Researchers at the University for Development Studies in Ghana’s Northern Region collected fresh wet okara from two artisanal tofu and soymilk processors near their campus, combining the batches into a homogeneous supply within minutes of collection. They then designed a rigorous 2 × 5 factorial experiment: okara was either heated at 90 to 100 degrees Celsius for just five minutes, mimicking traditional hot processing, or left unheated, and both groups were spontaneously fermented at 25 degrees Celsius for 0, 12, 24, 36, or 48 hours. No starter culture was added, meaning the fermentation relied entirely on whatever microorganisms happened to be present. After fermentation, samples were sun-dried for four to five hours at roughly 35 to 38 degrees Celsius, consistent with traditional flour processing in northern Ghana, and milled into flour for analysis. Three independent processing replicates produced 30 distinct experimental samples.

The most striking finding concerned flavor. A panel of 25 semi-trained food science students, evaluating samples in an ISO-standard sensory laboratory over three days, rated beany-flavor intensity on 9-point scales. Fermentation duration turned out to be the overwhelming driver of this attribute, explaining nearly 98 percent of the variance. Scores stayed low at 0 and 12 hours, hovering around 2.2 to 2.7, but jumped dramatically to between 6.6 and 7.5 once fermentation extended past the 12-hour mark. Crucially, the five-minute heat pretreatment itself had no significant effect, and there was no meaningful interaction between heating and fermentation time. In other words, the clock, not the heat, controlled the flavor trajectory, and the sweet spot was unmistakably early fermentation.

The authors are careful about what those rising scores actually mean. Because the panel was trained only on beany-flavor references, the elevated ratings after 24 hours may partly reflect the accumulation of new fermentation-derived odorants, such as the cheesy, natto-like, or fermented notes that other studies have linked to microbial metabolites like acetoin, pyrazines, and branched-chain fatty acids, rather than a true intensification of soybean beany character. Previous work has shown that fermentation can reduce aldehydes like hexanal, a major beany compound, while simultaneously generating entirely new aroma-active molecules. Without volatile profiling, the study cannot separate these contributions, but the practical message stands: extending spontaneous fermentation beyond 12 hours did not progressively improve the sensory profile.

Beneath the sensory story, the fermentation chemistry told a more complicated tale. Both pH and titratable acidity responded strongly to fermentation duration, but with a surprising twist: a significant interaction with thermal pretreatment. In unheated okara, pH fell from 5.35 to 4.48 within 12 hours, then rebounded sharply to 6.52 at 24 hours before dropping again to 4.24 by 48 hours. Heated samples followed the same phasic pattern, dipping to 4.78 at 12 hours, spiking to 6.74 at 24 hours, and ending at 4.59. This 24-hour rebound suggests that acidification was not sustained, possibly because the dominant microbial populations shifted toward proteolytic activity, generating alkaline nitrogenous compounds such as ammonia, a phenomenon documented in other spontaneous legume fermentations. The researchers emphasize, however, that they did not characterize the microbial community, so these mechanisms remain plausible explanations rather than proven causes.

Nutritionally, the flour held its own throughout. Crude protein ranged from about 24.5 to nearly 30 grams per 100 grams, peaking in the heated 12-hour sample, while carbohydrate, fat, ash, and calculated energy all shifted with fermentation time in treatment-dependent ways. The authors caution that these increases in measured protein and fat proportions do not indicate net synthesis of new nutrients; they more likely reflect proportional redistribution as microbes consumed organic matter and altered extractability. Similarly, ash fluctuations represent shifts in the inorganic fraction relative to declining organic components, not mineral generation. The study’s honesty on these interpretive limits is refreshing in a field where compositional gains are often oversold.

Phytochemical measurements added another layer of nuance. Total phenolic content rose from around 5.8 to over 12 micrograms of gallic acid equivalents per gram during the first 24 hours, consistent with microbial enzymes liberating phenolics from bound forms, but then declined by 48 hours, with the trajectory diverging between heated and unheated samples. Phytate, the notorious antinutrient that binds minerals, followed a non-linear path, peaking around 24 hours before dropping to its lowest levels, roughly 272 milligrams per 100 grams, at 48 hours in heated samples. Tannins showed similarly complex, pretreatment-dependent swings. None of these changes translate directly into claims about mineral bioavailability, which was not measured, but they do suggest that fermentation time reshapes the phytochemical landscape in ways that simple longer-is-better thinking misses entirely.

Functional properties, which determine how a flour behaves in real food formulations, also responded in treatment-specific ways. Water absorption capacity ranged from 3.75 to 5.02 grams of water per gram of flour, with the highest value in the heated 24-hour sample, while oil absorption and swelling power varied more modestly. Bulk and tapped densities shifted with fermentation, though the Hausner ratio and Carr’s index, the standard powder-flow metrics, remained statistically unchanged. The authors note that okara’s swelling behavior cannot be attributed to starch gelatinization alone, since the material is dominated by insoluble fiber and protein, and that heat and fermentation likely modified the accessibility and organization of water-binding components in the matrix.

Perhaps the most elegant analytical touch was the principal component analysis, which mapped all ten treatment combinations across eight quality variables. The first three components captured over 81 percent of the variation, and the patterns revealed that prolonged fermentation pushed quality attributes in different directions simultaneously rather than producing a uniform improvement. Beany flavor loaded almost entirely onto a third dimension separate from the dominant physicochemical variation, underscoring that flavor and composition respond to fermentation on partly independent tracks. The heated 12-hour sample occupied a favorable region combining low beany intensity with comparatively high measured protein.

The bottom line is a practical recommendation wrapped in appropriate caveats: heating okara at 90 to 100 degrees Celsius for five minutes followed by roughly 12 hours of spontaneous fermentation offered the best balance of low beany flavor, solid protein content, and acceptable functional properties among the tested conditions. The researchers are explicit that this is not proof of universal optimality, consumer acceptance, or microbiological safety. The pH rebound at 24 hours, combined with an uncharacterized natural microbial community and no starter culture, means safety verification is essential before any food application. Future work, they argue, should profile the microorganisms, metabolites, and volatile compounds driving these changes, and test the selected treatment in real formulated foods with actual consumers. For now, the study transforms a waste-stream liability into a data-backed ingredient candidate, and it does so with a methodological rigor that other valorization studies would do well to imitate.

Subject of Research: Effects of thermal pretreatment and spontaneous fermentation duration on the sensory, physicochemical, phytochemical, and functional quality of okara flour

Article Title: Effects of Thermal Pretreatment and Spontaneous Fermentation on the Quality Characteristics of Okara Flour

Article References: Bawa, A.-A., Akabanda, F., & Afoakwah, N. A. (2026). Effects of Thermal Pretreatment and Spontaneous Fermentation on the Quality Characteristics of Okara Flour. Food Science & Nutrition, 14(10), Article e72453. https://doi.org/10.1002/fsn3.72453

Image Credits: AI Generated

DOI: 10.1002/fsn3.72453

Keywords: okara, soybean by-product, spontaneous fermentation, thermal pretreatment, beany flavor, food valorization, phytochemicals, phytate, water absorption capacity, flour quality, food science, Ghana

The Norwegian brewery, which also produces cider from its own orchard, prepared 1100 liters of wort from malted barley, 30 percent unmalted wheat, and aged dry hops, acidified to pH 4.9 with lactic acid in imitation of Belgian practice. The wort cooled overnight in a metal coolship open to the ambient air, and two hours before the end of cooling the brewers added 40 kilograms of freshly harvested whole Aroma apples. The cooled wort was then split into three similar 225-liter oak barrels that had previously held American Merlot wine and a plum beer, cleaned with warm water at 100 bar pressure before filling. These barrels, sampled repeatedly from two hours to 27 months after filling, served as biological triplicates for the study.

The analytical arsenal deployed was formidable. Selective agar media allowed researchers to count presumptive lactic acid bacteria, acetic acid bacteria, yeasts, and enterobacteria. Amplicon sequencing of the bacterial 16S rRNA gene and the fungal ITS1 region revealed which genera were present, while shotgun metagenomic sequencing of whole-community DNA, yielding more than 134 million high-quality reads and 41.7 billion base pairs, enabled species-level identification through fragment recruitment plotting. Cell-free supernatants were analyzed for carbohydrates, sugar alcohols, organic acids, amino acids, biogenic amines, and dozens of volatile organic compounds, giving a parallel picture of the chemistry unfolding inside each barrel.

The results confirmed that the same microbial succession known from Belgian lambic breweries took place in Norway, but with blurred boundaries. In classic lambic production, four phases follow one another: an initial fermentation dominated by enterobacteria and non-Saccharomyces yeasts, a main fermentation driven by Saccharomyces, an acidification phase led by Pediococcus and acetic acid bacteria, and a long maturation dominated by Brettanomyces. In the Norwegian barrels, these phases occurred but overlapped considerably. In barrel 2, for example, Brettanomyces species appeared as early as day 27, coinciding with Saccharomyces and enterobacteria, so that main fermentation, acidification, and maturation ran simultaneously.

The early days of fermentation were dominated by enterobacteria, reaching plate counts as high as log 9.6 colony-forming units per milliliter despite the manual acidification of the wort, which was clearly insufficient to suppress them. Shotgun metagenomics identified Rahnella variigena, an unidentified Pseudescherichia species close to Pseudescherichia vulneris, an unidentified Leclercia species, and a Klebsiella species related to Klebsiella huaxiensis. Remarkably, Rahnella variigena has never before been identified in beer. Its varying abundance across the three barrels correlated with concentrations of dimethyl sulfide, acetoin, and 2,3-butanediol, compounds that enterobacteria produce from monosaccharides and that can impart off-flavors. The cadaverine detected in all barrels was plausibly linked to the Klebsiella species, which carries a lysine decarboxylase gene.

Acetic acid bacteria made an early but barrel-dependent appearance. Acetobacter cerevisiae dominated barrel 2 in the first hours, reaching 59.7 percent relative abundance, while barrel 3 hosted Gluconobacter cerinus at 44.6 percent. Barrel 1, filled first from the bottom of the coolship, contained no acetic acid bacteria at all, likely because these aerobic microbes congregated at the liquid-air interface during cooling. Intriguingly, both A. cerevisiae and G. cerinus are known from apple surfaces and cider fermentations, hinting that the added apples may have contributed them, although the absence of an apple-free control fermentation means the researchers cannot be certain. The apple addition did visibly boost fruit-surface microbes such as Metschnikowia, Aureobasidium, and Sporobolomyces at the coolship stage, but only the bacterial family Erwiniaceae, including Pantoea agglomerans and Erwinia billingiae, persisted beyond it.

The main fermentation revealed a surprise: two different Saccharomyces species appeared in different barrels. Saccharomyces bayanus emerged early in barrels 1 and 3, possibly because of its greater tolerance of the cellar’s cool temperatures, while S. cerevisiae peaked at month 6 in barrel 2. Since the brewery also ferments cider spontaneously, and S. bayanus is a common cider yeast, the wood of the barrels may have harbored it. The Saccharomyces yeasts consumed fructose, glucose, and sucrose within three days, then maltose by month 2 or 12 depending on the barrel, producing ethanol concentrations near 50 grams per liter along with higher alcohols and esters that shape the beer’s aroma.

The acidification and maturation phases then converged on the familiar lambic key players. Pediococcus damnosus, the great acidifier of lambic beer, appeared from month 6 or 12 onward and reached nearly half of the microbial community in barrel 3, driving the pH down from 4.9 to about 3.6 and pushing total lactic acid to roughly 5.7 to 6.0 grams per liter in an almost equimolar mix of the L and D forms. Brettanomyces species, chiefly B. custersianus, B. bruxellensis, and B. anomalus, degraded the leftover maltooligosaccharides and generated the signature volatile phenols 4-ethylphenol and 4-ethylguaiacol, which appeared within the first month at constant levels of about 1 to 2 milligrams per liter. Ethyl acetate and ethyl lactate rose in two waves, one coinciding with Saccharomyces and the later one with Brettanomyces, while acetoin produced by P. damnosus was converted into 2,3-butanediol by the Brettanomyces yeasts as part of their redox balancing.

Perhaps the most sobering finding is that the three barrels, filled with identical wort and stored side by side in the same cellar, developed measurably different microbial communities and metabolite profiles, underscoring that each wooden barrel acts as its own reservoir of resident microbes and its own micro-environment. The researchers caution that without an apple-free control, their conclusions apply only to this specific apple batch from this specific orchard, and that batch-to-batch variability at the brewery remains unexplored. Still, the message is clear: the essence of lambic, a self-organizing community of enterobacteria, Saccharomyces, Pediococcus, and Brettanomyces orchestrated by wooden barrels rather than by any magical valley air, can be recreated far from Brussels. The Senne valley may hold the tradition, but the microbes, it turns out, are willing travelers.

Subject of Research: Microbial ecology and metabolite dynamics of spontaneously fermented lambic-style acidic beer produced in Norway

Article Title: Norwegian lambic-style acidic beers share key microorganisms and metabolites with Belgian lambic beers, although typical microbial phases are less pronounced

Article References: Vermote, L., Jenkins, A., Moen, B., Axelsson, L., De Vuyst, L., & Weckx, S. (2026). Norwegian lambic-style acidic beers share key microorganisms and metabolites with Belgian lambic beers, although typical microbial phases are less pronounced. Heliyon, 12(15), Article e45521. https://doi.org/10.1016/j.heliyon.2026.e45521

Image Credits: AI Generated

DOI: 10.1016/j.heliyon.2026.e45521

Keywords: lambic beer, spontaneous fermentation, sour beer, Brettanomyces, Pediococcus damnosus, Saccharomyces, enterobacteria, acetic acid bacteria, shotgun metagenomics, wooden barrels, apple addition, food microbiology