Digital Repository @ Iowa State University Analysis of Corn Distillers Dried Grains With Solubles (DDGS) / Flour Mixtures, and Subsequent Bread Baking Trials

Grains offer a variety of nutrients; it is thought that through the addition of distillers dried grains with solubles (DDGS) the fiber and protein in baked products may be improved. In this study, all-purpose flour and bread flour were tested with various DDGS substitution levels (0%, 25%, or 50% flour substitution) with the dough conditioner sodium stearoyl lactylate (SSL) (0%, 0.15%, or 0.3% flour weight basis). SSL is surfactant produced from reacting stearic acid with food grade lactic acid in the presence of sodium. Overall, as the substitution level of DDGS increased, so did protein, moisture, ash, and Hunter-a values. Peak height, side height, width, and length of baked loaves decreased as DDGS quantity increased, however. Baked bread containing SSL had enhanced quality, with increasing dough strength, rate of hydration, mixing tolerance, crumb strength, loaf volume, and shelf life. Overall, 25-50% DDGS substitution appeared to have a negative effect on physical features despite the fact that the nutrient content was enhanced. Less than 25% DDGS had minimal negative effects on bread properties.


Introduction
Grains such as wheat, corn, rice, and oats have traditionally been major constituents in the human diet. Historically this has been due to intrinsic nutrients as well as functionality. Wheat is very versatile, as it forms a complex known as gluten, which contains 80% -90% glutenin and gliadin proteins (Lilliard, 2000;Wang et al., 2004). Many factors impact gluten structure and function in food systems. Gluten proteins must be hydrated and then physically manipulated to form both the gluten complex and ultimately the starch-protein matrix. Gluten offers structure to baked products, such as breads, by trapping fermentation gases thus allowing dough to rise. Weak gluten structure can result in excessive expansion and uneven texture, whereas too strong a gluten structure results in decreased expansion and low loaf volume (Schofield & Booth, 1983). Ultimately, protein content and quality impact bread quality and structure.
Not only does the preparation process affect the final product, but ingredients used and flour quality also have effects. To overcome potential variations and improve quality, a variety of dough conditioners, hydrocolloids, and enzyme treatments are often used in products to overcome differences and deficiencies in flour qualities (Azizi & Rao, 2004). These additives have a macroscopic effect on dough by inducing structural changes to the flour . Hydrocolloids are water-soluble polysaccharides that provide various functional properties, including gelling, thickening, emulsifying, stabilizing, and foaming; they can also be used for syneresis inhibition, water retention, and improving textural properties. Guar gum is known for enhancing bread loaf volume and texture (Ribotta et al., 2004). Hydroxypropylmethylcellulose (HPMC) creates a softer bread crumb, increased loaf volume, enhanced shelf life, and improved sensory characteristics (Barcenas & Rosell, 2005;Collar et al., 1998;Rosell et al., 2001). Other examples include K-carrageenan, sodium alginate, and xanthan gum, all of which have been found to retard crumb hardening of bread (Guarda et al., 2004). Biochemical mechanisms allow enzymes to favor covalent bonding of proteins, thus improving functional properties of breads (Caballero et al., 2006). A dough conditioner commonly used in industry is sodium stearoyl lactylate (SSL). SSL is a surfactant that is produced by reacting stearic acid (50-90%) with food grade lactic acid in the presence of sodium (Krog & Lauridsen, 1976). Surfactants increase the quality of baked bread by Once bread loaves are made, quality analyses can be completed by examining loaf weight, loaf volume, specific volume, and loaf height/width ratio of the central slice (Caballero et al., 2006;Shittu et al., 2007;Soulaka & Morrison, 1985). A texture analyzer can be used to quantify characteristics such as compression, penetration, three-point bending, and force-displacement. Other common tests for baked bread include crumb firmness, color indices (colorimeter), crumb moisture, crumb hardness (penetrometer), density, porosity, and softness (AACC, 2000). These tests give researchers quantifiable methods to quantify properties such as crispy, cracky, and hard (Pamies et al., 2000).
The aforementioned methods are essential for assessing the effectiveness of new ingredients. For example, distillers dried grains with solubles (DDGS) is an ingredient which can provide substantial protein and fiber to baked products, but to date has seen limited use in foods. DDGS is a cereal-based ingredient high in fiber but absent of gluten-forming proteins. Over the last few decades, many studies have been completed which have examined flour replacement with DDGS for a variety of food products. Tsen et al. (1982) used corn DDGS at 15% in molasses and spice cookies, and at 25% in chocolate chip cookies. Rasco et al. (1987) used wheat DDGS at 30% in chocolate chip cookies and banana bread. Abbot et al. (1991) used 36% untreated DDGS in oatmeal muffins. Reddy et al. (1986) used distillers dried grains (DDG) at 10% in wheat muffins and 15% in muffins that had blueberries or raisins added; DDG improved physical characteristics and increased consumer acceptance. Brochetti et al. (1991) used 5-10% DDG in bread, which resulted in acceptable sensory attributes; they also incorporated up to 15% DDG in bread, which was also acceptable, with the exception of differences in appearance compared to the control. Brewers' grains, similar to DDGS and DDG, have been successfully substituted at 6% (Finley & Hanamoto, 1980) and 10% (Tsen et al., 1982) into breads.
Depending on DDGS substitution levels, bread fermentation may be hindered due to dilution of gluten, thus altering the quality of final products. A study completed by Aamodt et al. (2005) which investigated bread characteristics as a function of protein quality indicated that flour blends with strengthened protein quality resulted in bread loaves with a larger loaf volume and bread slice area, and formed a higher height to width ratio. Rasco et al. (1990) hypothesized that wheat DDGS may have a reduced negative effect on gluten formation compared to other high-fiber ingredients. Wheat DDGS may have enhanced electrostatic interactions during protein hydration, which could improve gluten functionality. Additionally, high temperature post-fermentation treatments may soften the fiber in the DDGS, allowing faster hydration, which may decrease its inclination to puncture developing air cells.
found that DDGS produced using harsh heat treatments were less suitable for bakery products than DDGS produced using drum dryers at a lower temperatures. Food matrices using the harsh drying were of significant poorer quality. Temperatures during thermal processes such as extrusion (100-190 °C), flaking (30-150 °C), biscuit or cracker processing (100-140 °C), and commercial baking (190-260 °C) affect the resulting physical and functional properties (Hansen et al., 1975;Donovan, 1977;Li & Lee, 1996). Hydration levels and processing temperatures can determine whether gluten or starch will contribute to bread structure or hinder it. High moisture systems (45-50%) are commonly used for products such as breads, noodles, pasta, and biscuits, and they utilize gluten to provide the majority of the structure (MacRitchie, 1992). Low moisture systems (< 35%), often used for extruded products, generally depend on starch for structure (Chanvrier et al., 2006).
In recent years, the ethanol industry has evolved, and most DDGS is now produced from corn using the dry grind manufacturing process. But, very little research has examined use of this ingredient in human foods. And no studies have yet been conducted to improve DDGS functionality using dough conditioners. Perhaps these can improve DDGS use in bread systems. Therefore, the objectives of this study were to 1) examine the effects of corn-based DDGS on functionality of wheat flour/DDGS blends, 2) examine the impacts of wheat flour/DDGS blends on bread properties, and 3) determine if SSL can improve DDGS functionality and improve breads made with this ingredient.

Flour Mixtures
DDGS for this experiment was acquired from a commercial fuel ethanol plant and was milled in a Glen Mills Inc. (Clifton, NJ, USA) mill, providing a mean particle diameter size of 0.384 mm. Using a factorial experimental design (Table 1), all-purpose flour and bread flour were tested with various DDGS substitution levels (0, 25, or 50% flour substitution) and various levels of the dough conditioner sodium stearoyl lactylate (SSL) (0, 0.15, or 0.3% flour weight basis). Independent variables were DDGS replacement (3 levels), SSL inclusion (3 levels), and type of flour (all-purpose vs. bread). Dependent variables included protein (% db, dry basis) using AACC method 46-30 (2000), lipid (% db) using method 920.39 (AOAC, 2003), ash (% db) using method 08-03 (AACC, 2000), and moisture (% wb, wet basis) using method 44-19 (AACC, 2000). All flours were acquired from a local grocery store. Mixtures of the various treatment combinations were stored at 20.5 ± °C until use. Physical parameters of the flour were analyzed using the following Mixolab properties: amplitude (Nm), stability (s), water absorption, α (Nm/min), β (Nm/min), and γ (Nm/min). The Chopin+ protocol's total analysis time was 45 minutes and it utilized an 80-rpm mixing speed, 1.1 Nm target torque (for C1), 30 °C tank temperature, and 75 g dough weight.

Baking Trials
Two bread machines (OSTER® 2 lb. EXPRESSBAKE, Sunbeam Corporation, 1999) were used for the bread baking experiment to reduce baking differences and decrease error. The machines had the versatility of 1.5 lb or 2 lb loaves, 18 bread settings, 3 crust color selections, delay baker timer, cool touch exterior and Express Bake cycle options (60 min). The baking pan was a non-stick, oval aluminum bread pan with the following dimensions: 18 cm length, 13.6 cm width, and 13.2 cm height. A hook attached to the bottom mixed and kneaded the dough.
Ten flour mixtures were baked (replicated twice), for a total of 20 loaves of bread. For this experiment, the machines were adjusted to the "Basic" bread making setting (1.5 lb loaf), with a medium crust color, which allowed the breads to be baked in 3 hours. The following formulation was used to produce each 1.5 lb loaf: 240 g cold water (17 °C), 20 g vegetable (soybean) oil, 30 g sugar, 10 g salt, 400 g flour mixture, and 8 g quick acting dried yeast. Liquid ingredients were added first, dry ingredients second, and yeast last. Bread machine stages were as follows: 10 min dough was first kneaded, 20 min dough begins to rise, 15 min dough was kneaded a second time, 20 min dough continues to rise, 30 sec dough was "punched down", 55 min dough rises final time, 60 min bread bakes (Sunbeam Corporation, 1999). During the process, the bread loaves with 25% or 50% DDGS replacements were closely monitored to ensure the doughs formed appropriate balls.
Proximate composition of the baked breads included protein (%db) using method 46-30 (AACC, 2000), ash (%db) using method 08-03 (AACC, 2000), and lipid (%db) using method 920.39 (AOAC, 2003). Physical properties of the bread included side height (height of sides of loaves until the top starts to mound, cm), peak height (height of loaves to the top of the mound (i.e. at center), cm), width (cm), length (cm), mass (g). Moisture (%wb) was determined using method 44-19 (AACC, 2000). Water activity was determined using a water activity meter (AW Sprint). Strength (MPa) and stiffness (MPa) were determined using an Instron testing machine. Hunter L, a, and b values, for both internal crumb and exterior crust color, were measured with a Hunter spectrocolorimeter (LabScan XE, Hunter Associates, Reston, VA, USA).

Statistical Analysis
The completely randomized, factorial experimental design was a 2 × 2 × 2 with additional 2 center points for a total of 10 treatment combinations replicated three times to total 30 runs (Table 1). Statistical analyses on all collected data were performed via SAS v.8.0 (SAS Institute, Cary, NC) and Microsoft Excel v.2003 (Microsoft Corp., Redmond, WA) software, using a Type I error rate (α) of 0.05, and included summary statistics, general linear models to test for differences between experimental treatments (i.e., main, interaction, and treatment effects), and linear correlations. Table 2 shows the main effects on proximate composition of all ten flour mixtures. The quantity of DDGS replacement resulted in significant differences in protein, lipid, ash and moisture. The percentage of sodium stearyl lactylate (SSL) included resulted in significant differences for protein, lipid, and moisture. Finally, the type of flour used resulted in significant differences in protein, ash, and moisture. Treatment combination effects were significant as well ( Figure 1). Figure 1. Treatment combination effects on ash, moisture, and protein due to flour type (1 is all-purpose flour, 2 is bread flour), dough conditioner levels, and DDGS levels. (A); between moisture content, flour, and DDGS levels (B); between moisture content, flour, and SSL levels (C); between protein, flour, and DDGS levels (D); and between protein, SSL, and DDGS levels (E). Overall, as DDGS levels increase the quantity of protein and ash increase while moisture levels decrease. Error bars represent ± 1 standard deviation  There was a statistically significant increasing trend in ash content as the percentage of DDGS substitution increased. Tsen et al. (1983) also found that the addition of DDG into flour formulations increased ash content compared to white flour. Ash content is heavily impacted by environmental factors such as the fertility of the soil and genetics of the plant (Fennema, 1996). Other constituents may include quantity of sunlight received, access to water, and stage of growth during harvest. Flour mixtures that had 0% DDGS had a mean ash content of 0.66%, 25% DDGS mean ash content of 1.63%, and 50% DDGS had a mean ash content of 2.54%. This was reasonable, as the addition of DDGS, a fibrous material, to flour resulted in more nutritive and non-nutritive residues left after ashing.

Chemical Properties
As the quantity of DDGS increased, moisture levels decreased. In fact, the initial moisture content of DDGS (12.9%) had a higher percentage of water compared to the all-purpose (7.9%) or bread flours (7.7%). Therefore, it may be logical to assume that as the percentage of DDGS (which contained greater initial moisture), replaced more all-purpose or bread flour (which were drier than the DDGS), the flour mixture moisture content would increase. DDGS also contains a high level of fiber, which can readily absorb free water. Dreese and Hoseney (1982) concluded that products high in fiber also had increased quantity of water absorption. Therefore, in order to achieve an ideal product, the quantity of liquid included in the formulation may need to be increased. Once the fiber molecules are softened, the fiber should incorporate more easily into the dough system. Fiber plays many roles in food systems, such as providing structure and bulk, modification of rheological properties, as well as other functions (Fennema, 1996). Several studies have advocated the idea that DDGS would provide an excellent fiber supplement for baked products (Brochetti et al., 1991;Waelti & Ebeling, 1982;Wu et al., 1984;Rasco et al., 1987).
The quantity of moisture slightly increased as the SSL in bread flour mixtures increased. SSL is a food additive, which is categorized as a surfactant. It is amphiphilic, and exhibits both hydrophilic and hydrophobic properties. Its major function is to reduce interfacial tension between two fluids. Other functional properties of surfactants include emulsifying, foaming, solubilizing, wetting, and dispersing (Lilliard, 2000). The bakery industry uses surfactants to improve dough-mixing qualities and to enhance loaf volume. Azizi and Rao (2004) completed a study that analyzed the combination of surfactant gels and gums on dough and bread systems. Farinograph results revealed that surfactant gels actually decrease a flours ability to absorb water by 0.4-1.2%. Also, alveograph characteristics showed a 0.4-1.6 mL increase in the swelling index of dough's using surfactants. Less variation in moisture content was shown by all-purpose flour mixtures and likely due to a decrease in water absorption. Moisture content increased with increased SSL levels. The bread flours slightly increased swelling ability may explain this slight increase in moisture content. Perhaps additional water was captured into the food matrix and was not able to be dissipated as steam during baking. For the most part, the function of the surfactants is to reduce interfacial tension in food systems, decreasing the amount of water that flour can absorb. The protein  (Kahraman et al., 2008).
However, this eventual increase in consistency is mostly contributed to the beginning of stage C3, starch gelatinization. The quality of starch determines the rise in consistency. Starch granules absorb water and swell, while viscosity increases due to leaching of amylose (Kahraman et al., 2008). Starch gelatinization results in peak torque during the heating stages, comprising initial and final pasting . Gelatinization occurs when amylose and/or amylopectin (insoluble in cold water) molecules are suspended in water, producing a starch slurry. Once heating begins, water becomes absorbed and granules hydrate. Continued heating weakens hydrogen bonds as molecules swell resulting in irreversible changes to the starch structure and eventually some granules will burst (White & Johnson, 2003).
In stage C4, endogenous or amylastic activity results in minimum torque during the cooling stage when reaching stability . This activity ultimately determines how large the decrease in consistency will be. Larger decreases are proportional to a greater quantity of amylastic activity (Anonymous, 2005;Kahraman et al., 2008). Two types of amylases include α-amylase and β-amylase. Alpha amylase hydrolyzes interior α-1,4-glucosidic bonds of starch, glycogen, and cyclodextrins. These enzymes are endo-splitting which acts to increase viscosity. Beta-amylase hydrolyze the α-1,4-glucosidic bonds of starch beginning at the non-reducing end to result in β-maltose. Since they are exo-splitting, many bonds need to be hydrolyzed before a significant impact on viscosity can be seen (Fennema, 1996).
Finally, stage C5, gel formation, which is related to retrogradation, causes an increase in consistency among the dough as the temperature decreases for the cooling stage (Kahraman et al., 2008). This cooling stage allows the starch to retreat and thus increase product consistency. Staling may be delayed through the addition of certain chemicals that enhance product flexibility (Anonymous, 2005). Retrogradation occurs after maximum viscosity is reached, in which some granules have broken or burst. Upon cooling of the starch solution, some starch granules will partially reassociate to form a gel. The retrogradation of amylopectin requires more time than amylose, giving it the distinction for causing part of staling (Fennema, 1996).
The Mixolab collects additional properties, comparable to the Farinograph (Brabender Instruments, New Jersey). A Farinograph measures flour and dough indices from the first 8 min of a sample run, whereas the Mixolab further analyzes flour's performance throughout the entire bread making process, including the heating and cooling phases. These extra phases have the potential to tie research and industry together, as the curve can relate differences between flours and traditional industrial baking conditions. The Mixolab instrument aids in explaining baking performance differences due to starch-protein interactions, enzyme activity, gelatinization, gelling of starch, and environmental factors (Saunders et al., 2007). A study completed by Kahraman et al. (2008) tested the possibility of using the Mixolab versus Zeleny sedimentation or Alveoconsistograph values to predict cake-baking quality of various flours. Results showed that Mixolab analysis proved to be a useful tool whereas the Alveoconsistograph results and cake characteristics were not significant. Overall, the Mixolab technique has been viewed as comparable to classical instruments such as the Farinograph, Mixograph, Extensograph, or Alveograph (Bloskma & Bushuk, 1988;Chiotelli et al., 2004). Table 3 illustrates the main effects for the Mixolab analysis of the flour mixtures. The percentage of DDGS substitution resulted in significant differences among amplitude, stability, water absorption, protein breakdown (α), and cooking stability rate (γ). The percentage of SSL resulted in significant differences among water absorption, α, and gelatinization (β). Finally, the type of flour included in the flour mixtures resulted in significant differences among stability, water absorption and γ.
Treatment combinations were also significant ( Figure 3). As the percentage of DDGS increased, the stability of the dough decreased. Since DDGS material contains no gluten forming proteins, it makes sense that the dough system would have less stability. A slight increasing trend showed that as the quantity of DDGS increased, water absorption also increased. An increased amount of water was absorbed due to the increased levels of fiber provided from the DDGS. Fiber required additional water in order to soften and to be incorporated into a dough ball. Also, as the quantity of DDGS increased, so did γ values, except for the 0.15% whose values were not available through data analysis. γ values are indicative of the rate in which a dough system reaches cooking stability. As DDGS increased, it took less time to stabilize the cooking process, which caused γ rate values to increase. This might influence the acceptance of DDGS in industry, as less time is required to reach stability Collar et al. (2007) defined α, β and γ on a Mixolab curve as ascending and descending curves to be protein breakdown, gelatinization, and cooking stability rate, respectively. Table 3 illustrates these values found in the treatments, and Figure 4 shows respective Mixolab curves. The bowl temperature was greater than dough temperature because a heating element was applied to the mixing bowl, which simulates the baking process, thus the bowl temperature was always higher than the dough temperature.
α values that are higher are indicative of protein breakdown, or weakening of the flour mixture. C2 curves that decrease less could be good indicators of protein quality. As the quantity of DDGS increased, the slope at which protein breakdown occurred also increased by the α values becoming more negative. Even though gluten forming proteins decreased as DDGS increased, the DDGS significantly adds other proteins to the flour mixture, thus there is more protein in the system to potentially breakdown. The bread flour mixtures had a higher initial protein content than the all-purpose mixtures, thus showing slightly increased protein weakening. Bread flour curves demonstrate a slightly lower decreased slope and curve than all-purpose flour.
A B C Figure 3. Treatment combination effects on stability, gamma, and water absorption between due to flour (1=all-purpose & 2=bread) and DDGS levels (A); between γ, flour, and DDGS levels (B); between water absorption, SSL, and DDGS levels (C). γ values were not produced for the center points (25% DDGS) during the Mixolab analysis. Overall, as the quantity of DDGS increases water absorption and γ values increase while time to reach stability decreases. Error bars represent ± 1 standard deviation   Positive β values are good indicators of starch gelatinization. However, if the β values become too high it is possible that the starch slurry may become over-gelatinized. Overcooked starch granules tend to swell and burst resulting in a complete loss of crystallinity and birefringence. Such starch molecules are increasingly susceptible to hydrolysis via enzymes (Nielsen, 2003). On the other hand, lower β values indicate little to no starch gelatinization or slower rate. This would then indicate that the flour does not contain enough reducing sugars. Not all flour mixtures will complete sufficient starch gelatinization to become viscous. Even though they may not be used in some baked products, industry can find other uses for thin starch slurries (e.g., in pie fillings). Figure 4 shows that as the percentage of DDGS increased, the quantity of starch gelatinization decreased. DDGS is low in sugar and starch, as these were used for ethanol production. DDGS is low in reducing sugars necessary for starch gelatinization. γ values indicate the slope and the rate that cooking stability is reached. As the quantity of DDGS increased, γ values decreased, indicating less time needed to reach cooking stability. This may be an industrial advantage, in that less time needed to reach cooking stability could be favorable in a bakery setting. Table 4 showed the main effects of the independent variables on the proximate composition of the bread loaves.

Chemical Properties
As the percentage of DDGS replacement increased, significant differences were found among protein, ash, and lipid values. A significant difference was also found between type of flour and protein. As DDGS substitution increased, so did protein content. This reaffirms the study completed by Reddy et al. (1986), which showed an increase in protein content of muffins as DDGS substitution increased. The addition of proteins may have strengthened the food system by offering extra proteins to promote functional tasks such dispersibility, swelling, water-holding capacity, gelation, and viscosity (Fennema, 1996). Table 5 shows the main effects of the independent variables on the physical properties of the bread loaves. As the quantity of DDGS increased, significant differences could be found in strength, water activity, a-crust, b-crust, L-crumb, a-crumb, b-crumb, side height, peak height, width, length, and mass. As the percentage of SSL increased, significant interactions between strength, a-crust, L-crumb, a-crumb, peak height, width, and mass were found. Finally, the type of flour showed significant differences among strength, water activity, a-crumb, side height, peak height, and length.

Physical Properties
Treatment combination effects were also significant ( Figure 5). Hunter a crust values decreased as the quantity of DDGS and SSL increased. As the quantity of red pigments decrease, the product becomes more brown. As the level of DDGS increases, it appears that the Maillard reactions and carmelization during baking are factors that may contribute to browning. Hunter a values decreased as did Hunter L values for the crust color, also indicating the presence of Maillard browning. SSL may also contribute to the decrease in Hunter a values as the dough conditioner may alter the way molecules interact during baking. Hunter a crumb values significantly increased from 0% to 25% DDGS, which indicated more red pigments. All in all, DDGS has more initial red pigments than found in all-purpose or bread flour. Therefore, as the level of DDGS increased, the Hunter a value of the interior crumb should become redder as well. These values did not decrease as the Hunter crust values do, because the interior crumb was protected and encountered a decreased amount of Maillard browning compared to the crust. Rasco et al. (1990) determined how the addition of a variety of distillers grain products from wheat and barley would affect mixing and baking properties of breads and cookies. Breads included 4 or 8% of various types of distillers grains, while the cookies substituted 2, 4, or 8%. Color analysis of bread loaves crust showed that almost all loaves had decreased Hunter a values. Loaves were found to become greener as the level of DDGS substitution increased. Color analysis of interior structure of these loaves revealed that all loaves exhibited increased Hunter a values. The interior structure became notably darker and redder as the level of DDGS increased. These trends found by Rasco et al. (1990) parallel the trends found in our study.
An increase in dark pigments (due to the addition of DDGS) has also been found in a study completed by Brochetti et al. (1991). Breadcrumb color darkened and Hunter L values decreased as DDG level increased. Our breadcrumb and crust Hunter L values also showed a decreasing trend as DDGS substitution increased. Hunter L values are indicative of product brightness, as a value of 0 indicates black and 100 indicates white. The DDGS added was "golden yellow" from dark pigments mostly located in the corn lipids. Darker colors may also have been due to heat (during drying), which caused undesirable Maillard browning. This type of browning is a chemical reaction between reducing sugars (i.e. D-glucose) and a free amino acid or amino group (Fennema, 1996).  Figure 5. Treatment combination effects on physical properties due to flour type, SSL, and DDGS levels (A); between crumb Hunter a values, flour (1=all-purpose & 2=bread), and DDGS levels (B); between peak height, Flour, and DDGS levels (C); between side height, flour, and DDGS levels (D); between width, SSL, and DDGS levels (E); between length, flour, and DDGS levels (F); and between peak height, SSL, and DDGS levels (G). Overall, as DDGS levels increase red pigments in the crumb also increase while peak height, side height, and length decrease. As SSL increases red pigments in the crust, width, and peak height decrease. Error bars represent ± 1 standard deviation  1 for a given main effect, differing letters between levels for a given property signify significant differences (p < 0.05, LSD); St. Dev. is +/-1 standard deviation.  1 for a given main effect, differing letters between levels for a given property signify significant differences (p < 0.05, LSD); St. Dev. is +/-1 standard deviation.  1 for a given main effect, differing letters between levels for a given property signify significant differences (p < 0.05, LSD); St. Dev. is +/-1 standard deviation.
Peak height and side height decreased as the percentage of DDGS increased. Rasco et al. (1990), also found a decreased loaf volume on the majority of bread loaves with 4 or 8% distillers grains substitution. O'Palka et al.
(1989) completed a study of baked products containing sour mash corn distillers dried grains. Three of four baked products with various distillers substitution levels resulted in decreased loaf volumes: dinner rolls (17 or 33%), nut rolls (33%), and carrot coconut bread (40%).
The decrease in volume was thought to have been due to the dilution of gluten, particularly in yeast-leavened products, along with other factors that accompany the addition of DDGS to wheat dough systems. Increased concentrations of fiber results often absorb more water (Dreese & Hoseney, 1982). Fiber molecules can be incorporated more easily into a food system once the fiber molecules are softened. Fiber plays many roles in a food matrix, such as providing structure and bulk, as well as modifying rheological properties (Fennema, 1996). If fibrous materials are not softened via hydration, the structure may be compromised due to the cutting of gluten strands, thus diminishing inflation and structure. In this study, width decreased as the percentage of SSL and DDGS increased. The SSL's purpose was to condition and facilitate dough quality; however, increased quantity of DDGS had an opposite effect, altering bread functionality. Length of bread loaves decreased as the quantity of DDGS increased, which reinforced that higher replacement rates of DDGS impacted size and shape. Peak height decreased as SSL and DDGS substitution increased. This showed how bread loaves had a higher peak height when 0% SSL was used compared to 0.15% SSL, which indicated that 0.15% was not a significant quantity to make much difference in enhancing dough and loaf properties. Brochetti et al. (1991) analyzed bread with 5, 10, and 15% DDG substitution. Results indicated that at a 15% DDG substitution level, bread loaf volume also decreased. This volume decrease is reflected in decreased peak height, side height, length, and width parameters. Tsen et al. (1983) and Morad et al. (1984) have also found similar results. The more DDGS included into the bread, the less wheat flour, which dilutes the amount of gluten proteins available. It is also possible that increased water absorption can negatively affect the ratio of wheat flour-to-water (3:1), thus preventing ideal gluten formation (Fennema, 1996). Additional water may disrupt interactions that can bond the structure together. Bread loaves with higher DDGS content may have been manipulated past maximum resistance, thus resulting in a decrease in resistance, which breaks down the gluten structure. Finally, the addition of albumin-and globulintype proteins adversely affects bread volume and gluten structure. Therefore it is important to monitor the addition of these proteins into baked products (Fennema, 1996).
Height, shape, and color differed among the flour treatment combinations ( Figure 6). Additional images showed the grain quality, color, size, and structure of center cut slices ( Figure 7). As the quantity of DDGS replaced increased, cell structure of loaves became increasingly compact, dense, and thicker.  Table 6 shows results for the subjective baking quality analysis. As the percentage of DDGS included in the bread formation increased, significant differences could be found among uniformity, size, thickness, grain, tenderness, softness, and crumb color. No significant differences were found between the percentage of SSL or the other independent variables. The type of flour had a significant effect on softness. Figure 8 illustrates the effects on softness and DDGS substitution. As the percentage of DDGS increased, softness decreased. This was expected as the DDGS was highly fibrous, which has previously been shown to alter the grain structure of the food matrix. Brochetti et al. (1991) tested sensory characteristics of breads with DDG replacement. Breads with 10% and more DDG substitution had a less uniform cell distribution and harsher texture than bread that was considered "ideal". Their findings indicated adverse effects of increased DDG concentrations that affected sensory and textural characteristics. They also found that as DDGS replacement increased, the several properties were negatively affected, including cell uniformity, cell thickness, grain condition, texture tenderness, texture softness, and crumb color. This issue could possibly be fixed if DDGS refinement processes could result in extremely fine DDGS flour. Figure 6. Initial loaf profiles show differences in shape, color, and texture between experimental treatments. As the quantity of DDGS increased physical dimensions such as loaf volume, height, and length decreased. DDGS substitution resulted in a loaf color that was darker as well  1 for a given main effect, differing letters between levels for a given property signify significant differences (p < 0.05, LSD); St. Dev. is +/-1 standard deviation.  1 for a given main effect, differing letters between levels for a given property signify significant differences (p < 0.05, LSD); St. Dev. is +/-1 standard deviation.

Correlations and Multivariate Analysis
Linear correlations were determined (Table 7), these are not indicative of causation, nor do they represent relationships found in non-linear behavior. As bread mass increased so did bread protein (r=0.863), flour protein (r=0.864), bread ash (r=0.901), flour ash (r=0.893), bread lipid (r=0.856) and flour lipid (r=0.866). These chemical properties obviously impacted the final physical condition of bread loaves. Flours that are rich in nutrients may produce increasingly dense loaves of bread. Water absorption (r=0.804) of the flour also increased as bread mass increased. This may indicate an increased quantity of water retained and not released as steam during baking processes. Also, water absorption increased as bread protein (r=0.953) and flour protein (r=0.967) increased, indicating protein molecules may be increased building to hydration. Results also show Hunter a and b values were positively impacted by chemical properties as well. These values increased as the quantity of bread protein (a values r=0.835 and b values r=0.878), bread ash (a values r=0.878 and b values r=0.886), and bread lipid (a values r=0.820 and b values r=0.844) increased.
As bread protein content increased, stability in the dough system decreased (r=-0.855). As the replacement rate of DDGS increases, more protein is added to the food matrix. DDGS has many factors, particularly fiber, that may prevent the coproducts from being easily incorporated into dough. Also, as the content of flour ash increases, softness decreases (r=-0.973). This decrease in softness may be due to the increased quantity of nutrients and non-nutrient residues, mostly increasing from rising substitution of DDGS.
In terms of multidimensional space, Principal Components Analysis was used to examine the data distribution when considering all independent and dependent variables simultaneously ( Figure 9). Overall, the covariance matrix appeared to be better suited to examine this data set, as it required fewer components (2 vs. 10) to summarize the variance in the data. As shown, three variables were most influential, and thus had the greatest influence on the data, including DDGS level, bread mass, and stability. It also appeared that when considering the first two principal components there was clustering in the data, and this was specifically due to the DDGS level.  Figure 9. Principal components analysis of data indicates that, using the correlation matrix, A) less than 10 variables had significant effects in multi-dimensional space, B) many components were interrelated and influenced the data, and C) clustering according to DDGS level occurred. Using the covariance matrix, on the other hand, indicated that D) very few factors were required to explain the variance in the data, E) the data could be explained by DDGS, bread mass, and stability, and F) clustering according to DDGS level was observed in multivariate data space