Loquat (Eriobotrya japonica cv. Dawuxing) flowers were harvested from Xiluodu Town, Yongshan County (103.64°E, 28.24°N), Zhaotong City, Yunnan Province, China, on November 25, 2024. Flowers were collected at the partially bloomed bud stage to ensure uniform maturity and physiological consistency. To minimize variability, samples were pooled from six healthy parent plants grown under identical environmental conditions. Collection was performed using sterilized scissors with careful handling to maintain structural integrity, and specimens were transported in sterile collection receptacles to prevent contamination and physical damage.

Laboratory processing involved an initial purification protocol wherein specimens underwent immersion in deionized water with gentle mechanical agitation to remove particulate contaminants, followed by moisture removal via sterile absorbent material and ambient air desiccation. The processed specimens were subsequently allocated into three experimental groups based on preservation methodology: refrigerated storage (4°C), thermal dehydration (60°C for 6 h until complete moisture removal), and lyophilization (preliminary freezing at −20°C followed by vacuum dehydration at −50°C for 48 h).

The extraction procedure employed thermal aqueous extraction at 90°C with 30-min duration using a standardized biomass-to-solvent ratio (1:20 w/v). Post-extraction, the solution underwent gravity separation for 6 h, facilitating supernatant isolation. The aqueous extract subsequently underwent lyophilization via preliminary freezing (−40°C) followed by vacuum dehydration (48 h), yielding a stable powdered extract amenable to long-term storage.

Quality control was maintained throughout the experimental procedure through utilization of precision instrumentation (temperature-controlled drying apparatus, lyophilization equipment, and ultra-low temperature storage) and implementation of aseptic techniques to preserve sample integrity. The resultant extracts (fresh, thermally dehydrated, and lyophilized) were hermetically sealed, appropriately labeled, and preserved for subsequent phytochemical analysis, particularly for flavonoid compound characterization and quantification.

The antioxidant capacity of variously processed loquat flower preparations was assessed using the 2,2-diphenyl-1-picrylhydrazyl (DPPH) free radical scavenging assay (34). The samples analyzed included fresh flowers, freeze-dried specimens, oven-dried specimens, and their corresponding hot-water-extracted powders.

Sample preparation involved homogenizing 0.1 g of each specimen in 1 mL of 80% methanol. The mixtures underwent ultrasonic-assisted extraction at 60°C for a 30-min period with periodic agitation, followed by centrifugation at 12,000 rpm to harvest the supernatant. The analytical procedure consisted of combining equal volumes (150 μL) of the extracted supernatant and DPPH working solution (prepared in ethanol) in microtubes, while the control preparation contained only methanol. Following a 30-min dark incubation period, spectrophotometric measurements were taken at 517 nm wavelength using a microplate reader. Antioxidant potency was expressed as micro-gram Trolox equivalents per gram of sample (μg TE/g) through quantification against a standard curve prepared with Trolox concentrations ranging from 0 to 25 μg/mL.

Multivariate analytical approaches were employed to examine flavonoid and anthocyanin distribution patterns alongside their corresponding gene expression profiles. Prior to analysis, both metabolomic and transcriptomic datasets underwent Z-score normalization. To distinguish between experimental groups while reducing intra-group variability, hierarchical cluster analysis (HCA) and principal component analysis (PCA) were conducted using OmicShare tools, an open-access online analytical platform (35).⁴ Visual presentations and standard error computations were prepared using Microsoft Excel (Redmond, WA, USA). For statistical evaluation of biochemical parameters, Statistix 8.1 software (Tallahassee, FL, USA) was employed, with all experimental procedures performed in triplicate. The least significant difference (LSD) test was applied to assess statistical significance of variations in antioxidant activity and total flavonoid and anthocyanin concentrations across loquat samples.

Variable importance in projection (VIP) values were determined using orthogonal partial least squares-discriminant analysis via the MetaboAnalystR package within the R statistical environment (61). Metabolites exhibiting VIP values above 1.0 were classified as significant discriminatory biomarkers for distinguishing experimental groups (62). The selection of significantly modified metabolites followed a comprehensive dual-criterion approach: statistical significance was assessed using Student’s t-test with false discovery rate adjustment (p < 0.05) (63), while biological significance was determined through fold-change thresholds (|log2FC| ≥ 1, corresponding to FC ≥ 2 or ≤ 0.5) (64). Venn diagram visualization was performed using the EVenn web-based tool (35).⁵

Venn diagram analysis of loquat flower metabolites revealed distinct compositional differences between sample groups. In different samples, 10 compounds were common across all groups, while pairwise comparisons showed unique metabolites: HD vs. FS had 3, FD vs. FS had 13, and FD vs. HD had 6 (Figure 1A). The total number of significantly altered compounds was highest in FD vs. FS (47), followed by FD vs. HD (45) and HD vs. FS (28), suggesting that processing methods significantly influence metabolite profiles. These variations may stem from differences in enzymatic activity, thermal degradation, or extraction efficiency, highlighting the importance of processing conditions in preserving bioactive compounds.

A similar trend was observed in processed samples, where 10 shared compounds were detected (Figure 1B). However, fewer unique metabolites were identified: HDP vs. FSP had 3, FDP vs. FSP had only 1, and FDP vs. HDP had 3 (Figure 1B). The total altered compounds were highest in FDP vs. HDP (28), followed by HDP vs. FSP (24) and FDP vs. FSP (11). This reduction in unique compounds compared to raw samples suggests that processing may homogenize certain metabolites, possibly due to thermal degradation or chemical interactions. These findings provide valuable insights into how different treatments affect loquat flower chemistry, which could guide optimal processing methods for retaining functional compounds in food applications.

The comparative metabolomic analysis of loquat flower samples revealed significant variations in metabolite composition across different processing and storage conditions (Figure 2). In raw samples, the FD (freeze-dried) vs. FS (fresh) comparison exhibited the highest number of upregulated compounds (39 out of 47), suggesting that freeze-drying better preserves or even concentrates certain metabolites compared to fresh samples (Figure 2). Conversely, HD (heat-dried) samples showed fewer upregulated compounds (16 in HD vs. FS), implying potential thermal degradation (Figure 2). Interestingly, FD vs. HD had 38 upregulated metabolites, reinforcing that freeze-drying retains more bioactive compounds than heat-drying.

Processed samples displayed contrasting trends: FDP (freeze-dried powder) vs. HDP (heat-dried powder) had more downregulated compounds (20 vs. 8-upregulated). Notably, FDP vs. FD showed a dramatic shift, with 74 upregulated and 24 downregulated metabolites—the highest overall change (98 total), indicating that freeze-drying powder drastically alters the metabolic profile. Meanwhile, HDP vs. HD had 60 upregulated and only 3 downregulated compounds, suggesting heat-dried powder extract stabilizes certain metabolites. These findings highlight that processing methods—particularly fermentation and drying techniques—profoundly influence loquat flower chemistry, with implications for optimizing functional food production.

Hierarchical cluster analysis (HCA) of loquat flower metabolites revealed distinct grouping patterns based on processing methods (Figure 3). HCA analysis revealed that the freeze-dried samples (FD, FDP) exhibited the closest similarity, suggesting that freeze-drying preserves metabolite profiles more consistently than heat treatments. In contrast, heat-dried samples (HD, HDP) clustered distinctly, likely due to thermal degradation of heat-sensitive compounds such as flavonoid glycosides. Fresh samples (FS and FSP) formed a separate group, indicating that drying—regardless of method—significantly alters the metabolite composition compared to untreated flowers.

Within clusters, powdered extracts (FSP, HDP, FDP) showed tighter grouping than their non-powdered counterparts, suggesting that the hot-water extraction and freeze-drying process homogenizes certain metabolites (Figure 3). For example, violanthin and syringetin-3-O-glucoside were highly abundant in freeze-dried powders (FDP), while 5,7,2′,5′-tetrahydroxyflavanone and quercetin derivatives dominated fresh samples. Heat-dried powders (HDP) showed reduced levels of thermally labile compounds like myricetin glycosides but retained higher concentrations of heat-stable metabolites such as tricin derivatives (Figure 3). These findings underscore that processing methods—particularly drying temperature and extraction—profoundly influence loquat flower chemistry, with freeze-drying better preserving bioactive flavonoids.

Principal component analysis (PCA) of loquat flower metabolites revealed distinct clustering patterns based on processing methods (Figures 4A,B). In compound-wise PCA (Figure 4A, PC1: 71.99%, PC2: 15.43%), most compounds clustered tightly along the x-axis and y-axis intersection point, indicating high metabolite consistency (Figure 4A). Fresh samples (FS and FSP) formed a separate group, highlighting significant compositional changes induced by drying. Key compounds like violanthin (FC2) and syringetin-3-O-glucoside (FC29) heavily influenced FD/FDP clustering, underscoring their stability under freeze-drying. In contrast, heat-sensitive flavonoids (e.g., myricetin glycosides (Myricetin-3-O-β-D-glucoside - FC16)) were reduced in HD/HDP, aligning with PC2’s separation of thermally processed samples. Notably, FC12 (Okanin-4′-(6”-O-acetyl)glucoside) and FC55 (Cyanidin 3-O-(6-O-p-coumaroyl)glucoside) were pivotal in differentiating FD/FDP, correlating with their preservation of antioxidant polyphenols.

The group-wise PCA (Figure 4B, PC1: 57.56%, PC2: 14.14%) further emphasized processing-driven divergence (Figure 4B). Powdered extracts (FSP, HDP, FDP) clustered separately than their non-powdered counterparts (FS, HD, and FD), suggesting that hot-water extraction and freeze-drying extraction procedure significantly altered the metabolic compositions (Figure 4B). PC1 clearly separated freeze-dried from heat-dried groups, while PC2 distinguished fresh samples, likely due to retained volatiles and unaltered glycosides. These results demonstrate that freeze-drying better maintains bioactive integrity, while heat processing introduces variability, critical for optimizing loquat flower-based functional foods.

The analysis of key bioactive metabolites in loquat flower samples revealed significant variations influenced by processing methods (Figure 5). Violanthin, a major flavonoid, showed the highest percentage abundance in hot-water powdered extract samples, with FSP (4.50%) > FDP (3.38%) > HDP (2.80%), suggesting that extraction and powdering increased compound concentration. Interestingly, fresh samples contained moderate levels (FS:1.01%), while freeze-drying (FD:1.21%) better preserved violanthin compared to heat-drying (HD:1.48%) (Figure 5). Okanin and eupatolitin derivatives showed moderate processing effects, with powdered forms generally having higher concentrations than their non-powdered counterparts, likely due to concentration during extraction (Figure 5). For methyl hesperidin, heat-dried samples (HD:10.04) contained dramatically higher levels than other samples, likely due to thermal-induced conversion of precursors, while freeze-drying (FD:0.57%) and powdered forms showed reduced amounts (Figure 5). Cyanidin, an important anthocyanin, exhibited an opposite trend, with freeze-dried samples (FD:1.66%) and powders (FSP:3.27%, FDP:2.67%) containing substantially higher levels than fresh (FS:0.38%) or heat-dried (HD:0.45%) samples, suggesting freeze-drying better preserves this heat-sensitive pigment (Figure 5).

Notably, syringetin and apigenin-7-O-rutinoside showed complex patterns. Syringetin was abundant across all samples but highest in heat-dried (HD:8.98%) and powdered heat-dried (HDP:8.61%) forms, indicating thermal stability. The flavone 5,7,2′,5′-tetrahydroxy-8,6′-dimethoxyflavone demonstrated dramatic processing effects, with fresh samples (FS:23.85%) and head-dried (HD:22.99%) containing 4–5 times more than freeze-dried (FD:5.29) or powdered samples, suggesting significant degradation during processing (Figure 5). These findings highlight that processing methods differentially affect various bioactive classes—while some compounds like violanthin and cyanidin are preserved or enhanced in powdered freeze-dried forms, others like methyl hesperidin and certain flavones are either heat-converted or degraded.

The metabolomic profiling of loquat flowers revealed significant variations in bioactive compounds depending on the drying method. Comparative analysis revealed FD significantly enhanced flavonoid content relative to both HD and FS samples (Figure 2), with 24 and 23 compounds showing >5-fold increases in FD vs. HD and FD vs. FS comparisons, respectively (Table 1). Freeze-drying markedly increased the abundance of key flavonoids, including delphinidin 3-O-beta-D-sambubioside (49.85-fold vs. HD), epitheaflavic acid-3′-gallate (23.35-fold vs. HD), gossypetin 3,4′-dimethyl ether (21.09-fold vs. HD), and eriodictyol chalcone (18.6-fold vs. FS), while heat-drying caused significant losses in glycosylated compounds (Table 1). Notably, cyanidin, an anthocyanin critical for color and health benefits, was 6.62-fold higher in FD than HD, confirming freeze-drying’s superiority in preserving thermolabile pigments (Table 1). Meanwhile, FD vs. FS comparisons indicated that freeze-drying not only preserves but may even enhance certain metabolites, such as pedaliin (18.01-fold increase) and eriodictyol chalcone (18.62-fold increase), likely due to the inhibition of enzymatic degradation during processing (Table 1). In contrast, heat drying led to the degradation of certain thermolabile compounds in comparison with FD, including methyl hesperidin (−3.31 Log2FC), eriodictyol (−2.34 Log2FC), taxifolin (−1.72 Log2FC) and scutellarin-7-diglucosidic acid (−1.81 Log2FC), suggesting that high temperatures may compromise flavonoid stability (Supplementary Table S3).

Interestingly, HD vs. FS results showed a mixed response (Figure 2): while some compounds like 6-hydroxyluteolin (27.36-fold increase) and eriodictyol (11.72-fold increase) were significantly elevated, others, including tricin-7-O-(6”-O-acetyl)glucoside and gossypetin 3,4′-dimethyl ether, were markedly reduced (Supplementary Table S3). This suggests that heat drying may induce both thermal degradation and potential thermal activation of certain metabolic pathways. Overall, freeze-drying appears to be the most effective method for preserving bioactive flavonoids, whereas heat drying may be suitable for extracting specific compounds but risks losing others.

The DPPH radical scavenging assay revealed significant differences in antioxidant capacity among processed loquat flower samples (Figure 6). FDP exhibited the highest antioxidant activity (608.83 μg TE/g), followed by HDP (552.89 μg TE/g) and fresh sample powder (FSP, 423.21 μg TE/g) (Figure 6). Among non-powdered samples, freeze-dried flowers (FD) showed superior antioxidant capacity (372.78 μg TE/g) compared to fresh (FS, 268.87 μg TE/g) and heat-dried (HD, 209.82 μg TE/g) samples (Figure 6). Notably, powdered extracts consistently outperformed their whole-flower counterparts, suggesting that hot-water extraction and subsequent freeze-drying enhance antioxidant bioavailability.

The data underscore that drying methods critically influence the phytochemical profile of loquat flower products. Freeze-drying appears optimal for retaining structurally complex flavonoids, making it suitable for high-value nutraceutical applications (Supplementary Table S4). In contrast, heat drying may be more appropriate for enhancing specific metabolites (e.g., 6-hydroxyluteolin, quercetagetin derivatives), which could be beneficial for antioxidant-rich ingredient formulations (Supplementary Table S4). However, the significant loss of anthocyanins like delphinidin 3-O-sambubioside in HDP suggests that thermal processing may compromise color-related bioactive properties (Supplementary Table S4). For manufacturers, selecting a drying method should depend on the target bioactive compounds—freeze-drying for maximal flavonoid preservation or heat drying for cost-effective production of select antioxidants. Freeze-drying powder extract is the gold standard for maximizing flavonoid diversity (e.g., cyanidin, 15.3-fold increase in FSP vs. FS), making it ideal for high-value nutraceuticals (Supplementary Table S5). Heat drying, while less universally protective, may be cost-effective for products targeting specific metabolites like 6-hydroxyluteolin or quercetagetin derivatives (Supplementary Table S5). The dramatic increases in FSP vs. FS suggest that even minimal processing (e.g., fresh samples powder extracting) can enhance certain bioactive, but long-term stability requires further study. For industry applications, hybrid approaches—such as mild heat drying followed by freeze-drying—could balance cost and nutrient retention.

The elevated antioxidant activity in freeze-dried powders (FDP) aligns with their higher retention of key flavonoids like violanthin and cyanidin, as identified in metabolomic analysis. While heat-dried powder (HDP) also demonstrated strong antioxidant capacity (likely due to thermally stable compounds like syringetin) its lower flavonoid diversity may limit broader health benefits. These findings suggest that freeze-dried loquat flower extracts are ideal for premium nutraceuticals requiring maximal antioxidant potency, whereas heat-dried powders offer a cost-effective alternative for products targeting specific bioactive metabolites.