Evaluation of Oat Okara Sourdough for Its Potential Uses in Bread Making

Abstract

The growing over-75 population has increased the demand for functional foods tailored to the nutritional needs of the elderly. Within the AURA project, an innovative oat okara sourdough was developed to produce bread with enhanced nutritional and functional properties. Breads were produced using oat okara sourdough, oat sourdough, and wheat sourdough for comparison. All samples were subjected to microbiological, physical-chemical, technological, and metabolomic analysis. In addition, bread digestibility was evaluated. The results showed that oat okara flour is an excellent fermentable substrate, yielding sourdoughs with high counts of lactic acid bacteria and yeasts. The breads made with oat okara and oats were softer and brownish due to the oat presence and higher relative yeast. Moreover, oat okara bread exhibited a lower proportion of rapidly digestible starch (RDS) and a higher proportion of slowly digestible starch (SDS), suggesting potential benefits for post-prandial glycaemic control. Metabolomic profiling highlighted lipids, particularly steroidal glycosides (saponins) and fatty acyls, as discriminant metabolites in fermented samples, suggesting enhancement of bioactive compounds through sourdough fermentation. Overall, the use of oat okara in sourdough represents a sustainable approach to upcycle agro-industrial by-products while producing nutritionally valuable bakery products aligned with circular economy principles.

1. Introduction

Recently, increasing consumer demand for food products with improved nutritional quality and health benefits that also offer good sensory properties has posed new challenges to the baking industry. In this context, sourdough has gained renewed attention as a traditional bread-making method involving the natural fermentation of a mixture of flour and water [1]. It improves the nutritional value, taste, aroma, texture, and storage stability of bread but also induces changes in carbohydrate composition [2]. The bread-making process is highly dependent on the type of flour, the fermentation agent, and the fermentation time. These conditions may influence both bread quality and nutritional composition [3]. The fermentation process of sourdough involves the back-slopping technique. In this method, a small amount of the original fermentation product is used as a starter culture during the subsequent fermentation, promoting the synthesis of organic acids, enzymes, antifungal compounds, exopolymers, and saccharides, as well as proteolysis [4]. Mature sourdough is dominated by a microbial consortium comprised mainly of lactobacilli, obliged and/or facultative heterofermentative, and yeasts [5] that produce new nutritionally relevant compounds, such as peptides, amino acid derivatives, and potentially prebiotic exopolysaccharides, through their metabolism [6,7]. In this respect, for the health benefits of the final product, it is crucial that the positive nutrients are bio-accessible and bioavailable [8,9].

Fermentation is also widely used to recover food industry by-products that would otherwise go to waste. By using microorganisms, it may be possible to valorise by-products and implement low-cost bioprocesses, obtaining functional ingredients that can be used to produce traditional [10] as well as innovative foods [11]. The production of vegetable-based beverages, such as those made with oats, generates residues generally consisting of the insoluble part of the starting material: fibre, lipids, protein, and ashes. The high content of fibre and bioactive compounds, such as polyphenols and amino acids, makes these residues interesting by-products for use in the food sector. For each kilogram of oat-based beverage produced, approximately 0.85 kg of wet oat residue is generated, corresponding to 17–34% on a dry matter basis [12,13].

Oats (Avena sativa) are a nutritionally attractive cereal, and their use in food products has grown significantly in recent years due to the absence of gluten and high presence of dietary fibre, protein, vitamins, minerals, unsaturated fatty acids, and phenolic compounds [14,15]. While the absence of gluten protein makes oat-based products accessible to the celiac population, their application in baking remains challenging [16,17]. This is due to specific starch properties and the tendency of lipids to form off flavours. In fact, oats pose a higher risk for acrylamide formation than other cereals, such as wheat, due to their elevated levels of reducing sugars, amino acids (such as asparagine), and lipids. Okara, a by-product of vegetable beverage production, most commonly from soybeans, is rich in valuable components such as proteins, fibres, and lipids [18,19]. Oat okara is especially promising due to its gluten-free nature and the presence of beta-glucans, which may act as a natural alternative to commercial hydrocolloids [20,21], highlighting that oats are a rich source of high-quality protein, dietary fibre, and phytochemicals, offering a more balanced nutritional profile than animal-derived products. Oat protein is also emerging as a valid alternative to soy protein, with a quality profile like that of hulled soy flour [22]. Additionally, unlike many cereal prolamins that can trigger allergic reactions due to their high proline and glutamine content, oats predominantly contain globulin, which lacks allergenic potential, as their main storage protein [23].

Given the by-products of oat beverages, the formation of a sourdough and subsequent bread production could reduce acrylamide levels, because most of these molecules end up in the plant-based drink and therefore do not persist in the by-product. Furthermore, fermentation could also be an effective technique to reduce the accumulation of these substances, which are consumed by microbial metabolism. Therefore, considering that the microflora of sourdough and the production processes of leavened products are strongly influenced by the raw material used, it can be hypothesised that oat okara sourdough, with its ecological niches and nutritional composition, could produce quality baked goods with desirable nutritional characteristics.

To date, research on oat-based sourdoughs and okara derived from plant-based beverages has mainly addressed isolated technological or composition aspects, while an integrated evaluation is still lacking.

Following our previous work [24], traditional sourdough fermentation based on repeated back-slopping was adopted to obtain a stable sourdough system. This approach relies on the selection and adaptation of autochthonous microorganisms, which are known to enhance ecosystem robustness and process stability in Type I sourdoughs, according to the model described by De Vuyst et al. [25].

The aim of the present research was to evaluate the potential of a sourdough produced from oat okara flour for application in baked goods, with a focus on its rheological and functional properties. To gain a comprehensive understanding of its behaviour, the oat okara sourdough was compared with both wheat-based and oat-based sourdoughs. The study included a multidisciplinary set of analyses: microbiological profiling to assess the development of the sourdough microbiota, rheological tests to evaluate dough structure and performance, in vitro digestion assay to simulate gastrointestinal conditions, and metabolomic analyses to investigate biochemical transformations during fermentation.

2. Materials and Methods

2.1. Raw Materials

Oat okara flour, supplied by Packtin srl (Reggio Emilia, Italy), was used to produce the oat okara sourdough. According to the manufacturer’s specifications, the nutritional composition of the oat okara flour is as follows: energy 332 kcal, fat 3.93 g (of which saturated 0.71 g), carbohydrates 19.5 g (of which sugars 13.2 g), fibre 22.7 g, and protein 43.3 g. Two additional sourdoughs were prepared using oat flour and common white wheat flour type 00, both purchased from local markets. Buns were then produced by mixing each sourdough with flour, salt, extra virgin olive oil, and white refined sugar, all of which were also purchased from local markets.

2.2. Sourdough Preparation

The oat okara flour (OOF) sample was used to produce the oat okara sourdough (OOS), with only tap water added. The mixture of water and flour was prepared in a 2:1 proportion. The control sourdoughs (oat sourdough, OS; wheat sourdough, WS) were prepared with approximately the same OOS texture characteristics. The proportions used were as follows: oat sourdough with a water-to-flour ratio of 1:1, and wheat sourdough with a water-to-flour ratio of 2:1. Sourdough ingredients were mixed with a TK20 kneading machine (Tekno Stamap, Vicenza, Italy) for 5 min and kept at room temperature (25 ± 2 °C) for 24 h, enabling spontaneous fermentation. Sourdoughs were refreshed daily over a period of 30 days. At each refreshment step, a defined portion of fermented dough was used as an inoculum and mixed with fresh flour and water at the same formulation ratios described above.

2.3. Bread Making

Three different bread doughs were investigated in this study, obtained using oat okara sourdough (OOD), oat sourdough (OD), and wheat sourdough (WD). The doughs were prepared by mixing the ingredients in the proportions reported in

Table 1. Recipe formulations reporting the amount (g) used of each ingredient in the bread dough making.

Ingredients	OOD	OD	WD
Oat okara sourdough	250	-	-
Oat sourdough	-	250	-
White wheat sourdough	-	-	250
White wheat flour	500	500	500
Water	205	205	205
EVO oil	20	20	20
White sugar	20	20	20
Salt	5	5	5

using a professional mixer (Tk20 Tekno Stamap, Vicenza, Italy) until a homogeneous dough was obtained. The dough was leavened at 30 °C for 2 h (Polin Proofer, Verona, Italy) and then divided to form buns weighing 80 g each. Samples for subsequent analyses were taken every 30 min from the start of levitation: T0, T30, T60, T90, and T120. Finally, the bread doughs were baked at 190 °C for 27 min in a professional oven (Polin Wind 6040/5, Verona, Italy).

2.4. Microbiological Analyses

Microbiological analyses were performed on the sourdoughs (OOS, OS, and WS) and the doughs (OOD, OD, and WD) at various fermentation times (T0, T30, T60, T90, and T120). A total of 10 g of each sample was homogenised in 90 mL of physiological water (NaCl, 9 g/L) using a Stomacher machine (400 Circulator, International PBI, Milan, Italy) at 260 rpm for 120 s; this step was repeated twice. The decimal dilutions were executed and then plated on the following agar media: for yeast and mould counts, Rosa Bengala (RB, Oxoid, Italy) with the addition of chloramphenicol 0.01% (w/v) (Boehringer Ingelheim, Germany) was used, and the plates were incubated for 5 days at 25 °C; meanwhile lactic acid bacteria (LABs) were counted on MRS Agar (Oxoid) supplemented with cycloheximide 1% (w/v) (Oxoid), with the plates incubated at 37 °C for 48 h in anaerobic conditions using Anaerocult A (Merck, Darmstadt, Germany). Microbiological analyses were performed on three replicates for each sample.

2.5. Physical-Chemical Analyses Chemical–Physical Analyses

The physical-chemical analyses included measuring the pH and water activity (a_w) of all sourdough, dough, and bread samples. pH was measured with a HANNA pH meter HI-2202 Edge^®blu (Hanna Instruments, Padova, Italy) by putting the electrode directly in contact with each sample. pH measurements were done on the fresh doughs immediately after their preparation and once every 30 min until reaching 2 h leavening, as well as on the breads after cooling. For each sampling, three separate pH measurements were performed after calibration with commercial standard buffer solutions (pH 4.01, pH 7.00, and pH 9.21) and temperature compensation. Water activity (a_w) was measured using the Aqualab 4TE device (Meter Group, Munich, Germany). About 1 g of each sample was placed inside a small plastic holder. The analysis was carried out in triplicate, and each sample was analysed three times on each measurement.

2.6. Technological Evaluation

Samples were monitored in terms of moisture content, colour, texture, and viscoelasticity.

2.6.1. Moisture Content

Moisture content was determined following the AOAC official method 931.04 (AOAC, 2005). In detail, 3 g of fresh dough or bread was weighed in a crucible and maintained at 105 °C for 24 h up to a constant weight value. The moisture content percentage was evaluated by Equation (1):

$$\mathit{M}\mathit{o}\mathit{i}\mathit{s}\mathit{t}\mathit{u}\mathit{r}\mathit{e} \%={\displaystyle \frac{{\mathit{M}}_{\mathbf{0}}-{\mathit{M}}_{\mathbf{1}}}{{\mathit{M}}_{\mathbf{0}}}}·100$$

(1)

where M₁ represents the net weight (g) after 24 h in the oven and M₀ is the net initial weight of sample (g).

The dry matter content percentage was calculated through Equation (2):

$$\mathit{D}\mathit{r}\mathit{y} \mathit{M}\mathit{a}\mathit{t}\mathit{t}\mathit{e}\mathit{r} \%=\mathbf{100}-\% \mathit{M}\mathit{o}\mathit{i}\mathit{s}\mathit{t}\mathit{u}\mathit{r}\mathit{e}$$

(2)

Each measurement was carried out in triplicate.

2.6.2. Weight Loss

Bread weight loss after baking was measured by weighing breads before and after baking (parameters specified in Paragraph 2.3), using Equation (3):

$$\mathit{W}\mathit{e}\mathit{i}\mathit{g}\mathit{h}\mathit{t} \mathit{l}\mathit{o}\mathit{s}\mathit{s} \%={\displaystyle \frac{{\mathit{W}}_{\mathbf{0}}-{\mathit{W}}_{\mathbf{1}}}{{\mathit{W}}_{\mathbf{0}}}}·\mathbf{100}$$

(3)

where W₁ represents the weight of the baked bread (g) and W₀ is the weight of sample before baking (g).

The analysis was conducted using three replicates.

2.6.3. Specific Volume

The specific volume of the obtained bread was calculated with a geometric method. Given the flattened cylindrical shape and the fact that the bread had an oval cross-section, the formula for a cylinder with an elliptical base was used to calculate its volume, as shown in Equation (4).

$$\mathit{V}({\mathit{c}\mathit{m}}^3)={\displaystyle \frac{\mathit{\pi }}{4}}·\mathit{L}·\mathit{l}·\mathit{h}$$

(4)

where L is the length (cm), l is the width (cm), and h is the height (cm) of the obtained breads.

The specific volume was indeed calculated using Equation (5).

$$\mathit{Specific}\mathit{volume} ({\mathit{c}\mathit{m}}^{\mathbf{3}}/\mathit{g})={\displaystyle \frac{\mathit{l}\mathit{o}\mathit{a}\mathit{f} \mathit{v}\mathit{o}\mathit{l}\mathit{u}\mathit{m}\mathit{e} \mathit{o}\mathit{f} \mathit{b}\mathit{r}\mathit{e}\mathit{a}\mathit{d} ({\mathit{c}\mathit{m}}^{\mathbf{3}})}{\mathit{w}\mathit{e}\mathit{i}\mathit{g}\mathit{h}\mathit{t} \mathit{o}\mathit{f} \mathit{b}\mathit{r}\mathit{e}\mathit{a}\mathit{d} \left(\mathit{g}\right)}}$$

(5)

2.6.4. Colour Test

Colour evaluation of the fresh doughs was done using the Colorflex EZ colorimeter (HunterLab, Reston, Virginia), while the D25 NC colorimeter (HunterLab, Reston, Virginia) was employed for the measurements on the bread. Colorimeters provided the CIEL*a*b* colour space trichromatic coordinates: L* represents the lightness from 0 (black) to 100 (white); a* measures greenness (negative values, -a*) or redness (positive values, +a*); and b* measures blueness (negative values, -b*) or yellowness (positive values, +b*). Chroma (C*) and hue angle (h°) were also considered: C* represents the colour saturation and purity, and it is the distance from the L* axis (Equation (6)); meanwhile h° characterises the amplitude of the angle formed by taking the +a* axis as the x-axis (e.g., 0° is +a*, or red; 90° is +b*, or yellow) explaining the more dominant colour tones (Equation (7)). The colour evaluation was carried out taking five measurements for each sample.

$${\mathit{C}}^{\ast }={\mathbf{a}}^{\ast }\mathbf{2}+{\mathbf{b}}^{\ast }\mathbf{2}$$

(6)

$$\mathit{h} (°)={\mathit{t}\mathit{a}\mathit{n}}^{-1}\left({\displaystyle \frac{{\mathit{b}}^{\mathbf{\ast }}}{{\mathit{a}}^{\mathbf{\ast }}}}\right)$$

(7)

∆E value was also calculated through Equation (8) to determine if there was any colour difference between a sample and a reference.

$$∆\mathit{E}=\sqrt{{(∆{\mathit{L}}^{\mathbf{\ast }})}^{\mathbf{2}}+{(∆{\mathit{a}}^{\mathbf{\ast }})}^{\mathbf{2}}+{(∆{\mathit{b}}^{\mathbf{\ast }})}^{\mathbf{2}}}$$

(8)

∆E can assume different values. According to Mokrzycki et al. [26], the difference between two samples is not noticeable when ∆E is between 0 and 1; if 1 < ∆E < 2, the difference is noticeable only to experienced observers; when ∆E is between 2 and 3.5, the difference is also noticeable to inexperienced observers; if ∆E is between 3.5 and 5, there is a clear difference; and if ∆E > 5, the observer can notice two different colours.

2.6.5. Rheological Analysis

Rheological determinations were performed with a controlled stress and strain rheometer (302 MRC, Anton Paar, Gratz, Austria) thermally regulated by a Peltier plate and a circulating water bath (FP 50, Julabo, Milan, Italy). Dynamic state flow tests were done using a rough parallel plate geometry (PP25) with a 25 mm diameter probe and a gap size of 3 mm. Measurements were taken at 25 °C after a resting time of 120 s following Gruppi et al.’s [27] method. The amplitude sweep test was conducted by applying a constant frequency of 1 Hz and varying the strain amplitude from 0.01 to 100%. Based on the obtained linear viscoelastic region (LVER) results, a strain amplitude of 0.01% was set for the subsequent frequency sweep test, while varying frequency between 0.01 and 100 Hz. Measurements were taken immediately after the dough preparation (T0) and after 2 h leavening (T2), in duplicate. Obtained data were then elaborated by applying the power law model as a function of frequency (Equation (9)) according to Upadhyay et al. [28] to obtain more information about the flow properties of the doughs.

$${\mathit{G}}^{\prime }={\mathit{G}}_0^{\prime }{\mathit{\omega }}^{\mathit{n}}$$

(9)

where G′ is the storage modulus, G′₀ is the intercept of the power law model for frequency sweep, ω is frequency (Hz), and n represents the slope of G′.

2.6.6. Texture Evaluation

The breads’ textural characteristics were measured with a “holding-until-time” compression test, performed using the TVT 6700 Texture Analyser (Perten Instruments, Hägersten, Sweden). A single compression was applied to the bread without crust using a cylindric probe with a 36 mm diameter (part n°673036), applying 25% compression and a hold time of 60 s. The probe was set 5 mm away from the sample, which was previously prepared by removing the crust and cutting to obtain a bread slice of 25 mm height; measurement was done at 1 mm/s speed applying a trigger force of 5 g. This test was conducted to obtain information such as the firmness (g), denoted as the maximum force applied to compress the bread, and the springiness (%), represented by the sample recovery calculated as the force required for the sample to return to its initial shape after 60 s holding time. Texture analysis was carried out on seven measurements to ensure adequate data reliability.

2.7. In Vitro Starch Digestion

The standardised INFOGEST static in vitro digestion method [29] was applied, comprising sequential oral (2 min; 37 °C; pH 7.0), gastric (120 min; 37 °C; pH 3.0), and intestinal (120 min; 37 °C; pH 7.0) phases, as previously described and justified by Minekus et al. [30]. Rabbit gastric lipase (60 U/mL; Lypolytech, France) was included in the gastric phase. The protocol was scaled up to 50 g of sample, and aliquots (n = 3) were collected after each digestion phase and stored at −18 °C for subsequent analysis. In addition, rapidly digestible starch (RDS) and slowly digestible starch (SDS) fractions were assessed using the enzymatic method detailed by Englyst et al. [31]. RDS was measured by calculating the glucose released after 20 min, while SDS was measured as the glucose released after 100 min incubation. The resistant starch content (RS) was evaluated using a commercial assay kit (K-RSTAR 05/19, Megazyme International, Wicklow, Ireland; based on the AOAC official method 2002.02). Analyses were run in triplicate.

2.8. Extraction Method and Untargeted Metabolomics Analysis

A total of six different sourdough and dough formulations—including oat okara sourdough (OOS), oat flour sourdough (OS), wheat flour sourdough (WS), and doughs fermented for 120 h, comprising oat okara dough (OODT120), oat flour dough (ODT120) and wheat flour dough (WDT120)—were extracted in triplicate, following the procedure previously described by García-Pérez et al. [32]. Briefly, 1 g of each sample was combined with 10 mL of 80% (v/v) aqueous methanol acidified with 0.1% (v/v) formic acid. The mixture was then subjected to ultrasound-assisted extraction for 10 min at room temperature with a maximum power of 120 watts. Subsequently, samples were centrifuged at 5500 rpm for 15 min at 4 °C and stored at −18 °C overnight. The resulting supernatant was filtered through 0.22 μm cellulose syringe filters and collected in vials for metabolomics analysis. Afterwards, quality control (QC) samples were prepared by pooling equal aliquots of each sample into the same vial.

The untargeted metabolomics analysis of sourdough and dough samples was carried out through high-resolution mass spectrometry (HRMS) performed on a Q-Exactive™Focus Hybrid Quadrupole-Orbitrap Mass Spectrometer (Thermo Scientific, Waltham, MA, USA) coupled to a Vanquish ultra-high-pressure liquid chromatography (UHPLC) pump and equipped with heated electrospray ionisation (HESI)-II probe (Thermo Scientific, MA, USA). Chromatographic separation was performed using a gradient elution from 6 to 94% acetonitrile over 35 min. The mobile phase consisted of LC-MS-grade water–acetonitrile (Sigma-Aldrich, Milan, Italy), containing 0.1% formic acid as a phase modifier on a BEH C18 analytical column (2.1 × 100 mm, 1.7 μm), maintained at 35 °C. A sample volume of 6 µL was injected with a constant flow rate of 200 µL/min. Mass spectrometric detection was carried out in positive ionisation mode, with full scan acquisition over the m/z mass range 80–1200 and a mass resolution of 70,000 FWHM. Samples were injected in triplicate, following a randomised injection of pooled quality control (QC) in a data-dependent (Top N = 3) MS/MS mode. Afterwards, MS-DIAL software (version 4.80) was employed for automatic peak finding, LOWESS normalisation, and annotation via spectral matching against the comprehensive library FooDB.

2.9. Statistical Analysis

For each measurement, mean values and standard deviations were calculated and then analysed through analysis of variance (one-way ANOVA) followed by Tukey’s post hoc test performed at a p ≤ 0.05 level using the statistical software IBM SPSS^® (version Statistics 29.0, SPSS Inc., Chicago, IL, USA). Prior to one-way ANOVA, data were checked for homogeneity of variances using Levene’s test. The statistical evaluation of pH, water activity (a_w), and microbiological counts was performed by assessing significant differences at each sampling time.

The statistical multivariate analysis was performed using Mass Profiler Professional software (Agilent ^®, Santa Clara, CA, USA) for unsupervised hierarchical cluster (HCA) and principal component analysis (PCA) to compare samples based on treatments and fermentation metabolic profile (Euclidean distances and Ward’s algorithm). The raw data were normalised at the 75 h percentile, log2-transformed, and baselined against the median values.

Additionally, the effects of treatment, fermentation, and their interaction were analysed using supervised multi-block orthogonal partial least squares (AMOPLS), implemented through the “rAMOPLS” package (version 0.2) on R studio (version 4.2.3) software. The model was built based on 100 permutations. The results were expressed according to the following parameters: the relative sum of squares (RSS), representing the percentage of variability of each effect; residual structure ratio (RSR), indicating the ANOVA consistency of each effect with respect to residuals; p-value, which provides the statistical significance of each factor; and the principal predictive components, which represent the significant block contributions, in percentages, associated with each effect, and the contribution of the orthogonal predictive component. Furthermore, the variable importance in the projection (VIP) scores was employed to select the metabolites with the highest discriminant potential, selecting the top 30 VIP² markers. Finally, the application of these comprehensive statistical approaches enabled the identification of key metabolites, discriminating between fermentation and treatment conditions.

3. Results and Discussion

3.1. Microbiological Analyses

The microbiological analyses (Figure 1A), performed on the three types of sourdough and dough at different times of leavening, showed high counts of lactic acid bacteria (LAB) in both OS and OOS, reaching 9.63 and 9.32 log CFU/g, respectively, while WS exhibited a slightly lower count (8.62 log CFU/g). Yeast enumeration (Figure 1B) highlighted a comparable trend: OS and OOS exhibited similar counts (6.74 and 6.58 log CFU/g, respectively), while WS presented markedly lower values (5.10 log CFU/g). These findings align with data reported by Cera et al. [33], confirming that oat substrate composition plays a significant role in promoting the initial microbial ecology of sourdough systems. The LAB and yeast communities present in oat okara sourdough, after 30 days of back-slopping, were previously identified and described by Meanti et al. [24]. The dominant microbial species isolated included Lactiplantibacillus plantarum, Loigolactobacillus coryniformis, Enterococcus faecium and Saccharomyces cerevisiae.

The LAB dynamics during the fermentation of the three doughs (OOD, OD, and WD) reflected the trends observed in the corresponding sourdoughs: during all leavening stages (120 min), LAB counts remained relatively stable across all dough types, averaging approximately 8 log CFU/g, with minor variations (≤0.5 log CFU/g). Yeast populations in the doughs revealed a different evolution: OOD exhibited the highest yeast counts across all fermentation times, averaging 6.4 log CFU/g, followed by OD (5.57 log CFU/g) and WD (4.95 log CFU/g). These differences indicated that the incorporation of okara flour may promote yeast proliferation, likely due to its content of residual nutrients and fermentable fibres, resulting in an increase of approximately 1 or 2 log CFU/g compared to oat or wheat dough, respectively. The LAB and yeast counts observed in okara doughs were like those reported by Zielińska et al. [34] and Gül et al. [35] in wheat doughs.

Yeasts and LAB act synergistically in sourdough fermentation, influencing both leavening and the final quality of bread. LAB play a key role in flavour development through acidification, mainly lactic (“fresh”) and acetic (“sharp”) acids, and the accumulation of taste-active compounds like glutamate, glutathione, and glutamic dipeptides [25]. They also contribute to aroma via compounds such as 2-acetyl-1-pyrroline and drive a rapid pH drop that improves both sensory properties and microbial stability, extending shelf life [36]. Yeasts primarily support leavening by producing CO₂ via alcoholic fermentation [25] and contribute to mild acidification through the production of acetic and succinic acids. These acids enhance dough structure by promoting stable interactions within the protein network [37,38]. Additionally, yeasts influence the volatile profile of bread, with their metabolic activity shaped by both strain type and interactions with LAB species [39,40,41,42]. Moreover, Sayadi et al. [43] reported that sourdoughs obtained from Saccharomyces cerevisiae were softer and that increasing the water-holding capacity made the dough more extensible and elastic.

3.2. Physical-Chemical Analyses Chemical–Physical Analyses

Physical-chemical analyses (Figure 1C,D) were performed on the three types of sourdough, dough, and bread. OOS, OS, and WS reached low pH values (3.76, 3.85, and 3.70 respectively), and after two hours of leavening, all three types of dough maintained a stable pH of around 4.6. These findings are consistent with previous observations reported by García-Béjar et al. [44], who also noted similar pH profiles in oat- and wheat-based sourdough systems. The same pH value was found for breads of okara and oat, while that of wheat bread increased slightly to 4.77. According to Casado et al. [45], heterolactic fermentation of LAB typically decreases pH to around 5–4, which explains the relatively low pH observed in sourdough and dough samples.

Water activity (a_w) values were comparable across all sourdough samples (around 0.9940 for okara and oat and 0.9881 for wheat). The water activity of the raw doughs remained stable throughout the leavening period, with values around 0.98. In contrast, a marked decrease in water activity was observed in the bread samples, primarily due to the baking process and the associated moisture loss. The WD bread showed the most pronounced reduction in a_w (0.9118), followed by the oat OOD bread (0.9308) and the OD bread (0.9605). This higher a_w observed in oat and okara bread compared to wheat is likely attributable to the presence of β-glucans in oats, and consequently in oat-derived okara, which possess strong water-binding capacity. The differences in a_w values between oat and okara bread may result from the thermal processing applied during plant-based beverage production, which can degrade β-glucans either through heat-induced breakdown or enzymatic depolymerisation, thereby reducing their structural integrity and ability to retain water [46,47].

3.3. Technological Evaluation

3.3.1. Moisture Content

The moisture content of the fresh doughs was measured immediately after their preparation, after 2 h of leavening, and after the baking of the breads. The obtained results (

Table 2. Moisture content (%) and dry matter content (%) of the fresh doughs immediately after their preparation, after 2 h leavening, and after bread baking. Values are means ± standard deviations (n = 3). Within each row, different superscript letters indicate statistically different values according to post hoc comparison (Tukey’s test) at p ≤ 0.05.

	Times	OOD	OD	WD
Moisture (%)	Fresh dough	43.72 ± 0.13 c	40.39 ± 0.15 b	36.24 ± 0.22 a
	2 h leavened	43.60 ± 0.41 c	40.26 ± 0.17 b	37.12 ± 0.46 a
	Bread	41.55 ± 0.42 c	31.59 ± 1.95 b	24.01 ± 0.30 a
Dry Matter (%)	Fresh dough	56.28 ± 0.13 a	59.61 ± 0.15 b	63.76 ± 0.22 c
	2 h leavened	56.40 ± 0.41 a	59.74 ± 0.17 b	62.88 ± 0.46 c
	Bread	58.45 ± 0.42 a	68.41 ± 1.95 b	75.99 ± 0.30 c

) indicated that all samples exhibited different moisture contents; OOD exhibited the highest, while WD was the driest dough. Water constitutes about 40% of bread dough depending on the protein content of the flour used to produce it and the mechanical stress to which the dough is subjected [48]. The differences in moisture content are mainly related to the different sourdoughs used to make the breads. Moisture measured after 2 h of leavening showed no substantial differences compared to the corresponding fresh dough. As expected, the same behaviour was confirmed through the moisture content measurement of the bread, with a noticeably low moisture content for wheat bread (24.01%) compared to oat and okara breads (31.59 and 41.55%, respectively). Notably, measurements were conducted on both the crust and the crumbs of the breads, and the results aligned with Czuchajowska et al. [48] and Jeakel et al. [49] for wheat dough. The highest moisture content was found in OOD and its bread, followed by OD, then WD, and their relative breads: this can mainly be attributed to the differences in the sourdoughs used for dough production, as previously specified.

3.3.2. Weight Loss and Specific Volume

The percentages of bread weight loss after baking are reported in

Table 3. Weight loss (%), specific volume (cm³/g) and textural parameters (firmness, g; force B, g; springiness, %) of the different breads. Values are means ± standard deviations (n = 7). Within each row, different superscript letters indicate statistically different values according to post hoc comparison (Tukey’s test) at p ≤ 0.05.

		OOD	OD	WD
Weight loss (%)		19.29 ± 0.98 b	18.24 ± 0.96 ab	17.59 ± 0.74 a
Specific volume (cm3/g)		1.79 ± 0.13 a	2.22 ± 0.16 b	1.45 ± 0.10 c
Texture	Firmness (g)	4304.00 ± 893.53 a	4879.38 ± 479.32 a	9160.63 ± 728.53 b
	Force B (g)	2081.86 ± 391.07 a	2292.50 ± 173.41 a	4447.25 ± 470.92 b
	Springiness (%)	0.49 ± 0.01 a	0.48 ± 0.02 a	0.49 ± 0.02 a

. WD exhibited the lowest loss in weight, followed by OD; OOD bread showed the highest weight reduction, at almost 20%. These results aligned with those reported by Puhr and D’Appolonia et al. [50], whose research focused on the effect of baking on wheat bread yields. The specific volume of the breads is reported in Table 3. Significant differences were observed among samples (p ≤ 0.05). Breads containing oat-based ingredients exhibited higher specific volume values, with OD showing the highest value, followed by OOD, whereas WD presented the lowest specific volume. Specific volume is a common indicator of bread quality and is related to water retention and gas-holding capacity during fermentation [51]. The values obtained are consistent with those reported by Farkas et al. [52] for oat-based breads and by Verdonck et al. [38] for Type I sourdough products.

3.3.3. Texture Evaluation

Each bread was subjected to a single compression test to understand its firmness, the sample resistance after 60 s holding time (force B), and its springiness; the results are reported in Table 3. Regarding firmness, the bread obtained from WD was the hardest bread for both firmness and force B: the values were approximately double those obtained from OOD and OD, indicating that the crumb was very firm, possibly due to low humidity and reduced yeast development. The breads obtained from OOD and OD were comparable in technological terms. Regarding springiness, all samples were statistically similar, showing 50% recovery of their initial shape after compression holding. These results were comparable with those obtained by Pichler et al. [53], who performed the same test during the characterisation of a gluten-free bread with micronized oat husk fibre. Notably, the present study aimed to verify if the oat okara sourdough was suitable for bread making, and the textural analysis was needed only for bread characterisation. To improve springiness and obtain a soft and elastic bread, an increase in water content should be considered, for proper gluten network formulation. Springiness may have remained around 50% due to insufficient or incomplete starch swelling and gelatinisation during baking [54].

OOD and OD were softer than WD, probably due to the higher presence of yeast, resulting in a higher production of CO₂ and thus a more aerated structure. Sayadi et al. [43] stated that wheat bread with a lower yeast content may exhibit a denser and harder texture. Yeasts may release enzymes that free small peptides from proteins, resulting in weaker gluten development and reduced springiness; given that oat protein content is around 18% [55], while that of wheat varies from 9 to 12% [56], this could explain the higher softness of oat and oat okara breads. Additionally, the tenderness observed in both OOD and OD breads may be linked to the presence of β-glucans, which have been reported to enhance crumb and crust tenderness in bakery products [57,58].

3.3.4. Rheological Analysis

Oscillatory tests were conducted to determine the effect of sourdough fermentation on dough properties. All samples exhibited a suggested LVER of 0.01 (

Table 4. Rheological parameters measured on the dough through the amplitude sweep test (real and suggested LVER) and application of the power law to the data obtained through the frequency sweep test (n, log G′, G′, and R²). Values are means ± standard deviations (n = 2). Within each column, different superscript letters indicate statistically different values according to post hoc comparison (Tukey’s test) at p ≤ 0.05.

		Real LVER	Suggested LVER	n	Log G′	G′	R2
OOD	T0	0.041 ± 0.002 a	0.010 ± 0.000 a	0.212 ± 0.002 a	4.069 ± 0.049 b	11,747.7 ± 1309.4 a	0.96 ± 0.01 a
	T2	0.039 ± 0.004 a	0.010 ± 0.000 a	0.242 ± 0.002 a	3.902 ± 0.044 a	8008.4 ± 803.2 a	0.98 ± 0.01 a
OD	T0	0.041 ± 0.002 a	0.010 ± 0.000 a	0.277 ± 0.020 a	3.886 ± 0.039 c	7703.0 ± 683.9 b	0.97 ± 0.01 a
	T2	0.038 ± 0.001 a	0.010 ± 0.000 a	0.287 ± 0.005 a	3.670 ± 0.023 a	4677.9 ± 247.4 b	0.97 ± 0.01 a
WD	T0	0.042 ± 0.000 a	0.010 ± 0.000 a	0.218 ± 0.001 b	4.264 ± 0.035 a	18,401.6 ± 1481.5 a	0.96 ± 0.01 a
	T2	0.042 ± 0.001 a	0.010 ± 0.000 a	0.202 ± 0.047 a	4.390 ± 0.398 a	12,849.9 ± 0.0 a	0.99 ± 0.00 a

). All samples were characterised by a greater G′ than G″, indicating an elastic solid-like behaviour of the doughs. G′ and G″ were higher in both samples containing oats (OD and OOD) than in WD. The increase in both storage and loss modulus for OD and OOD can mainly be attributed to the greater presence of fibre and proteins that can interact among them [59]. The present results contrast with those reported by Hüttner et al. [60]; in fact, the sourdoughs showing the highest G′ were more elastic (Table 4). Linear regression data showed that all samples exhibited a slope below 0.4, as expected based on Georgopoulos et al. [61]. Particularly, n values approaching 0 indicate a good 3D network and more rubbery material; in our study all sourdoughs showed similar n values of about 0.20. The n increase also seems to respond to water content: an increase in water can increase n and thus the fraction of uncrossed linked material and weakness of the gel structure [61,62]. The results showed that all doughs can exhibit good bread-making ability. The leavening time did not affect the viscoelastic stability of the samples; no great changes were observed between T0 and T2 for the same sample.

3.3.5. Colour Evaluation

The colour of the doughs was measured immediately after their preparation, and the colour of the breads was assessed after baking. This evaluation is fundamental to assess bread production, product acceptability, process optimisation, and bread quality [63]. Results are reported in

Table 5. Colour evaluation (L*, a*, b*, C*, and h°) of the different doughs and the baked breads; ΔE values are expressed with OOD as a reference. Values are means ± standard deviations (n = 5). Within each row, different superscript letters indicate statistically different values according to post hoc comparison (Tukey’s test) at p ≤ 0.05.

		OOD	OD	WD
Colour of doughs	L*	73.71 ± 0.36 a	78.14 ± 0.21 b	81.69 ± 0.32 c
	a*	4.17 ± 0.09 c	2.43 ± 0.13 b	1.52 ± 0.05 a
	b*	21.65 ± 0.36 c	20.51 ± 0.17 b	19.38 ± 0.51 a
	C*	22.05 ± 0.37 c	20.66 ± 0.18 b	19.44 ± 0.51 a
	h°	79.09 ± 0.09 a	83.23 ± 0.30 b	85.52 ± 0.16 c
Colour of baked breads	L*	61.55 ± 0.45 a	64.41 ± 2.31 a	78.81 ± 4.52 b
	a*	6.38 ± 0.10 a	9.41 ± 1.29 b	9.54 ± 0.96 b
	b*	25.36 ± 0.21 b	23.76 ± 1.44 a	29.78 ± 0.66 c
	C*	26.15 ± 0.22 a	25.57 ± 1.71 a	31.29 ± 0.74 b
	h°	75.88 ± 0.18 c	68.46 ± 2.01 a	72.25 ± 1.66 b

. Samples really differed among the doughs, with all L* coordinates being statistically different. The lower L* value found in okara bread obtained from OOD with the highest yeast count aligns with Sayadi et al. [43]’s findings. Regarding the a* value, WD was the least red sample, followed by OD, and OOD was the reddest; b* values showed that OOD was the yellowest and the darkest overall. This result was confirmed by Chroma and hue angle, showing the greatest saturation and the lowest angle (closer to red, 0°, than to yellow, 90°). WD was the lightest dough, with the lowest a* and b* values: this dough was the brightest and the most yellow. OD dough was quite clear, with intermediate values of a* and b*.

Bread surface is subjected to the formation of colour during baking. This browning is mainly due to the Maillard reactions that occur between reducing sugars and amino compounds [64]. Baked breads are darker than dough due to baking, and the differences among breads can be attributed to the different sourdoughs used. Okara bread was the darkest but not the reddest and yellowest; wheat bread was the brightest due to the absence of oats [43], showing the highest b* value and one of the highest a* values (similar to OD). Oat bread was quite dark with an intermediate L* value (64.41), a strong red tone, and a low b* value. The colour differences between OOD, WD, and OD can be attributed to the synergistic effects of yeasts, sourdough fermentation, and the different flours used [43]. According to [63], consumers prefer breads with intermediate L*, with soft shades of brown (L* closer to 62.9) and light brown colourings (given by a combination of a* and b*). The obtained breads in this study presented good shades of brown, which were higher in OOD and OD samples due to oat presence. Strong colour differences were observed between OOD and OD, with ΔE values ranging from 3.5 to 5. In comparison, WD exhibited even larger deviations, with ΔE values exceeding 5 [65]. Overall, these variations in colour appeared to be mainly linked to the type of flour used.

3.4. In Vitro Bio-Accessibility of Bread After Simulated Starch Digestion

The total starch, RDS, and SDS values for each type of bread are shown in

Table 6. SDS/Av—slowly digestible starch on available starch; RDS/Av—rapidly digestible starch on available starch.

	Total Starch	SDS/Av	RDS/Av
Okara Bread	62.3 ± 1.14	11.21 ± 1.19	88.79 ± 1.19
Oat Bread	50.58 ± 2.40	8.72 ± 0.42	91.28 ± 0.42
Wheat Bread	53.31 ± 0.50	5.2 ± 0.8	94.8 ± 0.8

. After the samples were compared, differences in RDS and SDS values were observed. The OOD bread showed the lowest RDS value (88.79 ± 1.19 g/100 g) and highest SDS value (11.21 ± 1.19 g/100 g) of the sourdough breads, followed by the bread produced with oat sourdough (OD) (RDS: 91.28 ± 0.42 g/100 g; SDS: 8.72 ± 0.42 g/100 g) and wheat sourdough (WD) (RDS: 94.8 ± 0.8 g/100 g; SDS: 5.2 ± 0.8 g/100 g). The fermentation process increased the portion of RDS, making the breads more easily digestible. From a nutritional perspective, recent research suggests that an increased proportion of SDS in foods may be associated with a slower glucose release and a more favourable post-prandial glycaemic profile. In 2011, the European Food Safety Authority (EFSA) declared that the in vitro digestibility of cereal products could be used as an indicator of the in vivo effect on post-prandial glycaemic responses [66]. Recent studies have confirmed a positive relationship between dietary consumption of SDS and a reduction in post-prandial glycaemic response; foods with SDS also have a protective effect against chronic diseases [67,68], as they allow for slow absorption of glucose during digestion. Insulin hormones reduce gastric emptying, also associated with glycaemic control, weight loss, and lower blood insulin levels, which could lead to health benefits such as reduced risk of diabetes or metabolic syndrome [69]. The production of bread and other cereal-based foods using sourdough is gaining increasing attention, as it appears to enhance the SDS content in the final product. This effect is mainly attributed to the formation of organic acids (lactic, acetic, and propionic acids), which reduce the rate of starch digestion through chemical modifications occurring during sourdough fermentation and limit starch gelatinisation [70,71]. Consequently, fermented cereal-based products typically exhibit lower starch digestibility values [72]. However, in vitro starch digestibility represents an indirect measure, and confirmation of potential effects requires in vivo studies.

3.5. Multivariate Statistics on the Chemical Profile of Sourdough and Dough Samples

An untargeted metabolomics analysis was performed as a preliminary and exploratory approach to investigate the impact of different treatments (oat okara, oat flour, and wheat flour) and fermentation processes (sourdough and dough fermented for 120 h) on the chemical profile of the samples. Firstly, an unsupervised hierarchical cluster analysis (HCA) was performed based on the average fold-change variations across all samples. The resulting dendrogram revealed a clear separation between OOS and OS compared to the other samples. In contrast, WS and WDT120 were grouped together within a sub-cluster, whereas the oat dough samples fermented for 120 h were clustered independently from each other and from the sourdough samples (Supplementary Figure S1A). These findings suggest that the observed differences are primarily associated with the fermentation process as opposed to the type of treatment. These results were further supported by principal component analysis (PCA), where first and second principal components explained 31.47% and 13.18% of the variability, respectively (Supplementary Figure S1B). In parallel, a supervised ANOVA multiblock orthogonal projection to latent structures discriminant analysis (AMOPLS-DA, Supplementary Figure S1C) was applied to determine the specific influence of each factor contributing to the spontaneous fermentation in sourdough and dough samples and identify VIP² markers (

Table 7. Relative variability and block contributions of the AMOPLS analysis of sourdough and dough, considering the three factors under investigation.

Effect Name	RSS	p-Value	Block Contributions
			Tp1	Tp2	Tp3	Tp4	Tp5	To
Fermentation	16.8%	0.01	0	1	0.001	0.000	0.000	0.197
Treatment	36.3%	0.01	1	0	0.000	0.998	0.000	0.126
Fermentation × Treatment	14.8%	0.01	0	0	0.998	0.001	0.999	0.269
Residuals	32.1%	NA	0	0	0.001	0.001	0.001	0.408

RSS, relative sum of squares; To, orthogonal component; Tp, predictive component. The highest contribution for each component is highlighted in bold.

). The results showed that all factors involved in the study exhibited a statistically significant effect (p = 0.01) with the treatment exhibiting the highest influence on the observed differences (RSS = 36.3%). Moreover, this effect was further explained by two predictive components, tp1 and tp4, which accounted for the highest contribution (100% and 99.8%, respectively), whereas the orthogonal component showed the lowest contribution of 12.6%. In contrast, fermentation factor and the interaction between the two factors (fermentation × treatment) accounted for 14.8% and 16.8%, respectively. Finally, the residual variability accounted for the remaining 32.1%, suggesting that this portion of the data variability cannot be explained by the studied factors.

Additionally, a Variable Important in the Projection (VIP²) approach was applied to identify the markers for the discrimination among treatments (VIP score > 2). A comprehensive list of the 30 discriminant metabolites classified by superclass, with lipids as the predominant group, followed by benzenoids and organic acids and derivatives, is provided in Supplementary Table S1. The high discriminant potential of lipids can be attributed to the fact that oats tend to accumulate more lipids than wheat [73]. Delving deeper into this metabolite superclass, AMOPLS analysis revealed that steroidal compounds and fatty acyls were the most impacted classes, followed by prenol lipids. Among steroid compounds, most are represented by steroidal glycosides, which belong to the class of saponins. Oats are considered a good source of these metabolites. Regarding the treatment, 25-Epiruizgenin 3-[4″-rhamnosylglucoside] was most discriminant, with a VIP² score of 2.715. Notably, fermentation has been reported to enhance saponin content in oats, thereby improving the bioavailability of these health-promoting compounds [74,75]. However, the mechanism by which fermentation influences the production of steroidal saponins remains unclear [76]. Thus, further studies are needed to investigate the modifications occurring in fermented oats. Regarding fatty acyls, the unsaturated fatty acid 2-Octenedioic acid exhibited the highest impact with a VIP² score of 2.688, followed by (11E)-9,10,13-trihydroxyoctadecenoic acid (VIP² score = 2.673) and (Z)-15-Oxo-11-eicosenoic acid (VIP² score = 2.644). These findings are consistent with previous studies reporting oats as a rich source of lipids, characterised by a higher concentration of fatty acids than other cereals [73,77]. Overall, these trends highlight the potential of oat okara and oat flour to modulate the chemical profile of sourdough and dough. While some detected metabolites are commonly associated with bioactive properties, the lack of functional validation limits any conclusion on their health effects, which should therefore be considered preliminary. These findings nevertheless support microbiota-driven biochemical transformations during fermentation, although further integrated and targeted studies are needed to elucidate the underlying metabolic pathway.

4. Conclusions

This study demonstrated the potential of oat okara sourdough as a functional and sustainable ingredient in bread making. Compared to sourdoughs prepared with oat and wheat flours, oat okara sourdough exhibited high lactobacilli content and yeast count in all samples, indicating an excellent fermentation substrate. Similar results were obtained for oat sourdough, suggesting that the fermentable fibres present in oat-based flours, such as β-glucans, may serve as an effective carbon source supporting microbial growth in oat okara sourdough. Technological evaluations of texture and colour, supported by metabolomic profiling, showed that breads made with oat okara and oat flour performed better than wheat flour controls in terms of dough softness and the presence of positive metabolites. Metabolomic profiling further revealed differences among treatments. The VIP² analysis identified lipids as the main discriminant metabolites, followed by benzenoids and organic acids and their derivatives. Within the lipid fraction, steroidal glycosides (saponins) and fatty acyls were strongly influenced by fermentation, suggesting enhanced formation of bioactive compounds and modifications in the fatty acid profile. These changes indicate that the inclusion of oat okara in sourdough formulations can modulate the metabolic composition of the dough, improving its functional and nutritional potential. In vitro digestion analysis further revealed that bread made with oat okara sourdough exhibited a higher SDS content, which could be associated with an improved post-prandial glycaemic response and potential health benefits, although further research is needed in this area. Selecting specific microorganisms and optimising the fermentation time of the dough could possibly further enhance the quality of bread produced with okara sourdough. Overall, these findings support the use of oat okara in sourdough formulations as a promising strategy to improve the nutritional and functional quality of bread. This approach not only enhances the health potential of bakery products but also contributes to sustainable food production by upcycling oat by-products into high-value products, aligning with circular economy principles.