al. (2016) report a positive effect of the interaction between temperature and adding high fiber content spelt flour on the increase of “breakdown” values, which is attributed to the effect of fiber within the starch matrix that was enhanced by temperature. Breakdown values of extruded samples were lower regarding non-extruded samples and the positive effect of quinoa flour addition on viscosity stability was higher at 37% concentration.

“Setback” values are the indicator of retrogradation and rearrangement of starch molecules during the cooling process (Jozinović et al., 2016). “Setback” values decreased significantly after extrusion, showing that the extruded samples were less prone to retrogradation and hence extrusion may contribute for the final product's stabilization (Wani and Kumar, 2019).

The mixture of cereals and HPQF has a lower “retrogradation” value, regarding the extrudate mixture without quinoa. This may be due to a higher fiber content in the samples with quinoa flour, as presented in the proximate analysis, which is consistent with other studies' findings (Nascimento et al., 2012; Jozinović et al., 2016).