Quality and Technological Properties of Gluten-Free Biscuits Made with Pachyrhizus ahipa Flour as a Novel Ingredient

The development of gluten-free foodstuffs with high nutritional quality components is an important objective to achieve. Pachyrhizus ahipa is one of the few leguminous species that produce edible tuberous roots with high nutritional value. Thus, the aim of this work was to formulate biscuits containing corn starch and 125 350 g·kg of P. ahipa flour (AF) and to study the main physicochemical properties related to their nutritional and technological quality as well as their sensory acceptability. The formulation containing 250 g of AF kg was selected for improving the product nutritional quality (i.e. higher protein content) without an extensive modification of textural properties. A formulation replacing corn starch by cassava flour (250 g·kg) was also analyzed. In this case, ahipa and cassava biscuits showed maximum force and energy required to bite twice and three times higher than the control, respectively. Slight variations were observed in color (∆E and browning index). The overall acceptability of biscuits formulated with both flours was better scored than the control by a sensory panel. Principal component analysis allowed to relate quality attributes, chemical composition, and sensory characteristics of biscuits containing cassava or ahipa flours. The results indicated that ahipa flour could satisfactorily substitute part of the corn starch used in the formulations of gluten-free biscuits.


Introduction
In quantitative terms, the gluten-free food market has shown a big grow in the last years, evidencing an annual rate of 28% [1].For instance, the US gluten-free food and beverage sales are expected to surpass $6.6 billion by 2017 [2].Nevertheless, most of the time nutritional status and demands of celiac patients are not satisfactorily covered when formulation and production of gluten-free products is carried out [3].
People on a strict gluten-free diet are frequently undernourished since their rapidly available-energy intake, which in the western diet is largely taken from wheat-based foodstuffs, is reduced.This circumstance has prompted the development of gluten-free foodstuffs with high nutritional quality components [4] [5].In general, gluten-free baked goods comprises several raw materials like rice, corn, buckwheat, cassava, and potato starches and/or flours [6].At present, crops like amaranth, teff, and quinoa, which are comparatively novel or lesser known on the gluten-free market, are offered.
Thus, the variety of gluten-free products is expanding every year [7].
On the other hand, in many tropical areas where wheat does not grow well, attempts are made to encourage the use of flours from locally grown crops which could totally or partially replace wheat flour for use in baked goods.This fact would contribute to decrease the demand for imported wheat and to obtain protein enriched products [8].
Pachyrhizus is one of the few genera of the leguminous family that produce edible tuberous roots with high nutritional value [9].The species P. ahipa (Wedd.)Parodi produces thickened roots that accumulate low-amylose starch.Likewise, the protein content of P. ahipa roots has also motivated scientific and technological interest in this product [10] [11].From a nutritional point of view, P. ahipa flour showed a more balanced chemical composition than other root and tuber (R & T) flours.It might contribute protein, fiber, and minerals, such as potassium, calcium and iron to the diet.Ahipa flour can be simply obtained by a slicing procedure bringing a product with higher content of potassium, magnesium, calcium and protein together with a higher water-holding capacity than the flour obtained by grating plus pressing [11].Ahipa flour showed an acceptable stability when kept at 10˚C, 20˚C or 30˚C at variable relative humidity, information obtained from the sorption isotherms analysis [12].
The objective of the present work was to formulate and elaborate biscuits containing Pachyrhizus ahipa flour and to study the main physicochemical properties related to their nutritional and technological quality as well as their sensory acceptability.

Cooking Conditions and Biscuit Analysis
Biscuits were cooked in an electrical static oven (Ariston FM87-FC, Italy) under natural convection.Different baking conditions were assayed: 180˚C-10 min; 200˚C-7.5 min; and 200˚C-10 min.The main quality attributes (water activity, color, and texture) were analyzed, evaluating the biscuit characteristics for selecting the appropriate baking condition.Once the percentage of substitution of ahipa flour was selected, the formulated product was compared to the control and with a formulation made with the same proportion of cassava flour, which was chosen as a reference ingredient frequently incorporated in gluten-free product formulations.

Water Activity (aw)
Water activity of the samples was measured at 25˚C with a Water Activity Meter Aqualab series 3 (Decagon Devices Inc., Washington, USA).A solution of saturated K 2 SO 4 was used as a calibration standard (a w at 25˚C: 0.972).The average of three determinations, performed in two independent experimental batches, was informed.

Texture
Penetration tests were performed with a TAXT2i Texture Analyzer (Stable Micro Systems Ltd., Godalming, Surrey, UK).A 25 kg load cell was employed.The "Volodkevich Bite Jaws" probe was used in order to simulate the bite action of the incisive teeth on the sample.Parameters recorded were the maximum force (N) exerted when compressing 20% the sample, which is related to product firmness, and the area under the force curve (N mm) that is directly associated with the energy required in the process [14].Likewise, the number of peaks (variations in force higher than 0.02N) taken as an index of crunchiness were counted.The average of twelve determinations, performed in two independent experimental batches, was informed.

Color
Surface color was measured using a Chroma Meter CR 400 (Konica Minolta Sensing Inc., Japan) calibrated with a standard white plate (Y = 93.2,x = 0.3133, y = 0.3192).
Lightness (L*), red-green coordinate (a* value), and blue-yellow coordinate (b*) were registered.Color measurements were expressed as the total color difference (ΔE) calculated with respect to the coordinates (a 0 *, b 0 * y L 0 *) that characterized the color of the control baked cookies (C) (Equation (1)): Likewise, the browning index (BI) [15] was calculated using the following equations: The average of twelve determinations, performed in two independent experimental batches, was informed.
The liposoluble fraction (g•kg −1 on a dry basis) was extracted with hexane from samples of grinded biscuits (12 g) in a Soxhlet apparatus.Likewise, baked biscuits (0.8 g) were analyzed for total nitrogen content by the Kjeldahl method [16] and the results were expressed as crude protein (g•kg −1 on a dry basis).The content of total dietary fiber (TDF, g•kg −1 on a dry basis) was measured by the enzymatic kit K-TDFR 05/12 Megazyme © (Ireland) from samples (1.0000 g) of grinded biscuits, previously defatted with hexane by the Soxhlet method, until completing eight cycles of extraction.Chemical analyses were carried out at least by duplicate, in two independent experimental batches.

Sensory Evaluation
A first sensory analysis was carried out using a semi-structured hedonic scale to evaluate the acceptability of the biscuits made with the selected percentage of substitution, cooked at different baking conditions (200˚C-7.5 min or 200˚C-10 min).A total of 60 potential consumers evaluated biscuit color, texture, taste and overall acceptability according to a box-scale (1 -9) anchored in the following steps: "dislike very much", "indifferent" and "like very much".Samples were presented to the evaluators randomly arranged and coded with three digit numbers.
Once the baking time was selected, a second sensory analysis was carried out where the panelists were asked to evaluate color, texture, taste, and overall acceptability of control and the selected ahipa or cassava flour biscuit formulations.

Statistical Analysis
The InfoStat Software (Version 2008) (InfoStat Group, Agricultural Sciences College, National University of Cordoba, Argentina) was used.Analysis of variance (ANOVA) and comparison of means with the Fisher's least significant difference (LSD) test were conducted, at a significance level p = 0.05.The results were subjected to a Principal Components Analysis (PCA) and Cluster Analysis (CA) in order to observe the differences and similarities of the analyzed samples.

Biscuit Formulation and Baking Condition Selection
Figure 1 shows the formulated biscuits submitted to different baking conditions as well as their ∆E (Table insert in Figure 1).The conditions 180˚C-10 min and 200˚C-7.251 g of simple sugars kg −1 on dry basis [17], which support color observations (Figure 1).According to Chevallier et al. [18] cookie surface color is the result of non-enzymatic browning between reducing sugars and amino acids as well as from starch dextrinization and sugar caramelization.Concerning texture measurements, the curves for the different biscuit formulations lacked a "defined" profile pattern in the sense that they were variable in shape and the remaining curves lost the initial penetration peak, except for the 200˚C-10 min cooking condition.McWatters et al. [19] reported that the hardness of cookies is caused by the interaction of proteins and starch by hydrogen bonding during dough development and baking.On the other hand, the texture of gluten-free biscuits is primarily attributable to starch gelatinization and super-cooled sugar rather than to a protein-starch structure development [20].
The textural parameters analyzed were the maximum force associated with the strength of the sample and the area under the curve, which represents the work done on the break (Table 1).In all tested baking conditions, 350AF biscuits presented the highest values of both parameters.Gaines et al. [21] used a puncture test on various biscuits and reported that the fracture force rose with increasing flour protein content.On the other hand, de Simas et al. [20] found a similar trend working on gluten-free cookies containing up to 300 g•kg −1 of king palm flour as a rich-fiber ingredient.In this sense, it is worth noting that ahipa flour supplies both protein and fiber to the biscuits formulated with corn starch.for the 350AF formulation were 3.15, 3.2 and 4.3 times higher than the control for 180˚C-10 min, 200˚C-10 min and 200˚C-7.5 min, respectively.Conversely, for the 250AF formulation, maximum force was 1.38, 1.65 and 2.0 times higher than the control, for the same baking conditions respectively (Table 1).Thus, biscuit hardness estimated through maximum compression force was markedly higher for the highest ahipa flour substitution percentage.A comparable trend was observed for the area under the curve.

When comparing maximum force at different baking conditions the values obtained
Demirkesen [22] evaluated the influence of chestnut flour at different levels on the texture of gluten-free cookies and found that the cookies prepared using exclusively chestnut flour had the highest hardness values.The author mentioned that the very high viscosity of chestnut cookie dough prevented air incorporation during the mixing and resulted in hard texture.Likewise, the high sugar content of chestnut flour could restrict the starch gelatinization and caused harder texture in the cookies.
Table 2 shows a w and moisture content for the different formulations submitted to the assayed baking conditions.Except for control biscuits baked at 200˚C-7.5 min, the remaining baking conditions showed a w values lower than 0.600.For food products, a w > 0.6 could favor microbial growth and affect their stability and conservation.For all formulations, values of a w and moisture were lower in the condition of maximum temperature and time (200˚C-10 min), ranging in this case from 0.304 to 0.400 and 62.2 to 95.3g•kg −1 , respectively.
Considering that the inclusion of ahipa flour could improve the nutritional profile of the product on the basis of its higher protein and fiber content, the formulation 350AF might make the major contribution in this sense.However, textural measurements for this formulation revealed that hardness was significantly higher than the controls, added to the highest ΔE.Thus, the formulation 250AF was selected and characterized.
Owing to color and texture measurements (Figure 1 and Table 1), the baking conditions 200˚C-7.5 min and 200˚C-10 min were pre-selected.Then, one time-temperature combination was finally chosen from a sensory test, where the two alternatives for cooking the selected formulation (250AF) were evaluated.The attributes analyzed were color, taste, texture and overall acceptability (Figure 2).For the analysis of the results, values from the hedonic scale were grouped in three ranges: 1 -3 (dislike); 4 -6 (not like nor dislike, i.e. indifferent); and 7 -9 (like very much).
Figure 2(a) shows that 63% of the potential consumers gave higher score (7 -9) to the color of the biscuit baked at 200˚C-7.5 min, meanwhile 57% of the panelists gave those high punctuations to the color of the biscuit cooked for a longer time (10 min) at the same temperature.Concerning the attribute taste, 75% of the potential consumers selected the scores 7 -9 from the hedonic scale for the condition 200˚C-7.5 min (against the lower percentage of 58% in the case of biscuits baked at 200˚C-10 min).However, the texture of the product was better appreciated when biscuits were cooked for a longer time (Figure 2(a)).On the other hand, the score 8 for overall acceptability was chosen more frequently for the baking condition 200˚C-7.5 min, whereas for the condition 200˚C-10 min, the score 7 showed the highest frequency of choice (Figure 2(b)).Based on these results, the biscuits made with 250 g of AF kg −1 , baked at 200˚C-7.5 min were chosen as the most widely accepted by the evaluation panel.

Novel Ahipa Biscuit Formulation Compared to a Cassava Flour Equivalent Recipe
Chemical composition of the biscuits formulated with 1000 g corn starch kg −1 , and partially substituted (250 g•kg −1 ) with ahipa or cassava flour was evaluated (Table 3).Total ash content was higher in the cookies containing 250 g•kg −1 of flour (cassava or ahipa) than in the control (1000 g corn starch kg −1 ).This result is related to the botanical origin of the flours.Higher ash content is associated to flours obtained from tuberous roots while the contribution of corn starch to the ash level is much lower [23].
Total lipids found in the biscuits were similar and no significant differences (p > 0.05) were observed (Table 3), mainly due to the high fat content (butter) included in the formulations.Besides, the content of total carbohydrates (g•kg −1 ) showed no significant differences (p > 0.05) between the three analyzed formulations (Table 3).
Biscuits made with ahipa flour exhibited the highest protein content, due to the chemical composition of the flour [11].Cookies containing 1000 g corn starch kg −1   showed the lower values (13 g•kg −1 ) that indicate a comparatively lower nutritional quality of this product.Cassava flour did not contribute to increase the protein content of the biscuits, thus the values were similar to those of the control formulation.
Likewise, the total dietary fiber content in AF-biscuits was the highest (50 g•kg −1 ), because ahipa flour can be considered as a fiber source [13], enhancing the nutritional profile of the developed product.
Table 3. Parameters of the texture analysis, color, water activity, moisture content, and chemical composition of biscuits containing 250 g•kg −1 of ahipa or cassava flour in their formulations, baked at 200˚C-7.5 min.Concerning biscuit texture, the maximum force registered for the 250AF biscuits was higher (1.4 -2.0 times) than the control.Formulations including ahipa or cassava flour showed the same texture behavior, with regard to maximum force, area under the curve and crunchiness (Table 3), taking into account that peaks in biting test represent fracture events, where an association between the number of peaks and the perception of crunchiness has been mentioned [24].Thus, ahipa and cassava biscuit formulation showed similar performance in terms of texture, being the maximum force and energy required to bite twice and three times higher than the control, respectively.
Regarding color measurements, partially substituted formulations differed significantly from the control.Although statistically significant (p < 0.05), slight differences between cassava and ahipa flour formulation were found in terms of total color difference (ΔE) and browning index (BI) (Table 3).It is worth noting that the value used to determine whether the ΔE was visually obvious was ∆E > 3, which denotes color differences noticeable to the human eye [25].Particularly, color differences could be attributed to the chemical composition of ahipa flour, mainly to protein and fiber content.These darker tints have been explained by a greater number of reducing ends involved in a Maillard reaction [26].
Both moisture content and a w were higher for control biscuits (Table 3).Moreover, a w value for control formulation indicated that this product could have compromised its microbial stability during storage.Aparicio-Sanguilán et al. [27] reported increased moisture and a w for biscuits enriched with a resistant starch-containing product extracted from bananas; the increase was related to a relative abundance of amorphous starch zones, which had a considerable influence on water absorption.
Complementing chemical and physical analysis, results from the sensory tests performed in order to evaluate the acceptability of the developed product (250AF) in comparison to the control biscuits and the same percentage of substitution with cassava flour (250CF) are shown in Figure 3.The average scores for taste, color, texture and overall acceptability are presented in the Table insert in Figure 3.In this case, 75% of the panelists rated with scores > 7 the 250AF biscuits in terms of color, preferring them to the control (Figure 3(a)).Thus, the change in color was deemed acceptable for consumers.
For the attribute taste, the most accepted formulation was the one made with 1000 g of corn starch kg −1 (Figure 3(a)).Conversely, when evaluating the attribute texture, the biscuits that had higher acceptance were those made with cassava flour.Possibly, the panelists recognized control biscuits as "more familiar" in terms of taste, since it represents a product more frequently consumed even by non-celiac people.Texture of the control formulation was comparatively less appreciated, being associated by consumers to a "gritty" or less cohesive product.Thus, the association between crunchiness and force peak number does not seem so clear or evident in this case, according to texture and sensory results comparison.
Overall acceptability of the biscuits in the sensory panel was evaluated as good, given that the predominant score was 7 for this attribute (Figure 3(b)).This fact would be favorable for a potential market introduction, since the achieved general acceptance could be similar to biscuits formulated exclusively with corn starch, which are usually available in shops and shelves of products suitable for celiac patients.

Principal Component and Cluster Analysis
The results of quality attributes, chemical composition, and sensory characteristics were subjected to multivariate analysis by Principal Component Analysis (PCA) and Cluster Analysis (CA).In these analyses, the following variables: protein and fiber content, browning index, water activity, force, overall acceptability and sensorial attributes (color, texture and taste), which allowed to discriminate the samples, were considered.
There were two main components that explained the total variance (Figure 4(a)).

Figure 1 .
Figure 1.Photographs and mean values of color differences (∆E) of biscuits containing ahipa flour in their formulations, submitted to different baking conditions.

Figure 2 .
Figure 2. (a) Color, taste and texture of biscuits with 250 g of ahipa flour kg −1 baked at 200˚C for 7.5 or 10 min.Data reported are the percentage (%) obtained for the different scores from sensory evaluation of cookies.The values of the hedonic scale were grouped into three ranges: ∎ 1 -3 (dislike), ∎ 4 -6 (not like nor dislike) and ∎ 7 -9 (like very much); (b) General acceptability histograms of biscuits baked at 200˚C for 7.5 or 10 min.

Figure 3 .
Figure 3. (a) Color, taste and texture of biscuits with 100% corn starch (control), 250 g of ahipa flour kg −1 (250AF) and 250 g of cassava flour kg −1 (250CF) baked at 200˚C for 7.5 or 10 min.Data reported are the percentage (%) obtained for the different scores from sensory evaluation of cookies.The values of the hedonic scale used were grouped into three ranges: ∎ 1 -3 (dislike), ∎ 4 -6 (not like nor dislike) and ∎ 7 -9 (like very much); (b) General acceptability histograms of biscuits baked at 200˚C for 7.5 or 10 min.

Figure 4 .
Figure 4. (a) Principal component analysis (PCA) and (b) cluster analysis of quality attributes, chemical composition and sensory characteristics of biscuits containing nutritionally differentiated root flours (cassava or ahipa).

Table 1 .
Texture parameters (maximum force and area under the curve) of biscuits containing ahipa flour in their formulations and submitted to different baking conditions.: Control (1000 g•kg −1 corn starch); 250AF: 250 g•kg −1 ahipa flour + 750 g•kg −1 corn starch; 250CF: 250 g•kg −1 cassava flour + 750 g•kg −1 corn starch.Reported values correspond to the mean ± standard deviation.Data followed with the same letter within a column did not differ significantly (p > 0.05). C