Comprehensive analysis of changes in the quality, functional properties, flavor, and metabolome of blueberry pomace by different drying methods

Abstract

Blueberry pomace is rich in bioactive compounds but limited by high moisture content. This study evaluated eight drying methods, including natural sunlight drying (ND-S), no-sunlight natural drying (ND-NS), hot-air drying at 45/160 °C (HAD45/HAD160), microwave drying (MD), vacuum freeze-drying (VFD), and combined steam explosion drying (SE-HAD and SE-VFD), for their effects on physicochemical properties, bioactive compounds, flavor, and inhibitory activity against α-glucosidase of blueberry pomace powder. MD and HAD160 caused severe browning and significant bioactive losses, while VFD best preserved anthocyanins. SE-VFD showed the highest phenolic and flavonoid contents and achieved 96.37% α-glucosidase inhibition. Metabolomics revealed that steam explosion contributes to the enrichment of metabolites associated with flavonoid biosynthesis pathways, and molecular docking demonstrated that procyanidins B1 and B3 strongly interacted with α-glucosidase through hydrogen bonds and van der Waals forces. Entropy analysis ranked SE-VFD as the highest in overall functional quality, followed by HAD45/VFD and SE-HAD. SE-VFD is promising for functional product development, whereas HAD45 may be a promising candidate for industrial production.

1. Introduction

Blueberries (Vaccinium corymbosum) are among the most widely consumed berries globally, prized for their distinctive flavor and dense nutritional profile. Beyond fresh consumption, blueberries are extensively processed into juice, jam, or wine (Wang et al., 2023), generating substantial quantities of byproducts pomace that is typically discarded or relegated to low-value animal feed. This not only squanders a rich source of bioactive compounds but also places a strain on the environment (Li et al., 2026).

In recent years, the comprehensive utilization of blueberry pomace has attracted increasing attention and shown tremendous application potential (Miriam et al., 2026). Blueberry pomace is rich in dietary fiber, anthocyanins, flavonoids, polyphenols, proteins, and confers multiple health benefits including antioxidant, antidiabetic, antiobesogenic, and hyperlipidemia, improvement of cognitive impairment, and antibacterial properties (Huang, Zhang, et al., 2025). Due to its high moisture content and unstable physicochemical properties, fresh blueberry pomace is difficult to use directly in food processing and requires efficient and sustainable drying methods for processing (José et al., 2024). Numerous studies have shown that different drying methods can preserve the nutritional and functional properties of fruit by-products. These methods primarily include hot-air drying (HAD), natural drying (ND), vacuum freeze-drying (VFD), microwave drying (MD), and combined drying technologies (such as ultrasound, ultra-high pressure, and infrared radiation) (José et al., 2024; Xu et al., 2022; Ying et al., 2026; Zang et al., 2024; Zeng et al., 2025; Zhang et al., 2025). The ND is simple and low-cost, but it is highly susceptible to weather conditions, often resulting in inconsistent quality (Li et al., 2026). Studies have found that ND can increase the lycopene and vitamin C content in tomato powder (Aderaye et al., 2026). The HAD can improve the emulsifying activity, swelling capacity, and crystallinity of orange pulp, but antioxidant capacity and vitamin C content decrease as the hot-air drying temperature increases (Syeda et al., 2021). The VFD improves the physical properties, nutritional components, functional components, and antioxidant activity of berries such as blueberries and mulberries (Wang, Wen, Yang, & Li, 2023; Zhang et al., 2025). The MD improves the retention of phenolics, vitamins, and antioxidants in fruits (Ali et al., 2026). Ultrasonic and ultra-high-pressure-assisted drying significantly increases the content of pectin, hemicellulose, total phenolics, total flavonoids, phenolic acids, monosaccharides, and antioxidants in dried peaches (Zang et al., 2024). Steam explosion (SE) technology is an emerging pretreatment method that shows great promise in food processing due to its ability to effectively disrupt the natural barrier structures of materials (Wang, Lin, Yang, et al., 2023). Although studies have found that steam explosion can improve the functional properties of dietary fiber from orange peel, coffee husks, and hawthorn pomace, demonstrating high cholesterol adsorption, glucose adsorption, nitrite adsorption, and hydroxyl radical scavenging (Liu et al., 2021; Wang et al., 2015; Sun, Shao, et al., 2026). However, these studies used SE technology solely to extract bioactive components from fruit peels or pomace, rather than to supplement traditional drying techniques. Furthermore, no studies have yet been reported on the drying of fruits and vegetables using SE technology to supplement traditional drying methods. As different drying methods have varied effects, when selecting a drying technology suitable for blueberry pomace, it is important to consider not only product quality but also energy consumption, processing costs, and industrial feasibility (Li et al., 2026). Therefore, a comprehensive evaluation that balances product quality preservation and processing efficiency across different drying methods is crucial.

This study employed eight different drying methods, i.e., ND-S, ND-NS, HAD45, HAD160, MD, VFD, SE-HAD, and SE-VFD, to process blueberry pomace for a comprehensive evaluation of the quality characteristics of blueberry pomace. Through a comprehensive analysis of key indicators, including microstructure, color, bioactive components, antioxidant activity, enzyme inhibition rates, and flavor, the optimal drying technology was identified for various application scenarios. This work provides a scientific basis for selecting appropriate drying technologies and promoting the sustainable, high-value utilization of blueberry pomace.

2. Materials and methods

2.1. Materials

Blueberries were harvested in May 2025 from the blueberry farm of ZiYue Agricultural Technology Group Co., Ltd. and stored frozen at −20 °C until use. Prior to drying, the blueberries were thawed and thoroughly washed with deionized water. They were then processed using a juicer (JE-31, Ningbo Tails Electronics Industrial Co., Ltd., China). After being crushed and pressed, the blueberries were passed through a 2-mm-diameter sieve to obtain blueberry pomace. The blueberry pomace consisted primarily of fruit peel, pulp, and a small number of seeds. All reagents were of analytical grade.

2.2. Sample preparation

2.2.1. Natural drying (ND) of blueberry pomace

The ND was conducted under both direct sunlight and non-direct sunlight conditions. For natural sunlight drying (ND-S), 150 g of blueberry pomace was spread evenly on a tray with a thickness of 5 mm. The blueberry pomace was exposed to direct sunlight for 9 h, and the following parameters were recorded: light intensity, UV intensity, humidity, wind speed, atmospheric pressure, and temperature (perceived temperature and actual temperature) (Supplementary material 1). The samples were then placed in an oven at 45 °C to complete drying. The masses before and after drying were recorded. For no-sunlight natural drying (ND-NS), 150 g of pomace was spread evenly on a 5-mm-thick tray and shielded from direct sunlight using a light-blocking panel. The subsequent steps were identical to those for ND-S.

2.2.2. Microwave drying (MD) of blueberry pomace

For microwave drying (MD) 150 g blueberry pomace was divided into 10 equal portions. Each portion was spread in a glass dish and dried in a microwave oven (JC-101 W, Qingdao Juchuang Environmental Protection Equipment Co., Ltd.) at 800 W for 5 min. The mass before and after drying was recorded.

2.2.3. Vacuum freeze-drying (VFD) of blueberry pomace

The 150 g of blueberry pomace was pre-frozen at −20 °C for 12 h and then transferred to a vacuum freeze dryer (Eco mini, Nanjing GENSCI Co., Ltd., China) with a condenser temperature of −70 °C and an operating pressure of 5 Pa for 48 h.

2.2.4. Hot-air drying (HAD) of blueberry pomace

The HAD was performed at two different temperatures. For low-temperature hot-air drying, 150 g of blueberry pomace was spread evenly on a 5-mm-thick tray and dried in an oven (MGF—9240B, Shanghai Mulang Instrument Manufacturing Co., Ltd.) at 45 °C for 24 h. For high-temperature flash drying, 150 g of blueberry pomace was spread evenly on a 5-mm-thick tray and dried in an oven at 160 °C for 35 min. The dried samples were labeled HAD45 and HAD160, respectively.

2.2.5. Steam explosion (SE) of blueberry pomace

The SE was performed by placing 150 g of blueberry pomace into a steam-explosion machine (QBS-80, Qingzheng Eco-Technology Co., Ltd., China), setting the explosion pressure to 0.8 MPa and the duration to 10 s. The steam-burst-treated pomace samples were then placed in an oven at 45 °C for drying for 20 h, resulting in steam-burst-dried blueberry pomace powder (SE-HAD). The steam-burst-treated pomace was placed in a vacuum freeze-dryer for drying for 40 h, resulting in steam-burst-freeze-dried blueberry pomace powder (SE-VFD). All dried samples were ground using a grinder (FW100, Tianjin Tester Instrument Co., Ltd.) and sieved through an 80-mesh screen. The final samples were individually sealed in light-proof plastic bags and stored at −80 °C for further analysis. Each experimental group included three experimental replicates. All replicates were subsamples from one large homogenized batch of fresh blueberry pomace. Each subsample completed the full pretreatment-drying workflow independently.

2.3. Yield of blueberry pomace powder and moisture content

The masses of the blueberry pomace before and after drying were recorded, and the yield was calculated using the following formula:

$$\mathit{\Delta m}=\frac{m_2}{m_1}x100\%$$

where, m₁ and m₂ represent the masses (g) of the blueberry pomace before and after drying, respectively.

The moisture content was measured using the moisture analyzer (HE53, Mettler-Toledo Instruments Co., Ltd.).

2.4. Color, degree of browning, and microstructure

A colorimeter (CM-23d, Huangzhou Ke Sheng Instrument Co., Ltd., China) was used to measure the L, a, and b values of blueberry pomace powder. A digital camera was used to record the color and macroscopic morphology. The microstructure was observed using scanning electron microscopy (SEM). A small amount of sample was evenly spread on an adhesive pad and gold-sputter-coated. Observations were conducted at an acceleration voltage of 5.0 kV and a magnification of 500×. The degree of browning (DOB) was determined according to the method reported by Liu et al. (2025).

2.5. Determination of pH and soluble solids content

Exactly 0.5 g of blueberry pomace powder was weighed into a 15 mL centrifuge tube, dissolved in 10 mL of ultrapure water, and sonicated at 25 °C for 30 min. The mixture was then centrifuged (8000 rpm, 20 min) and the supernatant was collected for pH and soluble solids content analysis. The pH was measured using a calibrated pH meter (PHS-3C, Shanghai Precision Instrument Co., Ltd., China). The soluble solids content was determined using a refractometer (SW-55S, Guangzhou SWEVY Electronic Technology Co., Ltd., China).

2.6. Determination of total polyphenol content

Exactly 0.2 g of blueberry pomace powder was weighed into a 5 mL centrifuge tube and extracted with 4 mL of 80% ethanol. The mixture was sonicated at 25 °C for 30 min and centrifuged at 8000 rpm for 20 min. The supernatant was collected. Subsequently, 25 μL of the supernatant and 125 μL of 10% Folin's reagent were added to a 96-well plate and incubated for 10 min. Then, 125 μL of 7.5% Na₂CO₃ solution was added, and the mixture was incubated in the dark for 30 min. Absorbance was measured at 765 nm and total phenolic content was calculated using a standard curve prepared with gallic acid and expressed as gallic acid equivalents.

2.7. Determination of total flavonoid content

A 50 μL aliquot of the supernatant was placed in a 96-well plate, followed by the addition of 20 μL of 0.5 mol/L NaNO₂. After 5 min, 20 μL of 0.3 mol/L AlCl₃ was added and allowed to stand for 6 min. Then, 200 μL of 0.5 mol/L NaOH solution was added, mixed thoroughly, and allowed to stand for 15 min. Absorbance was measured at 510 nm. Total flavonoid content was calculated using a standard curve prepared with rutin and expressed as rutin equivalents.

2.8. Determination of total anthocyanin content

Total anthocyanin content was determined using the pH differential method. The supernatant of the sample was added with KCl buffer (pH = 1.0) and NaAC buffer (pH = 4.5), respectively, and allowed to stand in the dark for 15 min. Absorbance was then measured at 530 nm and 700 nm. The total anthocyanin content was calculated using the following formula and expressed as cyanidin-3-O-glucoside (C3G) equivalents:

$$\mathit{TA}\left(\mathit{mg}/g\right)=\frac{\left(A\times \mathit{Mw}\times \mathit{DF}\times V\times 1000\right)}{\mathit{Ma}\times L\times m}$$

where A (absorbance) = (A₅₃₀, _pH1.0 – A₇₀₀, _pH1.0) - (A₅₃₀, _pH4.5 - A₇₀₀, _pH4.5), Mw = 449.2 g/mol, DF = dilution factor, Ma (extinction coefficient) = 269,000 mol/L/cm, L (path length) = 1 cm, V = the total volume (mL), m = mass of blueberry pomace powder (g). The factor of 1000 was used to convert grams to milligrams. Results were defined as milligrams of C3G equivalents per gram of blueberry pomace powder.

2.9. Enzyme inhibition assays

The enzyme inhibition assays were modified according to the method described by Ruan et al. (2026). Exactly 0.5 g of blueberry pomace powder was weighed into a 15 mL centrifuge tube, dissolved in 10 mL of 0.02 M PBS (pH = 6.9), sonicated at 25 °C for 30 min, then centrifuged at 8000 rpm for 20 min. The supernatant was collected for analysis.

2.9.1. Α-amylase inhibitory activity assay

A 200 μL aliquot of the supernatant was mixed with 200 μL of α-amylase solution and incubated in a 37 °C water bath for 10 min. Then, 600 μL of 1% starch solution (preheated to 37 °C) was added, mixed thoroughly, and incubated at 37 °C for 10 min. The reaction was terminated by adding 1 mL of DNS reagent, followed by heating in a boiling water bath for 5 min. After cooling, the absorbance was measured at 540 nm. Acarbose (1 mg/mL) was used as the positive control.

2.9.2. Α-glucosidase inhibitory activity assay

The supernatant was diluted 3-fold with PBS. A 100 μL aliquot of the diluted sample was mixed with 100 μL of 0.2 U/mL α-glucosidase, and incubated at 37 °C for 10 min. Then, 300 μL of p-nitrophenyl-α-D-glucopyranoside (pNPG) was added and incubated at 37 °C for 15 min. The reaction was terminated by adding 500 μL of 0.2 mmol/L Na₂CO₃ solution. A sample blank was prepared to eliminate background interference from the sample color. A 200 μL aliquot of the reaction mixture was transferred to a 96-well plate, and absorbance was measured at 405 nm. Acarbose (1 mg/mL) was used as the positive control.

2.9.3. Lipase inhibitory activity assay

The supernatant was diluted 3-fold with PBS. A 100 μL aliquot of the diluted sample was mixed with 100 μL of 30 U/mL lipase and 1 mL of p-nitrophenyl palmitate (pNPP) and incubate at 40 °C for 10 min. The reaction was terminated by adding 500 μL of 1% SDS solution. A sample blank was prepared to eliminate background interference. A 200 μL aliquot of the mixture was transferred to a 96-well plate, and absorbance was measured at 405 nm. Orlistat (1 mg/mL) was used as the positive control. The inhibition rate was calculated as:

$$\text{Inhibition}\left(\%\right)=\left[1-\left({\mathrm{A}}_{\text{Sample}}–{\mathrm{A}}_{\text{Sample blank}}\right)/\left({\mathrm{A}}_{\text{Control}}-{\mathrm{A}}_{\text{Blank}}\right)\right]\times 100\%.$$

2.10. Determination of antioxidant activity

The supernatant obtained as described above was diluted 100-fold with deionized water to obtain the test solution. DPPH and ABTS radical scavenging activity were determined strictly according to the methods described in our previous study (Sun, Lu, et al., 2026). Ferric Reducing Antioxidant Power (FRAP) was assessed as follows: The FRAP working solution was prepared by mixing 0.3 mol/L sodium acetate buffer (pH 3.6), 10 mmol/L TPTZ solution, and 20 mmol/L FeCl₃ solution in a 10:1:1 volume ratio. A 500 μL aliquot of the diluted sample solution was mixed with 500 μL of FRAP working solution and incubated in a 37 °C water bath for 10 min. Absorbance was measured at 593 nm. A standard curve was plotted using Fe₂SO₄ standard solutions, and the results were expressed in Fe²⁺ equivalents.

2.11. Non-targeted metabolomics analysis

2.11.1. Metabolite extraction

A 10 mg sample was weighed into a 2-mL centrifuge tube, and 1 mL of pre-chilled methanol: acetonitrile (1:1, v/v) was added. The mixture was vortexed for 30 s and centrifuged at 12,000 rpm and 4 °C for 10 min. An 850 μL aliquot of the supernatant was transferred and concentrated to dryness under vacuum. The residue was redissolved in 150 μL of 50% methanol (containing 5 ppm 2-chlorophenylalanine), then filtered through a 0.22 μm filter membrane into a detection vial. A 10 μL aliquot from each sample was pooled to prepare a single quality control (QC) sample, which was used to evaluate instrument stability and data reliability.

2.11.2. LC-MS/MS analysis

Chromatographic separation was performed on an ACQUITY UPLC HSS T3 column (100 Å, 1.8 μm, 2.1 mm × 100 mm) at a flow rate of 0.4 mL/min, with a column temperature of 40 °C and an autosampler temperature of 8 °C. Prior to formal injection, 2–4 QC samples were injected to equilibrate the system; during the injection process, one QC sample was injected every 6–12 samples for quality control. The mobile phases consisted of (A) 0.1% formic acid in water and (B) acetonitrile with 0.1% formic acid for positive and negative ion modes. Mass spectrometry data were acquired using a Thermo Q Exactive HF-X mass spectrometer controlled by Xcalibur software (version 4.7, Thermo Fisher Scientific) in data-dependent acquisition (DDA) mode. A heated electrospray ionization (HESI) source was used with a spray voltage of 3.5 kV, sheath gas flow of 40 arb, auxiliary gas flow of 10 arb, capillary temperature of 320 °C, auxiliary gas temperature of 300 °C, first-stage resolution of 60,000, scan range 70–1000 m/z, AGC Target Standard, Max IT 100 ms, secondary fragmentation of the top 10 ions prior to the screening response, dynamic exclusion time 4 s, secondary resolution 15,000, and HCD collision energy 30%.

2.11.3. Database query and data analysis

Raw mass spectrometry data were imported into the MS-DIAL software (version 4.9.221218) for peak extraction, alignment, filtering, and metabolite identification. Metabolite identification was performed using the PSNGM (PerSonalbio Next-Generation Metabolomics Database). Peaks not detected in more than 50% of the QC samples were filtered out; for undetected peaks, missing values were imputed using the software's gap-filling algorithm, followed by normalization. The main database search parameters were as follows: MS1 tolerance for identification 0.01, MS2 tolerance for identification 0.05, smoothing level 3, minimum peak height 10,000, minimum peak width 5, mass slice width 0.05, and identification score cutoff 70. Metabolites with a VIP value >1, a P-value <0.05, and an FDR < 0.05 were considered significantly different. OPLS-DA was used to analyze the differences among the groups. Pathway analysis of the differentially expressed metabolites was performed using the KEGG database (https://www.kegg.jp). For machine learning analysis, we employed the recursive feature elimination (RFE) algorithm for feature selection within the random forest framework, measuring the importance of each metabolite in the random forest model by calculating the classification error (classif.ce). The most important features were selected by continuously removing the least important ones until the desired number of features was reached.

2.12. Electronic nose analysis

Exactly 2 g of blueberry pomace powder was weighed into a 20 mL headspace vial. The vial was equilibrated in a temperature-controlled water bath for 10 min to stabilize the volatile compounds. The probe of the electronic nose (PEN3, Airsense Analytics GmbH, Schwerin, Germany) was purged with synthetic air for 150 s to establish a stable baseline, followed by a 5 s sensor reset. The instrument was equipped with 10 sensors (Supplementary material 2). The pre-sampling time was set to 5 s, and the injection flow rate was 300 mL/min, controlled by a mass flow controller. The sampling measurement cycle was 120 s, with a data acquisition rate of 1 Hz.

2.13. Analysis of volatile compounds

The volatile components were analyzed using headspace solid-phase microextraction coupled with gas chromatography–mass spectrometry (HS-SPME/GC–MS). Exactly 1 g of blueberry pomace powder was weighed into a 20 mL headspace vial, and 10 μL of a 100 μg/mL 2-octanol solution and 5 mL of saturated NaCl solution were added. The mixture was then heated at 60 °C for equilibration, and volatiles were extracted using a SPME fiber at the same temperature for 30 min. GC–MS (Agilent, USA) analysis was performed using an HP-5 column. The initial temperature was set to 50 °C and held for 2 min; the temperature was then increased at a rate of 5 °C/min to 180 °C and held for 5 min; subsequently, the temperature was increased at a rate of 10 °C/min to 250 °C and held for 5 min. The carrier tube temperature was 280 °C, and the carrier gas flow rate was 1.0 mL/min. Mass spectrometry conditions were: ion source temperature 230 °C, quadrupole temperature 150 °C, EI ion source, and full-scan mode was employed in the 40–600 m/z range. Tentatively identified volatile compounds were compared with the NIST 17.0 spectral database. Data analysis was performed using MassHunter software (Agilent, USA).

2.14. Molecular docking analysis

Receptor-ligand molecular docking simulations were performed using Discovery Studio 2019 software. First, the crystal structure of the target protein (PDB ID: 8D43; PDB ID:1GPL) was obtained from the RCSB PDB (https://www.rcsb.org/) database. Using the “Prepare Protein” module, water molecules, impurity ions, and the native ligand were removed; amino acid residues were completed; and hydrogen atoms were added. The catalytic pockets of 8D43 and 1GPL do not contain any cocrystallized organic small-molecule ligands. Potential binding pockets were automatically found based on the geometric cavities of the protein using Discovery Studio 2019's “From Receptor Cavities” feature under the “Define and Edit Binding Site” tool. Small-molecule ligands were obtained from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/). The receptor-ligand molecular docking was performed using the CDOCKER flexible docking protocol, with the receptor backbone fixed rigidly and the side chains within the pocket set to be flexible. A full-conformational flexible search was enabled for the ligand, and the optimal binding conformation was output after high-temperature annealing and local optimization. The ligand interactions module was used to analyze intermolecular forces between the ligand and surrounding amino acid residues, including hydrogen bonds, hydrophobic interactions, π-bond interactions, and van der Waals forces. For CDOCKER, α-glucosidase (sphere center coordinates: X = 188.67, Y = 195.537, Z = 156.84 Å, radius = 13.6 Å) and lipase (with sphere center coordinates X = 31.2915, Y = 41.7128, Z = 57.9197 Å, radius = 7 Å) binding pockets were defined using site spheres. Uniform docking parameters were applied to both proteins: number of best matches = 10, number of random conformations = 10, and number of orientations to be optimized = 10. For each ligand, the conformation with the lowest CDOCKER energy was selected for ligand-receptor interaction analysis.

2.15. Comprehensive evaluation of different drying methods

Nine parameters—including enzyme activity inhibition rates, antioxidant capacity, and functional bioactive components—of blueberry pomace powder processed using eight different drying methods were subjected to range-standardization, and X_ij was obtained through data transformation. W_j was calculated using the entropy method, with weights determined in accordance with the approach described by Cao, Zhang, and Tong (2024). The formula for the comprehensive evaluation score (S_i) is as follows:

$$s_i={\Sigma }_{j=1}^n{x}_{\mathit{ij}}\ast w_j$$

where i represents the number of samples; j represents the number of the core indicators (j = 1,2,3…n).

2.16. Statistical analysis

All experiments were performed in triplicate. All data are expressed as mean ± standard deviation (mean ± S.D.). Experimental data were analyzed using SPSS 21.0 software (SPSS Inc., Chicago, IL, USA). Duncan's multiple range test was performed using one-way analysis of variance (ANOVA). The significance level of P < 0.05 was considered statistically significant.

3. Results and discussion

3.1. The effect of different drying methods on the basic quality of blueberry pomace powder

The macroscopic morphology of the blueberry pomace differed significantly after different drying treatments (Fig. 1A). The MD treatment had the greatest impact on the macroscopic morphology of the blueberry pomace, primarily in terms of color and microstructure. SEM observations showed that the blueberry pomace powder treated with ND-S formed large, dense agglomerates; the powder treated with ND-NS appeared as irregular, loose lumps, with a much lower degree of agglomeration than that of the ND-S-treated sample and no obvious dense, cohesive layers. After HAD45 treatment, the blueberry pomace powder appeared as intact, irregular lumps with relatively smooth surfaces and no obvious structural fragmentation. After HAD160 treatment, it exhibited sharply angular, fragmented particles with severe structural damage. After VFD treatment, the material exhibited a dense structure with a layered appearance. After MD treatment, it exhibited an irregular layered stacking structure with a dense surface and distinct shrinkage wrinkles. After SE-VFD treatment, the material consisted of extremely fine, amorphous powder with good uniformity. After SE-HAD treatment, the material appeared as irregular lumps with signs of fragmentation on the surface and no intact agglomerates (Fig. 1B).

Fig. 1.

Effects of different drying methods on the basic quality of blueberry pomace powder. (A) Macroscopic morphology; (B) Microstructure by SEM (500×); (C) Yield analysis; (D) Moisture content analysis; (E) pH and soluble solids content. The different lowercase letters in the graphs indicate significant differences between treatments (p < 0.05).

Table 1 shows the color parameters (L, a, b) and degree of browning (DOB) of the samples under different drying methods. SE-VFD exhibited the highest L value (47.17), indicating superior lightness, whereas ND-S, HAD45, and HAD160 yielded the lowest L values. The a value, proxy for redness, was highest for SE-VFD (11.74) and lowest for MD (3.69). While the b value was highest for HAD160 (7.44) and lowest for HAD45 (1.95). In terms of DOB, MD (0.274) and HAD160 (0.200) exhibited the most severe browning, consistent with accelerated Maillard and caramelization reactions at elevated temperatures. In summary, SE-VFD and VFD were most effective at preserving the native blue-purple color, whereas MD and HAD160 promoted extensive browning.

Table 1.

The effect of different drying methods on the color of blueberry pomace powder.

drying methods	L*	a*	b*	DOB /(Abs/g.d.)
ND-S	33.98 ± 1.47 d	9.93 ± 0.52 b	5.58 ± 0.28 b	0.099 ± 0.005 cd
ND-NS	39.74 ± 0.74 b	9.11 ± 0.23 cd	3.95 ± 0.44 c	0.096 ± 0.002 d
HAD45	32.96 ± 1.24 d	6.43 ± 0.29 f	1.95 ± 0.29 e	0.102 ± 0.002 cd
HAD160	34.48 ± 2.56 d	8.82 ± 0.74 de	7.44 ± 0.85 a	0.200 ± 0.017 b
VFD	39.78 ± 1.45 b	9.73 ± 0.62 bc	3.05 ± 0.36 d	0.117 ± 0.001 c
MD	37.12 ± 0.62 c	3.69 ± 0.13 g	3.82 ± 0.25 cd	0.274 ± 0.010 a
SE-VFD	47.17 ± 0.98 a	11.74 ± 0.41 a	6.18 ± 0.32 b	0.119 ± 0.003 c
SE-HAD	37.88 ± 0.99 bc	8.14 ± 0.15 e	3.24 ± 0.33 cd	0.102 ± 0.002 cd

To better compare the effects of different drying methods on blueberry pomace, we controlled the yield of each drying method to approximately 15.5% (Fig. 1C). In addition, our analysis using the moisture analyzer revealed that there were no significant differences in moisture content among the groups (Fig. 1D). In terms of pH, MD yielded the highest value (3.82); followed by HAD160 and SE-VFD, while ND-NS had the lowest (3.39) (Fig. 1D). In terms of soluble solids, VFD yielded the highest value (2.98), followed by HAD45 (2.80), while MD was lowest (1.45) (Fig. 1E).

Drying method profoundly alters the color, texture, pH, and soluble solids content of blueberry pomace (Ma et al., 2026). High temperatures may accelerate chemical reactions in the material (such as the Maillard reaction and caramelization), which typically result in darker colors, reduced saturation, decreased chroma, and lower soluble solids content (Lin et al., 2025). Air drying above 50 °C promotes the oxidation and degradation of anthocyanins, shifting the color from blue-purple to brown. Furthermore, higher cell shrinkage rates and lower porosity have been observed at high temperatures (Ma et al., 2026). Microwave drying causes surface cracking, leading to the loss of original morphology and natural color (Zia & Alibas, 2021). In contrast, SE offers higher energy and drying efficiency, without oxidative or combustion side reactions, achieving higher product quality (Marie et al., 2016); during SE treatment, hemicellulose-derived soluble sugars are released, increasing interparticle adhesion (Li et al., 2007). Drying under vacuum preserves the physical quality and color of most foods (Xu et al., 2020). Mechanistically, high temperatures disrupt cell walls and increased consumption of organic acids during the respiratory cycle, leading to a rise in pH, whereas VFD causes a decrease in pH (Kaveh et al., 2021). These findings are consistent with the results of our study.

3.2. The effects of different drying methods on functional components

As shown in Fig. 2A–C, different drying methods significantly affected the total polyphenol, total flavonoid, and total anthocyanin contents of blueberry pomace. Both SE-VFD and SE-HAD significantly increased the total polyphenol content compared with VFD and HAD45, by 24.46 and 22.38 mg/g, respectively. This enhancement is attributable to the cell-wall disruption induced by steam explosion, which liberates bound phenolics from the insoluble matrix (Wang et al., 2023). Studies have shown that SE can alter the matrix structure and affect the release of polyphenols, thereby triggering their transformation (Wan et al., 2022). VFD treatment yielded the highest total anthocyanin content (792.14 mg/g), followed by HAD45 (578.38 mg/g), reflecting the superior preservation of thermolabile pigments under low-temperature, low-oxygen conditions. Conversely, MD treatment resulted in the lowest levels of total polyphenols (17.11 mg/g), total flavonoids (19.07 mg/g), and total anthocyanins (41.47 mg/g), consistent with rapid thermal degradation under microwave irradiation. The total anthocyanin-preserving advantage of VFD aligns with the findings of Zhang et al. (2025), and the polyphenol-enriching effect of SE corroborates reports on pomegranate peel (Wang et al., 2023) and sea buckthorn pomace (Tu et al., 2018).

Fig. 2.

Effects of different drying methods on bioactive compounds and antioxidant capacity. (A) Total phenolic content; (B) Total flavonoid content; (C) Total anthocyanin content; (D) DPPH radical scavenging activity; (E) ABTS radical scavenging activity; (F) Fe²⁺ reduction capacity. The different lowercase letters in the graphs indicate significant differences between treatments (p < 0.05).

3.3. The effect of different drying methods on antioxidant capacity

The antioxidant capacity of blueberry pomace powder varied markedly across drying treatments (Fig. 2D–F). SE-VFD and HAD45 achieved the highest DPPH scavenging efficiency (50.68% and 49.52%, respectively), followed by VFD (41.84%), while MD (32.86%) performed poorest. Regarding ABTS scavenging efficiency, SE-VFD, HAD45, and VFD all demonstrated strong scavenging capabilities, at 62.13%, 57.57%, and 56.80%, respectively, while SE-HAD (49.54%) ranked next. HAD45 exhibited the strongest FRAP value (23.84 mmol Fe²⁺/L), followed by VFD and SE-VFD, whereas ND-S was weakest. The superior radical-scavenging performance of vacuum-dried samples are consistent with prior reports (Tan et al., 2020; Zhang et al., 2025), and the enhancement of DPPH and hydroxyl radical scavenging by SE corroborates findings in sea buckthorn pomace (Tu et al., 2018). These convergent results underscore that both low-temperature vacuum environments and steam explosion pretreatment preserve and even augment the antioxidant capacity of blueberry pomace.

3.4. The effect of different drying methods on enzyme activity

All drying treatments exhibited inhibitory activity against pancreatic lipase, α-glucosidase, and α-amylase (Fig. 3A-C). Acarbose inhibited α-glucosidase and α-amylase by 98.98% and 98.89%, respectively. Orlistat inhibited lipase by 95.37%. VFD and HAD45 showed the highest lipase inhibition rates (32.82% and 31.06%, respectively), followed by SE-VFD and SE-HAD (29.30% and 29.11%, respectively). Notably, SE-VFD exhibited the highest inhibition rate for α-glucosidase, reaching 96.37%, followed by SE-HAD (89.81%). VFD and SE-VFD exhibited the best inhibitory effects on α-amylase (77.15% and 70.04%, respectively), followed by HAD45, whereas MD showed the poorest inhibitory effects on both α-glucosidase and α-amylase. Correlation analysis (Fig. 3D) revealed that total polyphenol and total flavonoid contents were significantly and positively correlated with α-glucosidase and α-amylase inhibitory activities, while lipase inhibitory activity was positively correlated only with pH. Total anthocyanins were significantly correlated with ABTS, DPPH, FRAP, and α-amylase inhibitory activities. Mechanistically, low drying temperatures favor the retention of phenolic inhibitors of α-glucosidase and α-amylase (Wongsa et al., 2023), and freeze-dried products exhibit enhanced lipase inhibition, which is correlated with anthocyanin content (Fabroni et al., 2016). The elevated α-glucosidase inhibition by SE-treated samples is consistent with the report of Wang et al. (2023) and may be mediated by SE-induced release of flavonoid and polyphenolic inhibitors from the cell-wall matrix (Huang, Zhao, et al., 2025).

Fig. 3.

Effects of different drying methods on enzyme inhibitory activities. (A) Lipase inhibition rates; (B) α-glucosidase inhibition rates; (C) α-amylase inhibition rates; (D) Correlation analysis. The different lowercase letters in the graphs indicate significant differences between treatments (p < 0.05).

3.5. Flavor analysis of different drying methods

3.5.1. Analysis of electronic noses using different drying methods

The results of the odor characteristic data for different drying treatments are shown in Fig. 4A–B. Under all drying treatments, neither the W6S nor the W3S sensors exhibited a significant response. The VFD treatment elicited more pronounced responses from the W1C, W3C, W5C, W1S, and W2S sensors; the HAD160 and MD treatments elicited significant responses in the W5S and W1W sensors. This suggests that aromatic compounds, aromatic constituents (primarily ammonia), alkanes, aromatic compounds, broad methane, and broad alcohols are associated with sustained low-temperature drying environments. Nitrogen oxides and sulfur compounds are associated with high-temperature drying environments. These findings are consistent with the results reported by Chen et al. (2026). E-nose analysis outlined the differences in the odor of blueberry pomace subjected to different drying treatments, confirming significant differences in volatile organic compounds (VOCs). However, since specific compounds could not be distinguished, SPME-GC–MS was subsequently used to identify and characterize precise VOCs to supplement the E-nose results.

Fig. 4.

Flavor analysis of blueberry pomace powder under different drying methods. (A–B) Electronic nose analysis; (C) Cluster analysis of total VOCs; (D) Frequency distribution of VOCs; (E) PCA analysis of VOCs profiles; (F) Heatmap of the top 50 significantly different VOCs.

3.5.2. SPME-GC–MS analysis under different drying conditions

A total of 1687 VOCs were tentatively identified across all samples (Fig. 4C). Among all detected VOCs, esters were the dominant category with 247 compounds, excluding the “Others” category. Samples subjected to different drying treatments also had a significant impact primarily on esters, excluding the “Others” category. MD treatment primarily affected ethers, VFD treatment primarily affected esters, and ND-S treatment primarily affected alcohol and acids (Fig. 4D).

PCA of the VOC profiles revealed clear separation among drying treatments with MD, SE-VFD, and ND-S exerting the greater perturbation (Fig. 4E). To further identify the significantly different VOCs, the top 50 significantly altered compounds were visualized as a heatmap (Fig. 4F). Compared with HAD45, SE-HAD significantly increased the concentrations of benzenemethanethiol; tridecane-3-methyl-; phenol-4-butyl-; 3-cyclohexene-1-ethanol-β,4-dimethyl; and dihydro-5-pentyl-2(3H)-furanone, among others. This finding aligns with Zhang et al. (2022) found that SE increases the yields of higher alcohols and aldehydes. Compared with VFD, SE-VFD significantly increased 5-hydroxymethylfurfural; 2-pentylfuran; pentamethylbenzene; and 4-ethenyl-1,2-dimethylbenzene. As shown in Supplementary material 3, the steam-bursting technique can significantly increase the yields of 1H-Indene, 2,3-dihydro-4,6-dimethyl-; 5-azulenemethanol, 1,2,3,4,5,6,7,8-octahydro-α,α,3,8-, and other VOCs. These treatment-specific VOC signatures likely arise from differential activation of fatty acid oxidation, degradation, and Maillard pathways under distinct temperature and pressure regimes (Zhang et al., 2025).

3.6. Non-targeted metabolomics analysis using different drying methods

Untargeted metabolomic analysis revealed significant inter-group differences and good intra-group reproducibility in both the POS and NEG modes (Fig.S1 A–D). A total of 2963 metabolites were identified across all groups (Supplementary material 4), primarily classified as lipids and lipid-like molecules, organic heterocyclic compounds, and phenylpropanoids and ketones (Fig. S1 E–F). Among these, 2025 differentially expressed metabolites (DEMs) were identified, and 2012 DEMs were selected after FDR filtering (Supplementary material 5). In the OPLS-DA analysis (for the POS pattern: R²X = 0.362, R²Y = 0.285, and Q² = 0.13; for the NEG pattern: R²X = 0.353, R²Y = 0.285, and Q² = 0.128), the MD group and the HAD160 group showed more significant differences compared to the other groups (Figs. 5A–B), implicating sustained high temperatures as the dominant source of metabolic perturbation. Trend analysis of the DEMs revealed that they could be classified into 9 categories (Fig. 5C). Correlation analysis shows that there are significant correlations among substances such as phenylpropanoids and polyketides, organic acids and derivatives, benzenoids, and organic oxygen compounds (Fig. 5D). Machine learning analysis of the DEMs revealed that compounds such as MP16485, MN4763, and MN10673 were of higher importance (Fig. 5E). Finally, KEGG pathway enrichment analysis revealed that pathways such as ABC transporters, flavonoid biosynthesis, aminoacyl-tRNA biosynthesis, and flavone and flavonol biosynthesis were significantly enriched (Fig. 5F), consistent with the drying-induced metabolic shifts reported by Zhang et al. (2025).

Supplementary Figure S1.

Fig. S1

Fig. 5.

Non-targeted metabolomics analysis of different drying methods. (A–B) PCA of DEMs in positive and negative ionization modes; (C) Trend analysis of DEMs; (D) Correlation analysis of DEMs; (E) Machine learning feature importance analysis; (F) KEGG pathway enrichment analysis of DEMs.

3.6.1. The effects of blueberry pomace by steam explosion

OPLS-DA revealed that, regardless of whether in POS (R²X = 0.511, R²Y = 0.997 and Q² = 0.852) or NEG (R²X = 0.521, R²Y = 0.993 and Q² = 0.852) mode, SE primarily had a greater impact on the samples' PCA1 component (Fig. 6A–B). Cluster analysis of the DEMs identified MP8657, MP4741, MN19596, MP7309, MP12269, and MN8463 as having higher importance. SE increased the content of MN19596 and MN8463 (Fig. 5C). KEGG pathway enrichment analysis revealed that pathways such as flavonoid biosynthesis, ABC transporters, aminoacyl-tRNA biosynthesis, flavone and flavonol biosynthesis, and biosynthesis of amino acids were significantly enriched (Fig. 5D). The SE increases the yield of free phenolics and free flavonoids in pomace and enhances the abundance of compounds such as epicatechin (Cui et al., 2023; Dorado et al., 2021).

Fig. 6.

Metabolomic analysis of SE effects on blueberry pomace powder. (A–B) PCA analysis in positive and negative ionization modes; (C) Machine learning feature importance analysis of SE-responsive DEMs; (D) KEGG pathway enrichment analysis of SE-responsive DEMs; (E) DEMs in the flavone and flavonol biosynthesis pathways under different drying methods. Red represents DEMs, black represents non-DEMs; solid arrows indicate confirmed pathways, dashed arrows indicate uncertain pathways. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.6.2. The effect of temperature on blueberry pomace powder

To investigate the effects of different drying temperatures on the metabolites in blueberry pomace powder, we performed a machine learning analysis on the DEMs from the HAD45 and HAD160 groups and found significant changes in the levels of compounds such as MP17085, MN5354, MP10612, and MP11511, which were also found to be more important (Fig. S2 A). KEGG pathway enrichment analysis revealed that pathways such as aminoacyl-tRNA biosynthesis, biosynthesis of amino acids, ABC transporters, flavonoid biosynthesis, and flavone and flavonol biosynthesis were significantly enriched (Fig. S2 B). Heating affects the content of anthocyanins and proanthocyanidins in blueberry pomace (Ramesh et al., 2010). The retention of phenolic compounds in pomace is governed by the interaction of temperature, oxygen exposure, and residence time; HAD treatment causes a significant decrease in flavonoids, phenolic acids, and other compounds while also affecting the levels of free amino acids (Ali et al., 2026; Yin et al., 2022).

Supplementary Figure S2.

Fig. S2

3.6.3. The effect of light exposure on blueberry pomace powder

To assess the influence of light exposure, machine learning analysis was applied to DEMs from the ND-S and ND-NS groups. Compounds MP16883, MN13708, MN5264, MN9068, and MN10189 showed the most significant changes and highest feature importance (Fig. S2 C). KEGG enrichment identified significant enrichment of ABC transporters, purine metabolism, amino acids biosynthesis, and flavone and flavonol biosynthesis pathway (Fig. S2 D). Shade treatment increases the content of protein, amino acids, vitamin C, ash, fiber, and acidity (Kolla et al., 2021), and enriched flavonoids, anthocyanins, hydrocarbons, and alkaloids, alongside flavonoid biosynthesis, ABC transporters, arginine biosynthesis, and aminoacyl-tRNA biosynthesis pathways (Wu et al., 2022), consistent with the protective effect of light exclusion on photolabile metabolites. Comparative analysis between the ND-S and HAD45 groups showed substantial changes in metabolites MP15610, MN12351, and MN5459 (Fig. S2 E). KEGG pathway enrichment analysis highlighted enrichment of the flavone and flavonol biosynthesis pathway (Fig. S2 F). Distinct metabolite profiles, including MP14298, MN12767, and MN19502, were observed in the ND-NS group relative to the HAD45 group (Fig.S2 G), with KEGG enrichment revealing significant enrichment of ABC transporters and flavone and flavonol biosynthesis (Fig. S2 H). Huang et al. (2024) similarly found that the metabolites produced after hot-air drying are primarily involved in flavone and flavonol biosynthesis. In addition, studies have found that, compared with natural drying, low-temperature hot-air drying exhibits distinct lipid- and flavonoid-related metabolic profiles (Yang et al., 2026). Taken together, these results demonstrate that each drying method exerts a distinct signature on flavonoids and flavanols biothesis. HAD160 enriches isovitexin. MD enriches astragalin. SE preferentially enriched isoquercetin, rutin, kaempferol-3-O-xyloside, and kaempferol-3-glucoside-6″-p-coumaroylglucoside within the flavonoid and flavonol biosynthesis pathways (Fig. 6E). Since total flavonoid content is quantified on a rutin-equivalent basis, this further validates the observation that SE increases total flavonoid content. Notably, the bioactivity of individual flavonoids differs substantially: Barber et al. (2021) demonstrated that quercetagetin inhibits α-glucosidase more potently than kaempferol or galangin, while Yan et al. (2009) established the inhibitory quercetin > isoquercetin > rutin, providing a structure-activity context for interpreting the enzyme-inhibitory differences among drying treatments.

3.7. The effect of key DEMs on enzyme activity of human origin

Metabolomic analysis revealed that the flavonoid biosynthesis and flavone and flavonol biosynthesis pathways were significantly enriched in all drying treatments, particularly in SE. SE increased the content of total flavonoids and total polyphenols. Therefore, we further analyzed flavonoids and phenols and found that the major compounds with high total scores were epicatechin, procyanidin B1, 4-hydroxy-3-methoxycinnamaldehyde, 3,5-dimethoxy-4-hydroxycinnamaldehyde, procyanidin B3, and homoeriodictyol (Supplementary material 6). Correlation analysis (Fig. 7A) revealed that flavonoids (epicatechin, procyanidin B1, procyanidin B3, and homoeriodictyol) showed a significant positive correlation with α-glucosidase, DPPH, and ABTS (P < 0.01), while phenols (4-hydroxy-3-methoxycinnamaldehyde, 3,5-dimethoxy-4-hydroxycinnamaldehyde) showed significant positive correlations with lipase, DPPH, and ABTS (P < 0.01). We believe that while the Pearson correlation coefficient can indicate a statistical association between metabolite abundance and biological activity, it cannot prove a causal relationship. The drying process alters hundreds of metabolites simultaneously; multiple compounds may act in concert to influence enzyme inhibitory activity.

Fig. 7.

Effects of key DEMs on enzyme activity. (A) Correlation analysis of key DEMs with enzyme inhibitory activities and antioxidant capacity; (B–E) Molecular docking analysis of Procyanidin B1 (B), Procyanidin B3 (C), Epicatechin (D) and Homoeriodictyol (E) with α-glucosidase.

Molecular docking was performed to elucidate the binding modes of the six key DEMs with α-glucosidase and lipase. Among these, procyanidin B1 and procyanidin B3 exhibited the lowest CDOCHER-ENERGY and CDOCHER INTERACTION-ENERGY values for α-glucosidase binding (Supplementary material 7), indicating the most stable complexes. Procyanidin B1 engaged multiple α-glucosidase residues including PRO146, ASN97, GLU162, PRO370, ASP374, SER371, HIS308, and GLN735 via conventional hydrogen bonds, van der Waals interactions, π-anion contacts, and π-alkyl interactions (Fig. 7B). Procyanidin B3 formed hydrogen bonds and van der Waals interactions with the core residues GLU162, ARG309, and ARG374, and forms a π-alkyl bond with PRO370 (Fig. 7C). Epicatechin hydrogen-bonded with HIS308, ASP310, PRO370, and GLU162 and formed a π- alkyl bond with PRO370 (Fig. 7D); homoeriodictyol exhibited a comparable binding mode. The convergence of these interactions at the α-glucosidase active site, particularly the recurrent involvement of GLU162 and PRO370, provides a potential structural basis for the potent α-glucosidase inhibition observed in SE-VFD samples, which are enriched in procyanidins B1 and B3. Furthermore, we believe that to demonstrate the digestive enzyme inhibitory activity of these compounds, further in vitro experiments using standard samples should be conducted to confirm their actual contribution to enzyme inhibition.

These docking results indicate that these compounds exhibit favorable binding interactions with α-glucosidase. Studies have shown that flavonoids primarily inhibit α-glucosidase by forming hydrogen bonds with amino acid residues such as GLU, ARG, and ARG at different sites (Li et al., 2023). Phenolic acids, in turn, inhibit pancreatic lipase in a reversible and competitive manner, primarily by destabilizing its α-helix and occupying the active-site residues such as HIS, PHE, LYS, and SER (Cao, Mei, et al., 2024). In the present study, 3,5-Dimethoxy-4-hydroxycinnamaldehyde primarily interacts with lipase residues SER194, LYS163, GLU187, GLU188, and ASP195 via hydrogen bonds, van der Waals forces, and π-anion interactions, while 4-Hydroxy-3-methoxycinnamaldehyde bound GLU188 and SER194 via hydrogen bonds and van der Waals interactions (Fig. S3). By binding effectively to digestive enzymes involved in the hydrolysis of carbohydrates and lipids, these polyphenolic compounds may help alleviate obesity and improve dyslipidemia (He et al., 2023), highlighting the functional importance of the metabolite profiles formed by different drying methods.

Supplementary Figure S3.

Fig. S3

3.8. Comprehensive analysis of blueberry pomace powder processed using different drying methods

Integrating all quality, functional, and flavor metrics, SE-VFD emerged as the optimal treatment: produced low-browning powder with fine, uniform particle size; achieved the highest polyphenol and flavonoid content as well as antioxidant capacity (DPPH, ABTS, FRAP); and demonstrated an α-glucosidase inhibition rate as high as 96.37%. In terms of flavor, it is rich in volatile compounds such as furans and benzenes. Metabolomically, it significantly enriched flavonoid biosynthesis, accumulating bioactive monomers such as proanthocyanidins and epicatechins. Molecular docking simulations confirm that its polyphenols can stably bind to digestive enzymes to exert inhibitory effects. It is suitable for the production of high-quality blueberry pomace powder with specific functional properties (inhibitors of digestive enzymes). SE-HAD enhances phenolic compounds, but due to slight thermal oxidation, anthocyanin retention is lower than that of SE-VFD; color, microstructure, antioxidant activity, and enzyme inhibitory activity are above average, and flavor is superior to that of HAD and MD. Although SE can produce high-quality blueberry pomace powder, this process is extremely energy-intensive and may not be the preferred drying method for large-scale industrial production (An et al., 2024). The low-temperature vacuum environment of VFD maximizes the retention of anthocyanins and soluble solids, provides good lipase and α-amylase inhibitory effects, and offers good antioxidant capacity. However, freeze-drying equipment not only consumes a lot of energy but also involves long processing cycles and high production costs (An et al., 2024). The mild thermal environment of the HAD45 causes minimal damage to active compounds, resulting in good retention of anthocyanins and polyphenols, optimal FRAP reducing capacity, and outstanding lipase inhibitory activity; the samples retain their structural integrity, exhibit minimal agglomeration, and show no severe browning. Although the aerobic hot-air environment caused minor oxidation of phenolic compounds and some loss of flavor compounds, and its alpha-glucosidase inhibitory activity was weaker than that of the SE series and VFD, hot-air drying remains the most common post-processing method for agricultural products, with over 85% of industrial dryers employing this technology (Abhay et al., 2020). The ND-NS process involves no direct sunlight throughout, resulting in lower degradation of anthocyanins and flavonoids compared to ND-S, as well as lower browning. However, natural drying is highly dependent on weather conditions, involves a long drying cycle (An et al., 2024), carries a high risk of microbial contamination, and generally yields lower levels of polyphenols, antioxidant activity, and enzyme inhibitory activity. In ND-S, light exposure accelerates the oxidative degradation of anthocyanins (Ning et al., 2025) and flavonoids, leading to increased browning and a one-dimensional flavor profile, making it unsuitable for the production of functional ingredients. HAD160 and MD processes trigger intense Maillard and caramelization reactions due to high temperatures, resulting in extremely high browning levels and poor color. High temperatures cause significant degradation of polyphenols, flavonoids, and anthocyanins, leading to weak antioxidant and digestive enzyme inhibitory activity, as well as the production of undesirable flavors such as nitrogen oxides. These processes are only suitable for the processing of low-end by-products that do not require bioactivity and merely require rapid dehydration. Without considering energy costs, we conducted a comprehensive analysis using the entropy method on nine functional indicators (functional active ingredients, antioxidant capacity, and enzyme activity) and found that the performance rankings of the different drying methods were as follows: SE-VFD, HAD45, VFD, SE-HAD, ND-NS, ND-S, HAD160, and MD (Supplementary material 8). However, the choice of drying method should be determined based on the specific application scenario. The value-added utilization of food is also related to food packaging and related materials, in addition to drying processes (Said et al., 2026; Said & Andi, 2024). For blueberry pomace powder, if energy consumption is not a concern, the SE-VFD performs better in terms of functional quality. For industrial-scale production that still requires high quality and functional activity, the HAD45 may be a promising candidate. If the goal is an energy-efficient and environmentally friendly processing method, and time is not a factor, the ND-NS is the preferred choice. If time is the primary consideration, the HAD160 is recommended.

4. Conclusion

This study compared the effects of eight drying processes on the physicochemical properties, bioactive components, flavor, and digestive enzyme inhibition rate of blueberry pomace powder and elucidated the mechanisms of action using metabolomics and molecular docking. MD and HAD160 induced severe browning, significantly degraded polyphenols and anthocyanins, and exhibited the poorest antioxidant and digestive enzyme inhibitory activities; ND-S accelerates the oxidation of anthocyanins, leading to significant quality deterioration. VFD maximizes the retention of anthocyanins, while SE-VFD significantly enriches flavonoids and polyphenols, with an α-glucosidase inhibition rate of 96.37%; proanthocyanidins B1/B3 stably bind to α-glucosidase via hydrogen bonds and van der Waals forces, exhibiting the optimal digestive enzymes inhibitory activity. HAD45 strikes a balance between processing costs and functional activity, making it a promising candidate for industrial-scale production. Based on a comprehensive evaluation using the entropy method, SE-VFD exhibits the best overall quality but has relatively high energy consumption; in actual production, the drying process should be selected based on functional requirements and cost constraints.

The following are the supplementary data related to this article.

CRediT authorship contribution statement

Wu Sun: Writing – original draft, Investigation, Formal analysis, Data curation. Dong Huang: Methodology, Data curation. Yunfei Long: Investigation. Ming Bao: Validation, Data curation. Yuji Xu: Data curation. Hua Wei: Supervision, Conceptualization. Guoqiang Zhang: Writing – review & editing, Supervision, Resources, Funding acquisition. Xueying Tao: Writing – review & editing, Supervision, Resources, Funding acquisition.

Funding

This work was supported by the Jiangxi Province Graduate Student Innovation Special Foundation (YC2026-B036), Key Research and Development Projects in Anhui Province (No. 2023n06020045) and National Natural Science Foundation of China (grant no. NSF 32360026).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.