Exploration on the inhibition effect and pathway of Perilla leaf ethanol extract on flavor deterioration of refrigerated snakehead fillets Authors: Jinlin Li, Yuyan Jin, Huimin Ma, Bin Peng, Bizheng Zhong, Zongcai Tu, Cheng-wei Yu Published: 2026 ABSTRACT This study investigated the effect of Perilla frutescens leaf ethanol extract (PLEE) on inhibiting flavor deterioration in refrigerated snakehead (Channa argus) fillets which integrated physicochemical, microbiological, and lipidomic analyses. Results identified 12 key flavor-differentiating compounds, 8 key differential microorganisms and 6 key differential lipids. Correlation analysis revealed that PLEE preserves quality through triple synergistic pathways: (1) Significantly inhibiting spoilage bacteria proliferation (Pseudomonas abundance reduced by 11.23%), with lowering total volatile basic nitrogen (TVB-N: 13.33 vs. 20.21 mg/100 g) and trimethylamine (TMA: 4.10 vs. 9.86 mg/kg) at day 12; (2) Protecting key lipids such as PG (16:1/18:3) and PA (22:5/23:1) to inhibit lipid oxidation and the generation of fishy odorants like 1-octen-3-ol; (3) Enhancing sensory acceptance via olfactory contribution of its characteristic terpenoids, including perillaldehyde. These findings could elucidate the multi-target preservative effect of PLEE and support its potential application as a natural preservative in aquatic products. Introduction The prepared food industry in China has experienced substantial growth in recent years, with its market value projected to reach $150.29 billion by 2026. Aquatic-based prepared foods currently dominate this sector, constituting a significant market share. Among raw materials for prefabricated freshwater fish fillets, snakehead (Channa argus) is distinguished by its low bone content that facilitates processing efficiency, alongside superior nutritional properties and substantial economic value in aquaculture (Rahman et al., 2025). It is particularly notable for its high protein content and abundant polyunsaturated fatty acids, making it a valuable resource in the food industry (Deng et al., 2024). However, the high protein content accelerates postmortem muscle tissue deterioration in snakehead, leading to rapid quality loss due to microbial spoilage and lipid oxidation, generating unacceptable off-odors (fishy and putrid notes) within several hours (Sun et al., 2022). The rapid deterioration in both quality and flavor of snakehead fillets will cause the great economic loss and restriction of trans-regional market. Therefore, investigating the related refrigeration preservation methods is of vital importance. Preservation techniques for aquatic products are conventionally classified into physical, chemical, and biological methods (Chang et al., 2025; Francisco et al., 2024). In response to escalating consumer demand for natural and minimally processed foods, the adoption of eco-friendly preservatives has increased significantly (Ghamry et al., 2025). Perilla ethanol extracts contain significantly higher polyphenol and flavonoid concentrations than aqueous extracts due to solvent polarity differences, thereby exhibiting superior antioxidant and other functional efficacies. (Plaskova & Mlcek, 2023). Perilla leaf ethanol extracts (PLEE), enriched with polyphenolics (rosmarinic acid) and flavonoids (apigenin, luteolin), exhibit dual antioxidative and antimicrobial activities, rendering them effective in aquatic product preservation (Dang et al., 2024; Jin et al., 2023). Despite these benefits, the pathway basis for ability of perilla extract to retard flavor deterioration in refrigerated snakehead fillets remains uncharacterized. Resolving this knowledge gap would substantially contribute to the scientific framework for flavor stability control in perishable aquatic products. Postmortem flavor deterioration in fish muscle arises from lipid oxidation, enzymatic degradation, and microbial spoilage (Peng et al., 2025). Notably, enzymatic processes accelerate flavor degradation by promoting the proliferation and metabolic activity of spoilage microorganisms. Consequently, lipid oxidation and microbial metabolism constitute the primary drivers of postmortem flavor deterioration in fish (Dai et al., 2024; Zhou et al., 2024). Current evidence attributes the flavor-stabilizing effects of Perilla frutescens to its dual antioxidative and antimicrobial properties. For instance, Meng et al. (2008) identified polyphenol-rich compounds, including rosmarinic acid, apigenin 7-O-diglucuronide, and luteolin 7-O-diglucuronide, in dehydrated perilla aqueous extracts, which exhibit potent antioxidant activity. Kang reported that perillaldehyde from perilla leaves moderately inhibits diverse bacteria and fungi, while its synergy with polydatin extends efficacy against both Gram-positive and Gram-negative bacteria (Kang et al., 2024). Similarly, Meng isolated antioxidative constituents (rosmarinic acid, luteolin, apigenin and chrysoeriol) from perilla, all exhibiting robust radical-scavenging capacity (Meng et al., 2008). Given the intricate linkage between flavor deterioration in aquatic products and concurrent lipid oxidation/microbial proliferation, and the inherent complexity of these mechanisms, integrated multi-omics approaches have become essential. For instance, Chen et al. (2024) combined flavoromics and metabolomics to elucidate dynamic flavor changes during Pacific oyster purification. Further, Li et al. (2025) integrated lipidomics with flavoromics to reveal lipid-flavor transformation mechanisms in silver carp visceral oil under enzymatic hydrolysis. Peng et al. (2025) employed metabolomics-microbiomics coupling to decipher how low-salt dry-curing inhibits volatile flavor deterioration in refrigerated grass carp. Collectively, these studies underscore the critical role of multi-omics integration in unraveling complex flavor degradation pathways. Current research regarding the preservation effect of perilla leaf still mainly remains at a phenomenological level, lacking systematic investigation into underlying molecular mechanisms. Therefore, using multi-omics techniques to investigate the inhibitory effects of PLEE on the flavor deterioration of refrigerated snakehead fillets is feasible and has significant guidance. In this study, refrigerated snakehead fillets were employed as a research object. PLEE immersion treatment was implemented to integrate flavoromics with sensory-directed analysis for comprehensive flavor profiling. Concurrently, microbiomics and lipidomics analyses were synergistically applied to elucidate dynamic interrelationships among flavor compounds, spoilage-related microbial taxa, and lipid oxidation derivatives during refrigeration. Finally, the flavor compounds, microbial communities and lipid analyzed by correlation networks to explore the inhibition pathways about PLEE on flavor deterioration. The aim of this research was to elucidate the preservation pathway of PLEE in refrigerated snakehead fillets using an integrated multi-omics approach, thereby providing effective guidance for enhancing flavor preservation in the prepared fish fillet industry. Materials and methods Materials and sample preparation Fresh perilla leaves were obtained from Jiangxi Sunshine Agricultural Technology Co. LTD (Jiujiang, China). The leaves were first dried and then ground into fine powder using a mechanical grinder (JJ-699, Yongkang Hongtaiyang Electromechanical Co., LTD, Jinhua, China). For extraction, the powdered material was mixed with 90% ethanol solution at a ratio of 1:10 (w/v) and subjected to ultrasonic-assisted extraction (55 °C, 70 W) for 25 min. After cooling to room temperature, the mixture was vacuum-filtered, and the residual marc was re-extracted under identical conditions. The combined filtrates were concentrated using a rotary evaporator at 45 °C. The obtained ethanol extract should be stored at −20 °C until it is ready for use. Before use, it needs to be re-dissolved in distilled water to a concentration of 56 mg/mL. The concentration of PLEE (56 mg/mL) was selected based on preliminary optimization experiments, which identified this as the minimum effective concentration to significantly inhibit the increase of TVC and TVB-N in refrigerated snakehead fillets, while also considering cost-effectiveness for potential industrial application. Fresh snakehead (Channa argus, 0.80–1.00 kg per individual) were purchased from a local supermarket (Baijiahui Supermarket, Nanchang, China), which were maintained in a tank with clean water during transportation to the laboratory. Upon arrival, the live snakeheads were immediately subjected to electro-anesthetized euthanasia using AN50703S DC stabilized power source (Qingdao Ainuo Instruments Co., Ltd.) in accordance with the Guidelines for Experimental Animal Treatment Issued by the Ministry of Science and Technology of China (2006) and Ethical Statement guide lines for animal handling (No.20250606–012). For electro-anesthesia, the fish was immersed in a polyethylene plastic tank measuring 2.0 m in length, 1.0 m in width, and 1.5 m in depth, filled with freshwater containing 2 g NaCl/L to a depth of 50 cm. Two plate copper electrodes (40.0 × 2.5 × 1.2 cm) were vertically placed near both ends of the tank at a distance of 1.6 m apart, and direct current was applied using a DC stabilized power source (3 A/200 V for 1 min, followed by 3.5 A/220 V for 2 min). Subsequently, the snakeheads were immediately decapitated. Fresh snakehead fillets were prepared on a refrigerated workbench, with all tools and surfaces cleaned with sterile water. The preparation process included: (1) removal of scales, viscera, tail, and bones; (2) rinsing with running distilled water to eliminate surface impurities and blood residues; and (3) cutting into standardized portions (3 × 3 × 1 cm), each weighing 9.00 ± 1.00 g. The fillets were randomly divided into three experimental groups: Untreated control (UFF) group: no additional treatment; Sterile distilled water-treated (SDW) group: immersed in sterile distilled water for 30 min; Perilla ethanol extract-treated (PEE) group: immersed in reconstituted PLEE for 30 min, with 0.83 mL of PLEE corresponding to each gram of fish meat. Following treatment, excess moisture was drained and samples were vacuum-sealed in polyethylene bags using water displacement method to minimize air exposure. All samples were stored at 4 °C. Sampling was performed at 0, 3, 6, 9, and 12 days of refrigerated storage, with triplicate samples analyzed at each time point. All procedures complied with national animal welfare regulations and institutional ethical guidelines for experimental animal use. Detection of physicochemical indicators Color Color analysis of the samples was performed according to the standards established by the International Commission on Illumination (CIE, 2019). Briefly, the colorimeter (WSC-2B, Shanghai INESA Physico-Optical Instrument Co.,Ltd. Shanghai, China) was calibrated using the built-in white reference tile. Samples were placed on a white background, and the colorimeter was employed to determine the brightness (L* value), redness (a* value), and yellowness (b* value) of the sample cross-sections. Each sample was measured nine times to ensure data reproducibility. Texture profile A texture analyzer equipped with a P/36R probe (TA-XT Plus, Stable Micro Systems LTD, Godalming, UK) was used to determine the texture parameters of snakehead fillets, including hardness, springiness, cohesiveness, chewiness, and resilience (Zhang et al., 2024). The pretest speed, test speed and posttest speed were set at 1 mm/s, 1 mm/s and 1 mm/s, respectively. The interval time was 5 s, and the trigger force was 5 g. The compression ratio was set at 40%. Centrifugal loss rate Approximately 5 g of intact fish fillet (W1) was wrapped in double-layer filter paper, the sample was then centrifuged (H18.5B, Shanghai Lu Xiangyi Centrifuge Instrument Co., Ltd., Shanghai, China) at 13000 ×g for 10 min. Subsequently, the mass of the centrifuged fish fillet was accurately weighed (W2) (Lu et al., 2017). The centrifugal loss was calculated according to the following equation (Zhang et al., 2024): CLR (%) = (W1-W2) / W2 × 100%. In the above formula, W1 represents the weight before centrifuge of snakehead fillets, and W2 is the weight after centrifuge of snakehead fillets. pH The minced fish meat (4 g) was suspended in 40 mL of sterile distilled water and homogenized for 3 min (Zhang et al., 2024). The pH of the filtrate was then measured with a pH meter (Five Easy Plus, Mettler Toledo International Trade (Shanghai) Co., Ltd., Shanghai, China). Total volatile basic nitrogen (TVB-N) Approximately 1.50 g of homogenized fish muscle (accurately weighed to ±0.0001 g using an analytical balance) was mixed with 25 mL of ultrapure water in a sterile centrifuge tube. The mixture was vortexed for 30 s to ensure complete homogenization, followed by stationary extraction for 30 min at room temperature (25 ± 1 °C). After phase separation, exactly 20 mL of the supernatant was transferred to a digestion tube using a calibrated pipette (Wang et al., 2024). Subsequently, 1.00 ± 0.01 g of magnesium oxide (MgO, analytical grade) was added as an alkalizing agent. And preliminary tests have confirmed that 1.00 g of MgO is sufficient to completely alkalize 1.50 g of the fish sample without causing excessive alkalization. The TVB-N content was determined through automated Kjeldahl analysis (K9860, Hanon Group, Shandong, China) Calculation based on the following formula:TVB−N=V1−V0×c×14m×100 In the above formula, V1 represents the sample titration volume, V0 represents the blank titration volume, c represents the HCl concentration, and m represents the sample weight. Trimethylamine (TMA) TMA was analyzed by headspace gas chromatography (HS-GC) with modifications based on Chinese National Food Safety Standard Method (GB 5009.179–2016). The extraction solution was prepared following the GB. Approximately 10 g of prepared sample (±0.001 g) was weighed into a 50 mL polypropylene centrifuge tube. Then 20 mL of 5% trichloroacetic acid solution was added and homogenized for 1 min, followed by centrifugation at 1800 ×g for 5 min. The supernatant was filtered into a 50 mL volumetric flask through a glass funnel layered with defatted cotton wool. The residue was re-extracted twice with sequential additions of 15 mL and 10 mL of 5% trichloroacetic acid solution. Finally, all filtrates were pooled and brought to 50 mL volume with 5% trichloroacetic acid solution. Precisely 2.0 mL of the extract was transferred to a 20 mL headspace vial, which was immediately crimp-sealed. Using a medical plastic syringe, exactly 5.0 mL of 50% NaOH was injected into the vial. The mixture was then balanced in a water bath at 55 °C for 15 min. Subsequently, a solid-phase microextraction (SPME) fiber was inserted into the headspace vial for 20 min to adsorb TMA. The adsorbed analytes were desorbed at 250 °C and analyzed by GC–MS (7890 A/5975, Agilent Technologies, Santa Clara, USA) equipped with a DB-Wax capillary column (30 m × 0.25 mm × 0.25 μm). High-purity nitrogen was used as carrier gas at a constant flow rate of 2.5 mL/min. The injector temperature was maintained at 220 °C. The oven temperature program was as follows: initial hold at 40 °C for 3 min, followed by a temperature increase to 220 °C at a rate of 30 °C/min, with a final hold at 220 °C for 1 min. The flame ionization detector temperature was set at 220 °C. TMA quantification was performed using a standard curve generated from TMA hydrochloride, with results expressed as mg TMA per kg meat. The limit of detection (LOD) and limit of quantification (LOQ) of the method is 1.5 mg/kg and 5.0 mg/kg, respectively, based on a signal-to-noise ratio of 3:1 and 10:1. Free amino acid (FAA) Approximately 5 g of homogenized sample was accurately weighed and extracted with 50 mL of 0.1 mol/L HCL for 30 min. After thorough mixing, the solution was then filtered by a 0.22 μm filter, and precisely 2 mL of the clear filtrate was transferred to a centrifuge tube. Then, 2 mL of 8% sulfosalicylic acid was added, followed by vortex mixing and incubation for 15 min at room temperature. The mixture was subsequently centrifuged at 13000 ×g for 10 min using a high-speed centrifuge (Model H18.5B, Shanghai Lu Xiangyi Centrifuge Instrument Co., Ltd., China). The supernatant was carefully collected and filtered through a 0.22 μm aqueous-phase membrane filter prior for subsequent analysis (Wang et al., 2024). Amino acid composition was determined using an automated amino acid analyzer (L-8900, Hitachi, Tokyo, Japan) with dual-wavelength detection at 570 nm and 440 nm. Flavor assays Volatile flavor compounds (VFCs) Add approximately 5 g of ground snakehead fish fillets or 2 mL of PLEE to a 20 mL headspace vial, and then add 2 mL of 18% (w/v) sodium chloride solution, immediately following each sample, 10 mL of 1.306 μg/mL 1,2-dichlorobenzene internal standard was added. The mixture was thermally equilibrated in a water bath for 30 min at 55 °C (Xu et al., 2025). Subsequently, the SPME fibers were exposed to a headspace bottle that kept the water bath heated for 30 min to adsorb VFCs. The adsorbed analytes were thermally desorbed at 250 °C in the injection port of a gas chromatography–mass spectrometry (GC–MS) system for subsequent separation and identification (7890 A/5975, Agilent Technologies, Santa Clara, USA). The gas chromatography system was equipped with a DB-Wax capillary column (30 m × 0.25 mm × 0.25 μm; Agilent Technologies, Santa Clara, USA). Operational parameters included an inlet pressure of 7.361 psi and helium carrier gas (purity >99.99%) flowing at 1.0 mL/min. A split outlet purge flow of 100 mL/min was applied for 1 min. The injector and detector temperatures were maintained at 250 °C and 230 °C, respectively. The temperature program was as follows: Hold at 40 °C for 3 min, followed by a linear rise to 240 °C at 5 °C /min and a final hold at 240 °C for 15 min. The ionization voltage of the mass spectrometer was 70 eV, and the ion source and quadrupole temperatures were set at 230 °C and 150 °C, respectively. The mass spectrum was scanned from 35 to 500 m/z. Qualitative and quantitative analysis: Volatile compounds with round-trip matching greater than 800 were retrieved by computer and qualitatively compared with NIST 14 standard mass spectrometry combined with retention index (RI). The peak area ratio of each analyte to the internal standard was used to calculate its concentrations. OAV is used to evaluate the contribution of odor substances to the odor of a sample, and it is generally considered that substances with OAV ≥ 1 contribute more to the flavor of the sample. When 0 ≤ OAV < 1, the substance was used as a flavor ornament for the overall odor (Dai et al., 2024). The larger the OAV, the greater the contribution to the overall flavor. The calculation result is shown in the formula:OAV=CiTi In the above formula, Ci represents the actual concentration of each volatile compound in the sample, and Ti represents the odor threshold of the compound in an aqueous solution system. Sensory evaluation The sensory evaluation was conducted with approval from the ethics committee. A panel of ten trained assessors independently evaluated the snakehead fish fillets under controlled conditions, with no communication permitted between panelists during the assessment. The olfactory profile was characterized using six specific descriptors: fishy odor, earthy odor, grassy odor, fresh fish aroma, herbal odor, and citrus-like odor (Wang et al., 2024). Each assessor quantified odor intensity using a 10-point scale, where 0 represented undetectable odor and 10 indicated extremely intense odor. Ethical approval for sensory evaluation involving human participants was obtained from the Experimental Ethics and Welfare Committee of the School of Life Sciences, Jiangxi Normal University (Approval No. 20250512–001). The study adhered strictly to pertinent laws and regulations. Written informed consent was obtained from all participants prior to the study. Data were anonymized and participant privacy was strictly protected. Participation was voluntary, with the right to withdraw at any time without prejudice. Microbial diversity test Total viable counts (TVC) Total viable count was based on Zhang et al. (2024) with slight modifications. Approximately 5 g of the crushed sample was aseptically transferred into a 20 × 20 cm sterile homogeneous bag at a ratio of 1:9 (m/v). The mixture was homogenized for 6 min. Samples were homogenized and serially diluted 10 fold. Diluted samples were added to plates to determine the total viable count. The plates were then incubated at 37 °C for 3 days. TVC was expressed as log colony-forming units per gram (lg CFU/g). Microbiota diversity Microbiological analysis of chilled snakehead fillet was performed according to the protocol of (Peng et al., 2025). Genomic DNA was extracted and examined by electrophoresis on a 1% agarose gel (Peng et al., 2025). Primers 338F (5’-ACTCCTACGGGAGGCAGCA-3′) and 806R (5 ‘-GGACTACHVGGGTWTCTAAT-3′) were used to amplify the V3-V4 region of the bacterial 16S rRNA gene. The composition of the PCR reaction system is as follows: 4 μL of 5 × TransStart FastPfu buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each upstream and downstream primer (5 μM), 0.4 μL of TransStart FastPfu DNA polymerase, 10 ng of template DNA, and ddH₂O is added to make up to 20 μL. The reaction procedure is as follows: pre-denaturation at 95 °C for 3 min; 27 cycles including 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s; final extension at 72 °C for 10 min; stored at 4 °C (PCR instrument model: ABI GeneAmp® 9700). Polymerized amplification products were detected by electrophoresis on 2% agarose gels, then was recovered using the PCR Clean-Up Kit (YuHua, Shanghai, China) and eluted with Tris-HCl. Subsequent quantification was performed with a QuantiFluor™-ST blue fluorescence system (Promega) followed by proportional mixing according to sequencing requirements. Illumina adapter sequences were ligated to the outer end of the target region by PCR. The gel-purified amplification products were denatured with sodium hydroxide to produce single-stranded DNA, which was amplified by PCR to form DNA clusters. Sequencing utilizes fluorescently labeled dntp and modified polymerase with laser scanning after each nucleotide incorporation cycle. Purified amplicons were pooled in equimolar amounts and paired-end sequenced on an Illumina Nextseq2000 platform (Illumina, San Diego, USA) according to the standard protocols by Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China). Lipidomics test The lipid extraction was carried out as described by Zhao et al. (2025) with a slight modification. Lipid extraction was performed by weighing 50 mg solid samples into 2 mL microtubes and adding 280 μL methanol: water (2:5, v/v) and 400 μL MTBE. After adding a 6-mm grinding bead, samples were homogenized at −10 °C (50 Hz, 6 min) using a Wonbio-96c tissue crusher (Shanghai Wonbio Technology Co., China), sonicated at 5 °C (40 kHz, 30 min), and incubated at −20 °C for 30 min. Following centrifugation (13,000 ×g, 4 °C, 15 min), 350 μL of the upper-phase lipid extract was collected, evaporated under nitrogen stream, and reconstituted in 100 μL isopropanol: acetonitrile (1:1, v/v) with brief sonication at 5 °C. The solution was centrifuged again (13,000 ×g, 4 °C, 15 min) prior to UHPLC-MS/MS injection (2 μL). Chromatographic separation employed an Accucore™ C30 column (100 × 2.1 mm, 2.6 μm; Thermo) on a Thermo UHPLC-Q Exactive HF-X system, with mobile phases comprising: solvent A (10 mmol/L ammonium acetate in acetonitrile/water (1:1, v/v) with 0.1% formic acid) and solvent B (2 mmol/L ammonium acetate in acetonitrile/isopropanol/water (10:88:2, v/v/v) with 0.02% formic acid). Separation was achieved at 0.4 mL/min and 40 °C for 20 min, with samples maintained at 4 °C during analysis. Mass spectrometric analysis utilized a Thermo Q-Exactive HF-X Orbitrap system with H-ESI source operating in dual-polarity mode. Standard parameters included: sheath gas flow (60 psi), aux gas flow (20 psi), ion source temperature (370 °C), ion-spray voltage (± 3000 V alternating by polarity), and normalized collision energy (20–40-60 V). Data-dependent acquisition (DDA) covered m/z 200–2000. Raw data were processed in LipidSearch™ (Thermo) for peak alignment (precursor/fragment mass tolerance: 10 ppm) and lipid identification (m-score ≤ 2.0; ID grades A-D). Post-filtering, features detected in ≥80% of samples were retained through the Majorbio Cloud Platform (cloud.majorbio.com). Data underwent sum normalization, low-abundance metabolite imputation, RSD-based filtering (QC RSD >30% removed), and log10 transformation prior to statistical analysis. Statistical analysis All data were expressed as mean ± standard deviation (SD) from triplicate determinations. Statistical analysis was performed using SPSS 22.0 (IBM, USA), with one-way ANOVA followed by Tukey's post hoc test (P < 0.05 indicating significance). Cytoscape correlation analysis was conducted to assess relationships between key differential flavor compounds, microorganisms and lipids. Graphical representations were generated using Origin 2019. Results and discussion Changes in physicochemical properties Color, texture, centrifugal loss rate and pH L* (brightness) and W (whiteness) values declined consistently across samples during storage due to lipid oxidation/enzymatic browning, with PLEE significantly mitigating L loss (P < 0.05) (Table 1). Concurrent with chromatic degradation, irreversible textural deterioration (decreased hardness, springiness, cohesiveness, chewiness; Table 2) occurred via endogenous protease-mediated myofibrillar degradation, actomyosin dissociation-induced moisture migration, and collagen destabilization (Wang et al., 2022); despite superior initial hardness of UFF, PEE significantly exceeded UFF in late storage (from day 9 to day 12) by inhibiting microbial protease secretion and free radical protein cross-linking to mitigate structural collapse. Water-holding capacity (WHC), quantified via centrifugal loss rate (Fig. 1A), critically determines textural stability. All groups exhibited increased centrifugal loss rates with diminished increments over time: rapid escalation (between day 0 and day 3) followed by plateaued progression. By day 6, divergent trends emerged (UFF > SDW > PEE, P < 0.05). Integrated with texture analysis, PLEE decelerated WHC decline by retarding muscle tissue denaturation (Kucukoglu et al., 2024). Microenvironmental pH dynamics (Fig. 1B) reflected collective biological metabolisms: firstly, all groups showed an initial decline (glycolysis-dominated), then followed by a mid-term rise (microbial deamination-dominated) and finally showed a late-stage gradual reduction (acidic metabolite accumulation). Despite statistically significant differences in the PEE from day 3 to day 9 (P < 0.05), its limited overall modulation amplitude implies lower pH sensitivity to PLEE to other physicochemical indicators.Table 1Changes of color of snakehead in different groups during 4 °C cold storage.IndexGroupsStorage time0 d3 d6 d9 d12 dL*UFF60.94 ± 1.11Ba59.51 ± 0.25ABa58.92 ± 0.79ABa57.95 ± 1.31Aa57.55 ± 1.19AaSDW61.8 ± 1.76Ba59.41 ± 0.8ABa60.24 ± 0.93ABa59.28 ± 0.52ABa58.34 ± 1.15AaPEE59.99 ± 1.1Aa58.84 ± 1.51Aa58.62 ± 1.02Aa59.13 ± 0.43Aa58.28 ± 1.29Aaa*UFF2.15 ± 1.87Aa5.84 ± 0.45Ba4.2 ± 0.47ABb4.57 ± 0.95ABa3.04 ± 0.18AaSDW0.93 ± 2.5Aa2.54 ± 0.47Ab2.21 ± 1.13Aab4.37 ± 1.03Aa2.26 ± 0.37AaPEE−1.08 ± 0.72Aa1.64 ± 0.35Ba0.91 ± 1.11ABa2.73 ± 0.95Ba1.9 ± 1.09Bab*UFF3.65 ± 1.34Aa5.52 ± 1.18Ab5.38 ± 1.25Aab5.23 ± 1.17Aa5.05 ± 0.71AaSDW2.17 ± 1.08Aa2.87 ± 0.74Aa3.68 ± 1.69Aa3.96 ± 0.62Aa2.78 ± 1.78AaPEE8.59 ± 1.57Ab8.41 ± 0.83Ac8.28 ± 1.44Ab10.53 ± 0.86Ab8.63 ± 1.34AbWUFF60.66 ± 0.91Bab58.71 ± 0.44Aa58.34 ± 0.60Aa57.35 ± 1.06Aa57.20 ± 1.14AaSDW61.65 ± 1.58Bb59.23 ± 0.78ABa59.98 ± 0.81ABb58.84 ± 0.44Aa58.15 ± 1.01AaPEE59.04 ± 0.78Aa57.95 ± 1.31Aa57.76 ± 0.82Aa57.69 ± 0.59Aa57.33 ± 1.14AaNote: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05).Table 2Changes in texture characteristics of snakehead fillets in different groups during 4 °C cold storage.IndexGroupsStorage time (d)036912Hardness(g)UFF586.73 ± 56.5Ca415.62 ± 15.27Ba412.51 ± 2.29Ba378.64 ± 12.17ABb250.17 ± 15.03AaSDW471.65 ± 45.35Ca453.12 ± 61.3Ca428.2 ± 43.89Ca335.35 ± 9.35Ba256.76 ± 18.5AaPEE487.72 ± 57.61Ca463.18 ± 33.18Ca402.22 ± 11.25Ba400.6 ± 8.73Bb277.49 ± 27.17AaSpringinessUFF0.69 ± 0.08Ba0.62 ± 0.08ABa0.53 ± 0.07Aa0.51 ± 0.04Aa0.5 ± 0.04AaSDW0.78 ± 0.1Ba0.6 ± 0.09Aa0.5 ± 0.01Aa0.53 ± 0.05Aa0.52 ± 0.05AaPEE0.69 ± 0.04Ca0.64 ± 0.05BCa0.56 ± 0.07ABa0.61 ± 0.02BCa0.51 ± 0.07AaCohesivenessUFF0.64 ± 0.05Ba0.58 ± 0.09ABa0.54 ± 0.03ABa0.55 ± 0.03ABa0.49 ± 0.06AaSDW0.65 ± 0.06Ba0.57 ± 0.06Aa0.54 ± 0.01Aa0.54 ± 0.03Aa0.55 ± 0.02AaPEE0.58 ± 0.05Aa0.54 ± 0.05Aa0.59 ± 0.07Aa0.61 ± 0.06Aa0.54 ± 0.04AaChewinessUFF253.88 ± 48.32Ca150.86 ± 35.73Ba115.96 ± 11.92ABa105.81 ± 14.12ABa61.56 ± 12.58AaSDW247.64 ± 106.03Ba142.47 ± 51.08ABa115.41 ± 66.21Aa96.15 ± 14.79Aa73.2 ± 13.53AaPEE199.73 ± 63.47Ba161.68 ± 15.24Ba133.94 ± 34.29ABa147.73 ± 19.1Bb73.24 ± 9.31AaResilienceUFF0.39 ± 0.02Ca0.3 ± 0.07Ba0.26 ± 0.01Ba0.24 ± 0.02ABa0.18 ± 0.01AaSDW0.4 ± 0.07Ba0.28 ± 0.05Aa0.24 ± 0.04Aa0.23 ± 0.03Aa0.21 ± 0.02AaPEE0.34 ± 0.04Ba0.33 ± 0.05ABa0.32 ± 0.07ABa0.29 ± 0.06ABa0.2 ± 0.02AaNote: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05).Fig. 1Changes in centrifugal loss rate (A), pH (B), TVB-N (C), TMA (D), TVC (E) and sensory score (F) of fillets during 4°C cold storage.Fig. 1 Table 1Changes of color of snakehead in different groups during 4 °C cold storage.IndexGroupsStorage time0 d3 d6 d9 d12 dL*UFF60.94 ± 1.11Ba59.51 ± 0.25ABa58.92 ± 0.79ABa57.95 ± 1.31Aa57.55 ± 1.19AaSDW61.8 ± 1.76Ba59.41 ± 0.8ABa60.24 ± 0.93ABa59.28 ± 0.52ABa58.34 ± 1.15AaPEE59.99 ± 1.1Aa58.84 ± 1.51Aa58.62 ± 1.02Aa59.13 ± 0.43Aa58.28 ± 1.29Aaa*UFF2.15 ± 1.87Aa5.84 ± 0.45Ba4.2 ± 0.47ABb4.57 ± 0.95ABa3.04 ± 0.18AaSDW0.93 ± 2.5Aa2.54 ± 0.47Ab2.21 ± 1.13Aab4.37 ± 1.03Aa2.26 ± 0.37AaPEE−1.08 ± 0.72Aa1.64 ± 0.35Ba0.91 ± 1.11ABa2.73 ± 0.95Ba1.9 ± 1.09Bab*UFF3.65 ± 1.34Aa5.52 ± 1.18Ab5.38 ± 1.25Aab5.23 ± 1.17Aa5.05 ± 0.71AaSDW2.17 ± 1.08Aa2.87 ± 0.74Aa3.68 ± 1.69Aa3.96 ± 0.62Aa2.78 ± 1.78AaPEE8.59 ± 1.57Ab8.41 ± 0.83Ac8.28 ± 1.44Ab10.53 ± 0.86Ab8.63 ± 1.34AbWUFF60.66 ± 0.91Bab58.71 ± 0.44Aa58.34 ± 0.60Aa57.35 ± 1.06Aa57.20 ± 1.14AaSDW61.65 ± 1.58Bb59.23 ± 0.78ABa59.98 ± 0.81ABb58.84 ± 0.44Aa58.15 ± 1.01AaPEE59.04 ± 0.78Aa57.95 ± 1.31Aa57.76 ± 0.82Aa57.69 ± 0.59Aa57.33 ± 1.14AaNote: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05). Changes of color of snakehead in different groups during 4 °C cold storage. Note: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05). Table 2Changes in texture characteristics of snakehead fillets in different groups during 4 °C cold storage.IndexGroupsStorage time (d)036912Hardness(g)UFF586.73 ± 56.5Ca415.62 ± 15.27Ba412.51 ± 2.29Ba378.64 ± 12.17ABb250.17 ± 15.03AaSDW471.65 ± 45.35Ca453.12 ± 61.3Ca428.2 ± 43.89Ca335.35 ± 9.35Ba256.76 ± 18.5AaPEE487.72 ± 57.61Ca463.18 ± 33.18Ca402.22 ± 11.25Ba400.6 ± 8.73Bb277.49 ± 27.17AaSpringinessUFF0.69 ± 0.08Ba0.62 ± 0.08ABa0.53 ± 0.07Aa0.51 ± 0.04Aa0.5 ± 0.04AaSDW0.78 ± 0.1Ba0.6 ± 0.09Aa0.5 ± 0.01Aa0.53 ± 0.05Aa0.52 ± 0.05AaPEE0.69 ± 0.04Ca0.64 ± 0.05BCa0.56 ± 0.07ABa0.61 ± 0.02BCa0.51 ± 0.07AaCohesivenessUFF0.64 ± 0.05Ba0.58 ± 0.09ABa0.54 ± 0.03ABa0.55 ± 0.03ABa0.49 ± 0.06AaSDW0.65 ± 0.06Ba0.57 ± 0.06Aa0.54 ± 0.01Aa0.54 ± 0.03Aa0.55 ± 0.02AaPEE0.58 ± 0.05Aa0.54 ± 0.05Aa0.59 ± 0.07Aa0.61 ± 0.06Aa0.54 ± 0.04AaChewinessUFF253.88 ± 48.32Ca150.86 ± 35.73Ba115.96 ± 11.92ABa105.81 ± 14.12ABa61.56 ± 12.58AaSDW247.64 ± 106.03Ba142.47 ± 51.08ABa115.41 ± 66.21Aa96.15 ± 14.79Aa73.2 ± 13.53AaPEE199.73 ± 63.47Ba161.68 ± 15.24Ba133.94 ± 34.29ABa147.73 ± 19.1Bb73.24 ± 9.31AaResilienceUFF0.39 ± 0.02Ca0.3 ± 0.07Ba0.26 ± 0.01Ba0.24 ± 0.02ABa0.18 ± 0.01AaSDW0.4 ± 0.07Ba0.28 ± 0.05Aa0.24 ± 0.04Aa0.23 ± 0.03Aa0.21 ± 0.02AaPEE0.34 ± 0.04Ba0.33 ± 0.05ABa0.32 ± 0.07ABa0.29 ± 0.06ABa0.2 ± 0.02AaNote: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05). Changes in texture characteristics of snakehead fillets in different groups during 4 °C cold storage. Note: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05). Changes in centrifugal loss rate (A), pH (B), TVB-N (C), TMA (D), TVC (E) and sensory score (F) of fillets during 4°C cold storage. TVB-N value Total volatile basic nitrogen (TVB-N) serves as a critical indicator for assessing the freshness quality of aquatic products. As depicted in Fig. 1C, the initial TVB-N values across all three groups were below 10 mg/100 g with no statistically significant differences. By day 3 of storage, TVB-N values exhibited significant increases in all groups, a phenomenon linked to rapid microbial proliferation during early refrigeration stages and accelerated protein degradation by extracellular proteases (Du et al., 2025). Especially, the UFF group displayed significantly higher TVB-N values than both SDW and PEE groups at this stage, while no significant difference existed between SDW and PEE groups. This divergence may be attributed to the UFF group's lack of water-immersion pretreatment, resulting in higher surface microbial loads (Erkan et al., 2008). On day 6, although TVB-N levels continued to rise across all, the increments in SDW and PEE groups did not reach statistical significance. From the 9th day to 12th day, significant increases were observed for all groups, approximately triple their initial levels. The inhibition of TVB-N accumulation in the PEE group can be ascribed to the synergistic activity of compounds in PLEE: these agents disrupt microbial cell membrane structures to inhibit growth and metabolism while binding to hydrophobic protein domains to delay denaturation. These findings confirm that PLEE effectively impedes TVB-N accumulation in snakehead fillets, thereby preserving freshness during refrigerated storage (He et al., 2024). TMA Trimethylamine (TMA) directly induces fishy odors and signals spoilage onset as a key marker of quality deterioration, primarily generated by microbial reduction of TMAO via spoilage bacteria (Shewanella/Pseudomonas) (Chun et al., 2014; Zhou et al., 2024). In this study, TMA levels in all groups exhibited a triphasic trajectory: firstly a rapid increase, then a transient decline, and finally a sustained elevation (Fig. 1D). Remarkably, significant intergroup differences emerged during initial refrigeration (from day 0 to day 3), potentially attributable to rapid postmortem microbial proliferation and early antibacterial activity of PLEE in the PEE. The initial TMA surge reflects active metabolic processes of spoilage bacteria, while the subsequent temporary decrease may indicate ecological succession within microbial communities: substrate depletion and acidic metabolite accumulation inhibit certain TMA-producing bacteria, with concomitant TMA bioconversion by secondary microbiota (Wang et al., 2022). By storage endpoint (from day 9 to day 12), TMA content in UFF significantly surpassed that of PEE, with net increases measuring UFF (9.86 mg/kg), SDW (6.71 mg/kg), and PEE (4.10 mg/kg). These data confirmed efficacy of PLEE in inhibiting TMA accumulation (58.4% reduction compared with UFF), likely mediated through multi-target pathway: PLEE components disrupt microbial membrane structures to inhibit TMAO reductase activity, while simultaneously modulating microbiota composition to reduce TMA-producing bacteria abundance. Furthermore, antioxidant constituents in PLEE may interrupt lipid peroxidation chain reactions, limiting TMA precursor availability (Cai et al., 2024). Collectively, PLEE delays TMA-mediated flavor deterioration through coordinated regulation of microbial metabolism and oxidative stability. Free amino acids Free amino acids play a multifunctional role in fish, serving as key indicators for freshness, umami, and nutritional value, as well as precursors for flavor compounds like trimethylamine and aldehydes through decarboxylation and Strecker degradation. Strecker aldehydes typically exhibit potent aromas that significantly influence food flavor profiles. The dynamic changes in FAAs directly reflect the biochemical status of refrigerated fish fillets. As shown in Table 3, three snakehead fillet groups demonstrated distinct FAA evolution patterns: both UFF and SDW showed an initial decrease followed by an increase, whereas the PEE displayed fluctuating “increase-decrease-increase” dynamics. This phenomenon originates from the interplay between proteolysis and microbial metabolism, FAA accumulation primarily depends on endogenous proteases and microbial exoenzymes hydrolyzing peptide bonds, while their reduction stems from microbial FAA consumption. PLEE-treated samples exhibited significantly lower total FAAs than control groups at most timepoints, attributable to three synergistic pathway: (1) hydrophobic interactions and hydrogen bonding reducing free FAA concentrations; (2) inhibition of cathepsin activity decelerating proteolysis; (3) compromised microbial membrane integrity diminishing protease secretion. However, by storage endpoint, the PEE group demonstrated a net increase of 41.2%, significantly higher than UFF, confirming protein-protective effect of PEE through delayed myofibrillar protein denaturation and subsequent gradual FAA release. Among all groups, Gly, Cys, Ala, and His predominated with trends mirroring total FAAs, wherein Gly contributed critically to flavor balance as a sweet-tasting amino acid. At storage conclusion, the PEE group retained 1.29-fold higher Asp and showed a 0.38 mg/100 g increase in Glu versus initial values. Given their roles as core umami compounds, these results confirm that the elevated FAAs in PEE group signify flavor preservation rather than spoilage. Collectively, PLEE modulates FAA metabolism through multi-target regulation, simultaneously inhibiting early-stage spoilage while selectively enriching umami amino acids.Table 3Changes in FAAs of snakehead fillets with different treatments during 4 °C cold storage.GroupsFAAs (mg/100 g)AspSerThrGluGlyAlaCysValUFF_0d0.17 ± 0.06Aa2.7 ± 0.25Bb5.65 ± 0.7Ab4.42 ± 0.46Ab37.77 ± 3.7Ab9.35 ± 1.06ABb14.51 ± 1.54Ab0.73 ± 0.01AbSDW_0d0.08 ± 0.01Aa2.99 ± 0.16Db0.02 ± 0.01Aa3.99 ± 0.36Ab32.23 ± 1.4BCb7.88 ± 0.35Ab12.19 ± 0.45Cb0.8 ± 0.05AbPEE_0d0.1 ± 0.07Aa1.92 ± 0.22BCa2.31 ± 2Aa2.87 ± 0.11Aa22.2 ± 1.73Aa4.91 ± 0.37Aa9.02 ± 0.63ABa0.57 ± 0.05AaUFF_3d0.1 ± 0.03Aa2.45 ± 0.62Ba4.74 ± 1.18Aa5.11 ± 1.3Aa33.48 ± 9.57Aa8.34 ± 2.54Aa11.7 ± 3.36Aa0.9 ± 0.24AaSDW_3d0.06 ± 0.01Aa2.28 ± 0.11Ca4.55 ± 0.17Ca4.73 ± 0.23Ca29.05 ± 1.83ABa6.98 ± 0.47Aa10.3 ± 0.74ABa0.81 ± 0.07AaPEE_3d0.07 ± 0.01Aa2.25 ± 0.16Ca4.38 ± 0.32Aa4.28 ± 0.34Ba26.86 ± 2.03Aa6.65 ± 0.58Aa10.7 ± 1.1Ba0.87 ± 0.08BaUFF_6d0.08 ± 0.01Aa2.21 ± 0.01Bb4.75 ± 0.18Aa4.93 ± 0.19Ab39.96 ± 1.81Ab10.87 ± 0.66ABb14.05 ± 0.95Ab1.11 ± 0.07AbSDW_6d0.09 ± 0.07Aa1.23 ± 0.22Ba3.35 ± 0.39Ba2.88 ± 0.62Ba25.23 ± 3.21Aa6.63 ± 0.83Aa8.94 ± 1.03Aa0.66 ± 0.07AaPEE_6d0.1 ± 0.1Aa1.55 ± 0.45Bab2.2 ± 1.96Aa2.55 ± 0.58Aa23.71 ± 4.33Aa6.03 ± 1.14Aa8.5 ± 1.45ABa0.72 ± 0.13ABaUFF_9d1.25 ± 0.39Bb1.17 ± 0.11Ab5.61 ± 0.49Ab4.76 ± 0.62Ab43.99 ± 2.62Ac13.76 ± 1.32Bb13.92 ± 1.23Ac1.19 ± 0.13ABbSDW_9d1.51 ± 0.4Bb1.41 ± 0.01Bc4.97 ± 0.03Cb4.76 ± 0.04Cb34.12 ± 0.18Cb12.95 ± 0.06Bb10.66 ± 0.04BCb2.21 ± 0.01CcPEE_9d0.05 ± 0.01Aa0.73 ± 0.08Aa2.94 ± 0.27Aa2.25 ± 0.27Aa20.71 ± 1.68Aa6.27 ± 0.55Aa7.36 ± 0.55Aa0.8 ± 0.02ABaUFF_12d0.94 ± 0.21Ba0.96 ± 0.28Aa5.14 ± 0.86Aa4.31 ± 0.96Ab38.22 ± 5.82Ab13.24 ± 2.85ABa12.72 ± 2.58Ab1.83 ± 0.48BaSDW_12d1.03 ± 0.05Ba0.8 ± 0.05Aa4.74 ± 0.08Ca2.54 ± 0.13Ca33.57 ± 1.02BCab11.64 ± 0.41Ba11.14 ± 0.39BCab1.68 ± 0.09BaPEE_12d1.21 ± 0.39Ba1.26 ± 0.19ABa4.91 ± 0.09Aa3.25 ± 0.41Aab25.68 ± 1.63Aa9.88 ± 0.83Ba8.46 ± 0.76ABa1.7 ± 0.17CaGroupsFAAs (mg/100 g)LeuTyrPheHisLysArgPrototalUFF_0d1.28 ± 0.06Ab1.24 ± 0.08Ab0.66 ± 0.14Ab13.09 ± 1Ac4.77 ± 0.42Ab1.17 ± 0.11Ab2.37 ± 0.28Ac100.97 ± 9.97AcSDW_0d1.35 ± 0.13Bb1.19 ± 0.12ABCb0.77 ± 0.1Cb11.33 ± 0.57BCb4.41 ± 0.22Bb1.13 ± 0.05Ab1.81 ± 0.07Ab83.36 ± 4.3BbPEE_0d0.59 ± 0.02Aa0.7 ± 0.01ABa0.38 ± 0.02Aa8.36 ± 0.21ABa3.26 ± 0.36ABa0.91 ± 0.08Aa1.22 ± 0.04Aa59.92 ± 6.08ABaUFF_3d1.41 ± 0.55Aa1.38 ± 0.56Aa0.88 ± 0.41ABa11.25 ± 3.43Aa4.35 ± 1.31Aa1.09 ± 0.33Aa2.32 ± 0.84Aa90.4 ± 26.53AaSDW_3d1.49 ± 0.29Ba1.47 ± 0.26BCa0.74 ± 0.17Ba9.87 ± 0.7ABa4.16 ± 0.36Ba1.12 ± 0.11Ba1.87 ± 0.13Aa80.48 ± 6.04BaPEE_3d0.96 ± 0.21Aa0.93 ± 0.2ABa0.67 ± 0.09Aa9.74 ± 1.12ABa3.96 ± 0.37Ba1.15 ± 0.11ABa1.86 ± 0.22Ba76.19 ± 7.14BCaUFF_6d1.67 ± 0.08Ab1.07 ± 0.15Aa0.63 ± 0.01Aa12.48 ± 1.26Ab4.74 ± 0.29Ab0.57 ± 0.05Aa2.14 ± 0.16Ab102.49 ± 5.97AbSDW_6d0.85 ± 0.21Aa0.95 ± 0.26Aa0.65 ± 0.17Aa8.83 ± 1.28Aa3.29 ± 0.45Aa0.37 ± 0.06Ca1.63 ± 0.2Aab66.15 ± 9.15AaPEE_6d0.6 ± 0.36Aa0.99 ± 0.3ABa0.55 ± 0.03Aa8.41 ± 1.66ABa3.45 ± 0.62ABa1.87 ± 1.30ABCa1.37 ± 0.3ABa63.72 ± 15.45ABaUFF_9d1.81 ± 0.33Ab1.41 ± 0.21Ab1.51 ± 0.24BCb14.35 ± 1.27Ab4.15 ± 0.41Ab2.2 ± 1.65Aa2.77 ± 0.27ABa115.88 ± 11.58AbSDW_9d3.39 ± 0.01Dc1.58 ± 0.02Cb2.69 ± 0.02Bc13.23 ± 0.03CDb5.90 ± 0.01Cc0.53 ± 0.03Da4.46 ± 1.06Bb109 ± 2.14CbPEE_9d0.74 ± 0.12Aa0.61 ± 0.12Aa0.59 ± 0.07Aa7.22 ± 0.79Aa2.47 ± 0.19Aa2.52 ± 0.13BCa1.21 ± 0.16Aa57.14 ± 5.1AaUFF_12d2.7 ± 1.02Aa0.86 ± 0.33Aa1.76 ± 0.47Ca13 ± 2.88Aa3.11 ± 0.65Aa1.85 ± 1.17Aa4.96 ± 1.92Ba109.16 ± 23.76AaSDW_12d2.45 ± 0.14Ca1.01 ± 0.14ABa1.67 ± 0.17Aa13.36 ± 0.51Da3.65 ± 0.16ABa2.15 ± 0.12Ca3.91 ± 0.14Ba98.37 ± 3.77CaPEE_12d2.25 ± 0.35Ba1.13 ± 0.2Ba1.66 ± 0.26Ba10.33 ± 1.11Ba2.87 ± 0.27Aa2.81 ± 0.21Ca4.4 ± 0.17Ca84.6 ± 7.31CaNote: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05). Table 3Changes in FAAs of snakehead fillets with different treatments during 4 °C cold storage.GroupsFAAs (mg/100 g)AspSerThrGluGlyAlaCysValUFF_0d0.17 ± 0.06Aa2.7 ± 0.25Bb5.65 ± 0.7Ab4.42 ± 0.46Ab37.77 ± 3.7Ab9.35 ± 1.06ABb14.51 ± 1.54Ab0.73 ± 0.01AbSDW_0d0.08 ± 0.01Aa2.99 ± 0.16Db0.02 ± 0.01Aa3.99 ± 0.36Ab32.23 ± 1.4BCb7.88 ± 0.35Ab12.19 ± 0.45Cb0.8 ± 0.05AbPEE_0d0.1 ± 0.07Aa1.92 ± 0.22BCa2.31 ± 2Aa2.87 ± 0.11Aa22.2 ± 1.73Aa4.91 ± 0.37Aa9.02 ± 0.63ABa0.57 ± 0.05AaUFF_3d0.1 ± 0.03Aa2.45 ± 0.62Ba4.74 ± 1.18Aa5.11 ± 1.3Aa33.48 ± 9.57Aa8.34 ± 2.54Aa11.7 ± 3.36Aa0.9 ± 0.24AaSDW_3d0.06 ± 0.01Aa2.28 ± 0.11Ca4.55 ± 0.17Ca4.73 ± 0.23Ca29.05 ± 1.83ABa6.98 ± 0.47Aa10.3 ± 0.74ABa0.81 ± 0.07AaPEE_3d0.07 ± 0.01Aa2.25 ± 0.16Ca4.38 ± 0.32Aa4.28 ± 0.34Ba26.86 ± 2.03Aa6.65 ± 0.58Aa10.7 ± 1.1Ba0.87 ± 0.08BaUFF_6d0.08 ± 0.01Aa2.21 ± 0.01Bb4.75 ± 0.18Aa4.93 ± 0.19Ab39.96 ± 1.81Ab10.87 ± 0.66ABb14.05 ± 0.95Ab1.11 ± 0.07AbSDW_6d0.09 ± 0.07Aa1.23 ± 0.22Ba3.35 ± 0.39Ba2.88 ± 0.62Ba25.23 ± 3.21Aa6.63 ± 0.83Aa8.94 ± 1.03Aa0.66 ± 0.07AaPEE_6d0.1 ± 0.1Aa1.55 ± 0.45Bab2.2 ± 1.96Aa2.55 ± 0.58Aa23.71 ± 4.33Aa6.03 ± 1.14Aa8.5 ± 1.45ABa0.72 ± 0.13ABaUFF_9d1.25 ± 0.39Bb1.17 ± 0.11Ab5.61 ± 0.49Ab4.76 ± 0.62Ab43.99 ± 2.62Ac13.76 ± 1.32Bb13.92 ± 1.23Ac1.19 ± 0.13ABbSDW_9d1.51 ± 0.4Bb1.41 ± 0.01Bc4.97 ± 0.03Cb4.76 ± 0.04Cb34.12 ± 0.18Cb12.95 ± 0.06Bb10.66 ± 0.04BCb2.21 ± 0.01CcPEE_9d0.05 ± 0.01Aa0.73 ± 0.08Aa2.94 ± 0.27Aa2.25 ± 0.27Aa20.71 ± 1.68Aa6.27 ± 0.55Aa7.36 ± 0.55Aa0.8 ± 0.02ABaUFF_12d0.94 ± 0.21Ba0.96 ± 0.28Aa5.14 ± 0.86Aa4.31 ± 0.96Ab38.22 ± 5.82Ab13.24 ± 2.85ABa12.72 ± 2.58Ab1.83 ± 0.48BaSDW_12d1.03 ± 0.05Ba0.8 ± 0.05Aa4.74 ± 0.08Ca2.54 ± 0.13Ca33.57 ± 1.02BCab11.64 ± 0.41Ba11.14 ± 0.39BCab1.68 ± 0.09BaPEE_12d1.21 ± 0.39Ba1.26 ± 0.19ABa4.91 ± 0.09Aa3.25 ± 0.41Aab25.68 ± 1.63Aa9.88 ± 0.83Ba8.46 ± 0.76ABa1.7 ± 0.17CaGroupsFAAs (mg/100 g)LeuTyrPheHisLysArgPrototalUFF_0d1.28 ± 0.06Ab1.24 ± 0.08Ab0.66 ± 0.14Ab13.09 ± 1Ac4.77 ± 0.42Ab1.17 ± 0.11Ab2.37 ± 0.28Ac100.97 ± 9.97AcSDW_0d1.35 ± 0.13Bb1.19 ± 0.12ABCb0.77 ± 0.1Cb11.33 ± 0.57BCb4.41 ± 0.22Bb1.13 ± 0.05Ab1.81 ± 0.07Ab83.36 ± 4.3BbPEE_0d0.59 ± 0.02Aa0.7 ± 0.01ABa0.38 ± 0.02Aa8.36 ± 0.21ABa3.26 ± 0.36ABa0.91 ± 0.08Aa1.22 ± 0.04Aa59.92 ± 6.08ABaUFF_3d1.41 ± 0.55Aa1.38 ± 0.56Aa0.88 ± 0.41ABa11.25 ± 3.43Aa4.35 ± 1.31Aa1.09 ± 0.33Aa2.32 ± 0.84Aa90.4 ± 26.53AaSDW_3d1.49 ± 0.29Ba1.47 ± 0.26BCa0.74 ± 0.17Ba9.87 ± 0.7ABa4.16 ± 0.36Ba1.12 ± 0.11Ba1.87 ± 0.13Aa80.48 ± 6.04BaPEE_3d0.96 ± 0.21Aa0.93 ± 0.2ABa0.67 ± 0.09Aa9.74 ± 1.12ABa3.96 ± 0.37Ba1.15 ± 0.11ABa1.86 ± 0.22Ba76.19 ± 7.14BCaUFF_6d1.67 ± 0.08Ab1.07 ± 0.15Aa0.63 ± 0.01Aa12.48 ± 1.26Ab4.74 ± 0.29Ab0.57 ± 0.05Aa2.14 ± 0.16Ab102.49 ± 5.97AbSDW_6d0.85 ± 0.21Aa0.95 ± 0.26Aa0.65 ± 0.17Aa8.83 ± 1.28Aa3.29 ± 0.45Aa0.37 ± 0.06Ca1.63 ± 0.2Aab66.15 ± 9.15AaPEE_6d0.6 ± 0.36Aa0.99 ± 0.3ABa0.55 ± 0.03Aa8.41 ± 1.66ABa3.45 ± 0.62ABa1.87 ± 1.30ABCa1.37 ± 0.3ABa63.72 ± 15.45ABaUFF_9d1.81 ± 0.33Ab1.41 ± 0.21Ab1.51 ± 0.24BCb14.35 ± 1.27Ab4.15 ± 0.41Ab2.2 ± 1.65Aa2.77 ± 0.27ABa115.88 ± 11.58AbSDW_9d3.39 ± 0.01Dc1.58 ± 0.02Cb2.69 ± 0.02Bc13.23 ± 0.03CDb5.90 ± 0.01Cc0.53 ± 0.03Da4.46 ± 1.06Bb109 ± 2.14CbPEE_9d0.74 ± 0.12Aa0.61 ± 0.12Aa0.59 ± 0.07Aa7.22 ± 0.79Aa2.47 ± 0.19Aa2.52 ± 0.13BCa1.21 ± 0.16Aa57.14 ± 5.1AaUFF_12d2.7 ± 1.02Aa0.86 ± 0.33Aa1.76 ± 0.47Ca13 ± 2.88Aa3.11 ± 0.65Aa1.85 ± 1.17Aa4.96 ± 1.92Ba109.16 ± 23.76AaSDW_12d2.45 ± 0.14Ca1.01 ± 0.14ABa1.67 ± 0.17Aa13.36 ± 0.51Da3.65 ± 0.16ABa2.15 ± 0.12Ca3.91 ± 0.14Ba98.37 ± 3.77CaPEE_12d2.25 ± 0.35Ba1.13 ± 0.2Ba1.66 ± 0.26Ba10.33 ± 1.11Ba2.87 ± 0.27Aa2.81 ± 0.21Ca4.4 ± 0.17Ca84.6 ± 7.31CaNote: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05). Changes in FAAs of snakehead fillets with different treatments during 4 °C cold storage. Note: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05). Flavor analysis Sensory evaluation The dynamic evaluation of odor attributes in three groups of snakehead fillets during refrigerated storage, assessed via Quantitative Descriptive Analysis, is presented in Fig. 1(F). As storage time increased, fishy odor intensified significantly across all samples (P < 0.05), with the UFF group exhibiting the most pronounced increment. The PEE not only demonstrated significantly lower fishy odor scores than other groups (P < 0.01) but also showed a reduction in the rate of increase by 55.6% compared with the control group. This phenomenon confirms the efficacy of PLEE in mitigating fishy odor accumulation, attributable to their inhibitory effects on microorganisms and regulation of TMA generation as previously reported. Critically, the PEE displayed unique volatile characteristics during early storage, with herbaceous and citrus notes reaching intensities 3.92 and 8.25 times higher than the UFF, respectively. This divergence originates from the flavor-enhancing effects of characteristic aroma components ((−)-Perillaldehyde, Linalool) in perilla, where terpenoid compounds may reduce fishy odor perception through masking effects while simultaneously inhibiting spoilage metabolite formation via antimicrobial activity (Cetinkaya & Ayseli, 2024). Integrating lipidomic and microbial community analyses, PLEE maintained flavor stability through dual pathways: inhibiting the biosynthesis of fishy precursors (aldehydes, TMA) while enhancing overall aroma profiles through plant-derived volatiles. Volatile flavor compounds Volatile flavor compounds (VFCs) in PLEE were analyzed using SPME-GC–MS (Table 4), identifying 39 compounds comprising 2 hydrocarbons, 4 aldehydes, 4 ketones, 4 alcohols, 5 aromatics, 18 terpenoids, and 2 other categories. Aromatics constituted the predominant fraction (53.13%), with myristicin (52.65%) as the principal constituent. This phytochemical, ubiquitous in perilla and nutmeg, exhibits a spicy warm-woody aroma and demonstrates anti-proliferative, antimicrobial, and anti-inflammatory properties (Bi et al., 2025). Terpenoids represented the second largest group (33.59%), dominated by β-caryophyllene (13.43%) and caryophyllene oxide (8.88%). The sesquiterpene caryophyllene oxide can act as a natural antioxidant that scavenges free radicals and inhibit lipid peroxidation (Liu, Wen, et al., 2024). β-caryophyllene additionally inhibits periodontal pathogen activity and inflammatory mediator expression while conferring anti-inflammatory, analgesic, and hypolipidemic bioactivities (Spence, 2024).Table 4The content of VFCs in the PLEE.CategorynameCASformulaRTRIConcentrations (ug/mL)Hydrocarbons (2)Tetradecane629–59-4C14H3015.761440.480.1 ± 0.092,6-Dimethyl-2,4,6-octatriene673–84-7C10H1614.221380.850.004 ± 0.00Aldehydes (4)Nonanal124–19-6C9H18O14.661396.080.06 ± 0.02Benzaldehyde100–52-7C7H6O17.071490.260.38 ± 0.40Decanal112–31-2C10H20O17.451504.040.02 ± 0.01(−)-Perillaldehyde18,031–40-8C10H14O23.561768.292.72 ± 1.24Ketones (4)3,5-Dimethyl-2-cyclohexen-1-one1123-09-7C8H12O18.701557.630.01 ± 0.00(E)-6-Methyl-3,5-heptadien-2-one16,647–04-4C8H12O19.081572.980.01 ± 0.016,10-Dimethylundecan-2-one1604-34-8C13H26O22.141704.470.04 ± 0.00β-Ionone79–77-6C13H20O27.091935.841.55 ± 0.44Alcohols (4)Linalool78–70-6C10H18O18.351543.812.96 ± 0.20Citronellol106–22-9C10H20O23.401762.010.04 ± 0.00Nerol106–25-2C10H18O24.021789.141.2 ± 0.12Nerolidol7212-44-4C15H26O29.362047.821.74 ± 0.36Aromatics (5)1,2,3-Trimethylbenzene526–73-8C9H1212.801331.310.01 ± 0.001,6-Dimethylnaphthalene575–43-9C12H1227.981979.400.02 ± 0.01Methyleugenol93–15-2C11H14O228.321996.060.27 ± 0.06Elemicin487–11-6C12H16O332.412212.620.1 ± 0.02Myristicin607–91-0C11H12O332.992245.0843.67 ± 9.13Terpenes (18)β-Myrcene123–35-3C10H168.611180.510.02 ± 0.01β-Ocimene13,877–91-3C10H1610.971265.930.01 ± 0.00α-Copaene3856-25-5C15H2416.881483.900.11 ± 0.051-ethenyl-1-methyl-2-(1-methylethenyl)-4-(1-methylethylidene)-Cyclohexane3242-08-8C15H2417.351501.710.03 ± 0.02Cyclosativene22,469–52-9C15H2417.731517.530.2 ± 0.16β-Caryophyllene87–44-5C15H2420.201620.1911.14 ± 3.75α-Caryophyllene6753-98-6C15H2421.791689.402.13 ± 0.77(+)-Ledene21,747–46-6C15H2422.481720.370.04 ± 0.01Eremophilene10,219–75-7C15H2422.931740.790.05 ± 0.01β-Curcumene28,976–67-2C15H2423.221754.042.62 ± 0.38α-Farnesene502–61-4C15H2423.701775.480.37 ± 0.09(+)-d-Cadinene483–76-1C15H2423.841781.461.44 ± 0.30trans-Calamenene73,209–42-4C15H2225.291849.160.14 ± 0.12p-Menth-1(7)-en-9-ol29,548–16-1C10H18O26.131888.520.41 ± 0.09Caryophylleneoxide1139-30-6C15H24O28.151987.907.36 ± 1.76Salvial-4(14)-en-1-one73,809–82-2C15H24O28.652012.420.9 ± 0.19(1R,3E,7E,11R)-1,5,5,8-Tetramethyl-12-oxabicyclo[9.1.0]dodeca-3,7-diene19,888–34-7C15H24O29.212040.540.68 ± 0.13Eugenol97–53-0C10H12O230.922128.280.23 ± 0.11Others (2)4-Hydroxy-1-methyl-3-methallyl-benzol1786-08-9C10H16O16.531470.610.03 ± 0.00trans-Linalooloxide(furanoid)34,995–77-2C10H18O215.611435.030.15 ± 0.12 Table 4The content of VFCs in the PLEE.CategorynameCASformulaRTRIConcentrations (ug/mL)Hydrocarbons (2)Tetradecane629–59-4C14H3015.761440.480.1 ± 0.092,6-Dimethyl-2,4,6-octatriene673–84-7C10H1614.221380.850.004 ± 0.00Aldehydes (4)Nonanal124–19-6C9H18O14.661396.080.06 ± 0.02Benzaldehyde100–52-7C7H6O17.071490.260.38 ± 0.40Decanal112–31-2C10H20O17.451504.040.02 ± 0.01(−)-Perillaldehyde18,031–40-8C10H14O23.561768.292.72 ± 1.24Ketones (4)3,5-Dimethyl-2-cyclohexen-1-one1123-09-7C8H12O18.701557.630.01 ± 0.00(E)-6-Methyl-3,5-heptadien-2-one16,647–04-4C8H12O19.081572.980.01 ± 0.016,10-Dimethylundecan-2-one1604-34-8C13H26O22.141704.470.04 ± 0.00β-Ionone79–77-6C13H20O27.091935.841.55 ± 0.44Alcohols (4)Linalool78–70-6C10H18O18.351543.812.96 ± 0.20Citronellol106–22-9C10H20O23.401762.010.04 ± 0.00Nerol106–25-2C10H18O24.021789.141.2 ± 0.12Nerolidol7212-44-4C15H26O29.362047.821.74 ± 0.36Aromatics (5)1,2,3-Trimethylbenzene526–73-8C9H1212.801331.310.01 ± 0.001,6-Dimethylnaphthalene575–43-9C12H1227.981979.400.02 ± 0.01Methyleugenol93–15-2C11H14O228.321996.060.27 ± 0.06Elemicin487–11-6C12H16O332.412212.620.1 ± 0.02Myristicin607–91-0C11H12O332.992245.0843.67 ± 9.13Terpenes (18)β-Myrcene123–35-3C10H168.611180.510.02 ± 0.01β-Ocimene13,877–91-3C10H1610.971265.930.01 ± 0.00α-Copaene3856-25-5C15H2416.881483.900.11 ± 0.051-ethenyl-1-methyl-2-(1-methylethenyl)-4-(1-methylethylidene)-Cyclohexane3242-08-8C15H2417.351501.710.03 ± 0.02Cyclosativene22,469–52-9C15H2417.731517.530.2 ± 0.16β-Caryophyllene87–44-5C15H2420.201620.1911.14 ± 3.75α-Caryophyllene6753-98-6C15H2421.791689.402.13 ± 0.77(+)-Ledene21,747–46-6C15H2422.481720.370.04 ± 0.01Eremophilene10,219–75-7C15H2422.931740.790.05 ± 0.01β-Curcumene28,976–67-2C15H2423.221754.042.62 ± 0.38α-Farnesene502–61-4C15H2423.701775.480.37 ± 0.09(+)-d-Cadinene483–76-1C15H2423.841781.461.44 ± 0.30trans-Calamenene73,209–42-4C15H2225.291849.160.14 ± 0.12p-Menth-1(7)-en-9-ol29,548–16-1C10H18O26.131888.520.41 ± 0.09Caryophylleneoxide1139-30-6C15H24O28.151987.907.36 ± 1.76Salvial-4(14)-en-1-one73,809–82-2C15H24O28.652012.420.9 ± 0.19(1R,3E,7E,11R)-1,5,5,8-Tetramethyl-12-oxabicyclo[9.1.0]dodeca-3,7-diene19,888–34-7C15H24O29.212040.540.68 ± 0.13Eugenol97–53-0C10H12O230.922128.280.23 ± 0.11Others (2)4-Hydroxy-1-methyl-3-methallyl-benzol1786-08-9C10H16O16.531470.610.03 ± 0.00trans-Linalooloxide(furanoid)34,995–77-2C10H18O215.611435.030.15 ± 0.12 The content of VFCs in the PLEE. VFCs in three snakehead fillet groups were analyzed via SPME-GC–MS, detecting 95 compounds including 13 aldehydes, 13 ketones, 20 alcohols, 26 aromatics, 10 terpenoids, 4 hydrocarbons, and 4 other compounds (Table S1). PEE demonstrated superior total VFCs from initial storage: at day 0, its concentration reached 3308.46 μg/kg (28 compounds), significantly exceeding 200.04 μg/kg (26 compounds) of UFF, with 10 shared VFCs. For day 12, PEE increased to 4389.59 μg/kg (44 compounds) while UFF reached only 262.99 μg/kg (30 compounds), with shared compounds rising to 24. This enhancement was attributed to substantial flavor compounds introduced during initial PLEE treatment. Importantly, the PEE group consistently maintained higher VFC concentrations than other groups and exhibited distinct evolutionary characteristics compared to UFF: while UFF displayed its initial VFC rise by day 6, the PEE group remained in a decreasing phase at this timepoint. These differential change correlated directly with antioxidant and antimicrobial activities mediated by abundant volatile constituents in PLEE (β-caryophyllene, (−)-perillaldehyde), which effectively inhibited lipid oxidation-derived enzymatic degradation and microbial metabolism, thereby delaying undesirable flavor formation. Aldehydes, formed through fatty acid oxidation and amino acid degradation, are key components of aquatic flavor systems due to their low molecular weight and high volatility (Xi et al., 2023). These compounds exhibit dual effects on overall flavor profiles: while hexanal and nonanal impart characteristic aromas, certain short-chain oxidized aldehydes contribute to undesirable odors. In this experiment, UFF exhibited only a 10.14% increase in aldehydes at the end of storage, whereas the PEE group demonstrated a significant 946.04% surge. This disparity originated primarily from benzaldehyde, detected in PLEE, which contributes almond and nutty notes. Hexanal and heptanal were exclusively identified in UFF and SDW groups during early storage, conferring grassy aromas at low concentrations. It is of note while heptanal alone does not produce fishy odors, it synergistically enhances fishiness when combined with compounds such as (E, Z)-3,5-octadien-2-one (Dai et al., 2024). The absence of this compound in PEE may reflect odor-masking effects of PLEE. Crucially, perillaldehyde, a signature terpenoid aldehyde in PLEE, was unique to PEE. This compound optimized the overall flavor profile of snakehead fillets by imparting woody-citrus notes, highlighting the sensory-modulating advantages of PLEE treatment, findings consistent with the aforementioned sensory evaluations. Ketones emerged as common derivatives originating from lipid oxidation, microbial metabolism, and amino acid degradation (Zhou et al., 2024). During initial storage, UFF solely exhibited acetophenone (1.17 μg/kg) and m-ethylacetophenone (0.16 μg/kg), whereas PEE demonstrated detectable levels of 3,5-dimethyl-2-cyclohexen-1-one (1.25 μg/kg), 3′-methylacetophenone (3.16 μg/kg), naginataketone (2068.56 μg/kg), and β-ionone (1.66 μg/kg). By the terminal storage phase, all groups manifested increased naginataketone content, with increments of 41.56 μg/kg (UFF), 38.28 μg/kg (SDW), and 5.09 μg/kg (PEE). This progression unequivocally indicated that PLEE substantially inhibited the accumulation of this specific ketone. Regarding alcohols during the initial storage phase, the total alcohol content in PEE reached 4.68 times that of UFF. This disparity was primarily attributed to the introduction of aromatic alcohols, including linalool, nerol, benzyl alcohol, and phenethyl alcohol, from PLEE, which imparted pleasant aromatic notes to the fish fillets. Specifically, 1-octen-3-ol exhibited significant flavor influence owing to its low odor threshold, with its characteristic mushroom/earthy odor serving as a key contributor to fishy off-notes. In UFF, 1-octen-3-ol peaked at 21.09 μg/kg on day 3, whereas in PEE, the peak (9.08 μg/kg) was delayed until day 9, with a mere 1.37 μg/kg increase observed within the first 6 days. This temporal profile closely correlated with Pseudomonas population dynamics in the respective treatment groups, substantiating that PLEE effectively inhibited 1-octen-3-ol generation. Regarding aromatic compounds, UFF initially contained nine distinct species, whereas the PEE group exhibited only four types, predominantly myristicin (554.50 μg/kg). This compound imparts warm spicy notes with woody sweetness and functions as a signature component of Perilla. During storage, terpenes in the UFF group gradually decreased to only eugenol, whereas PEE showed smaller reduction in terpene diversity, with total content increasing by 14.80%. This increase primarily stemmed from the accumulation of caryophyllene oxide and eugenol. Esters and furans also demonstrated low flavor thresholds throughout storage. UFF produced merely one ester (methyl benzoate) during the terminal phase, while PEE yielded two esters (mipsin and dihydroactinidiolide) along with two furanoids: (Z)-linalool oxide (furanoid) and (E)-linalool oxide (furanoid). These compounds conferred subtle herbaceous notes to the fish fillets, suggesting their potential roles as active constituents for mitigating undesirable flavors in refrigerated snakehead fillets. The OAV served as a critical indicator for assessing the contribution of VFCs to the overall flavor profile in aquatic products. Table 5 demonstrated the identification of 29 key flavor-contributing compounds with OAV > 0.05 across three sample groups (Fig. 2A). In addition, 38 differential compounds with variable importance in projection (VIP) values >1 were further screened by orthogonal partial least squares discriminant analysis (OPLS-DA), as illustrated in the content heatmap (Fig. 2B). Among these, 12 key differential compounds simultaneously satisfied OAV > 0.05 and VIP > 1: octanal, nonanal, dodecanal, dl-perillaldehyde, acetophenone, 1-octen-3-ol, 1-heptanol, (E)-2-octen-1-ol, 1,4-diethylbenzene, indane, 2-methylnaphthalene, and (E)-linalool oxide (furanoid).Table 5OAV values of VFCs (OAV > 0.05) in snakehead with different treatments during 4 °C cold storage.CASCompoundsOdor threshold (μg/kg)OdorOAV valuesUFF_0dSDW_0dPEE_0dUFF_3dSDW_3dPEE_3dUFF_6dSDW_6dPEE_6dUFF_9dSDW_9dPEE_9dUFF_12dSDW_12dPEE_12d111–71-7Heptanal2.8Fresh, aldehydic, fatty, green, herbal and wine-lee ozone0.600.53–0.510.59––––––––––124–13-0Octanal0.587Green, waxy, fatty, citrus and orange–13.00–11.608.62–1.822.733.732.76–8.82––1.04124–19-6Nonanal1.1Waxy, rose, fresh, orris, orange and fatty46.6037.8820.45––18.6417.3058.5621.2729.95–50.5411.0514.0325.54100–52-7Benzaldehyde750.89sweet, oily, almond, cherry, nutty and woody0.010.020.030.010.000.020.010.020.030.030.021.350.080.050.96112–31-2Decanal3waxy, fatty, citrus and orange peel with a slight green melon nuance1.642.255.261.382.121.901.314.611.583.041.554.830.841.343.04112–54-9Dodecanal0.13soapy, waxy, citrus and orange12.238.38–25.922.31––––––13.31–––2111-75-3dl-Perillaldehyde30woody, spicy, waxy, sweet, citrus, lime and aldehydic––––––––0.05–––––0.1518,031–40-8(−)-Perillaldehyde30Aromatic, sweet, woody, spicy, waxy, orange, lime and aldehydic––0.23––0.08–––––0.09–––98–86-2Acetophenone65powdery, bitter almond cherry pit-like with coumarinic and fruity nuances0.020.02–0.010.01–0.07–––––0.010.02–122–00-94′-Methylacetophenone21Sweet, creamy, fruity, cherry, heliotropine–––––0.10–––––0.03––0.0679–77-6β-Ionone0.007woody, floral, berry and fruity with powdery nuances––237.14––212.86––150.00––––––3391-86-41-Octen-3-ol31Mushroom, earthy, green, oily, vegetative, umami and savory0.310.53–0.680.180.060.050.110.040.080.030.290.09–0.21111–70-61-Heptanol5.4Oily, nutty and green––––2.02––––––––––78–70-6Linalool30Citrus, orange, lemon, floral and woody––6.640.440.156.90––5.43–0.0712.860.030.0210.2818,409–17-1(E)-2-Octen-1-ol20Fatty, oily, sweet and fruity–––0.060.06––––––––––143–08-81-Nonanol45.5Waxy, citrus, rue aldehydic, orange, oily, fatty and spicy–––––0.300.05–0.050.000.010.08––0.0860–12-8Phenylethyl Alcohol564.23floral, sweet, rosey and bready––0.01––0.01––0.010.000.000.040.000.000.05128–37-0Butylated Hydroxytoluene1000Mild, phenolic and camphor0.070.040.250.030.090.190.140.120.110.050.050.140.080.100.11105–05-51,4-Diethylbenzene2.1––––––––––––0.460.380.90496–11-7Indane100.050.07––0.10–0.51––0.080.07–0.050.06–91–57-62-Methylnaphthalene3Sweet, floral, woody, oily and aromatic0.350.40–0.100.11–0.110.260.23––0.450.650.160.51607–91-0Myristicin88spicy warm balsamic and woody0.030.016.300.020.005.670.040.134.300.030.016.870.050.057.95123–35-3β-Myrcene1.2Peppery, terpene, spicy and balsam––––––––2.87––7.93–––100–42-5Styrene65Sweet, balsam and floral0.200.130.370.070.230.250.250.210.140.220.090.43––0.0487–44-5β-Caryophyllene64Spicy, clove, woody, nut and peppery––––––––0.23––0.12––0.101139-30-6Caryophyllene oxide410Sweet, fresh, dry, woody and spicy––0.04––0.05––0.02–––––0.0797–53-0Eugenol470Sweet, warm spicy, clove and woody0.010.010.040.000.000.030.000.010.020.000.000.050.010.010.085989-33-3(Z)-linalool oxide (furanoid)100Earthy, floral, sweet and woody–––––0.36–––––0.69––0.5534,995–77-2(E)-Linalool oxide (furanoid)190floral––0.15–––––0.20––––––Fig. 2Heatmap of VFCs with OAV > 0.05 (A); VIP analysis of VFCs (B).Fig. 2 Table 5OAV values of VFCs (OAV > 0.05) in snakehead with different treatments during 4 °C cold storage.CASCompoundsOdor threshold (μg/kg)OdorOAV valuesUFF_0dSDW_0dPEE_0dUFF_3dSDW_3dPEE_3dUFF_6dSDW_6dPEE_6dUFF_9dSDW_9dPEE_9dUFF_12dSDW_12dPEE_12d111–71-7Heptanal2.8Fresh, aldehydic, fatty, green, herbal and wine-lee ozone0.600.53–0.510.59––––––––––124–13-0Octanal0.587Green, waxy, fatty, citrus and orange–13.00–11.608.62–1.822.733.732.76–8.82––1.04124–19-6Nonanal1.1Waxy, rose, fresh, orris, orange and fatty46.6037.8820.45––18.6417.3058.5621.2729.95–50.5411.0514.0325.54100–52-7Benzaldehyde750.89sweet, oily, almond, cherry, nutty and woody0.010.020.030.010.000.020.010.020.030.030.021.350.080.050.96112–31-2Decanal3waxy, fatty, citrus and orange peel with a slight green melon nuance1.642.255.261.382.121.901.314.611.583.041.554.830.841.343.04112–54-9Dodecanal0.13soapy, waxy, citrus and orange12.238.38–25.922.31––––––13.31–––2111-75-3dl-Perillaldehyde30woody, spicy, waxy, sweet, citrus, lime and aldehydic––––––––0.05–––––0.1518,031–40-8(−)-Perillaldehyde30Aromatic, sweet, woody, spicy, waxy, orange, lime and aldehydic––0.23––0.08–––––0.09–––98–86-2Acetophenone65powdery, bitter almond cherry pit-like with coumarinic and fruity nuances0.020.02–0.010.01–0.07–––––0.010.02–122–00-94′-Methylacetophenone21Sweet, creamy, fruity, cherry, heliotropine–––––0.10–––––0.03––0.0679–77-6β-Ionone0.007woody, floral, berry and fruity with powdery nuances––237.14––212.86––150.00––––––3391-86-41-Octen-3-ol31Mushroom, earthy, green, oily, vegetative, umami and savory0.310.53–0.680.180.060.050.110.040.080.030.290.09–0.21111–70-61-Heptanol5.4Oily, nutty and green––––2.02––––––––––78–70-6Linalool30Citrus, orange, lemon, floral and woody––6.640.440.156.90––5.43–0.0712.860.030.0210.2818,409–17-1(E)-2-Octen-1-ol20Fatty, oily, sweet and fruity–––0.060.06––––––––––143–08-81-Nonanol45.5Waxy, citrus, rue aldehydic, orange, oily, fatty and spicy–––––0.300.05–0.050.000.010.08––0.0860–12-8Phenylethyl Alcohol564.23floral, sweet, rosey and bready––0.01––0.01––0.010.000.000.040.000.000.05128–37-0Butylated Hydroxytoluene1000Mild, phenolic and camphor0.070.040.250.030.090.190.140.120.110.050.050.140.080.100.11105–05-51,4-Diethylbenzene2.1––––––––––––0.460.380.90496–11-7Indane100.050.07––0.10–0.51––0.080.07–0.050.06–91–57-62-Methylnaphthalene3Sweet, floral, woody, oily and aromatic0.350.40–0.100.11–0.110.260.23––0.450.650.160.51607–91-0Myristicin88spicy warm balsamic and woody0.030.016.300.020.005.670.040.134.300.030.016.870.050.057.95123–35-3β-Myrcene1.2Peppery, terpene, spicy and balsam––––––––2.87––7.93–––100–42-5Styrene65Sweet, balsam and floral0.200.130.370.070.230.250.250.210.140.220.090.43––0.0487–44-5β-Caryophyllene64Spicy, clove, woody, nut and peppery––––––––0.23––0.12––0.101139-30-6Caryophyllene oxide410Sweet, fresh, dry, woody and spicy––0.04––0.05––0.02–––––0.0797–53-0Eugenol470Sweet, warm spicy, clove and woody0.010.010.040.000.000.030.000.010.020.000.000.050.010.010.085989-33-3(Z)-linalool oxide (furanoid)100Earthy, floral, sweet and woody–––––0.36–––––0.69––0.5534,995–77-2(E)-Linalool oxide (furanoid)190floral––0.15–––––0.20–––––– OAV values of VFCs (OAV > 0.05) in snakehead with different treatments during 4 °C cold storage. Heatmap of VFCs with OAV > 0.05 (A); VIP analysis of VFCs (B). Octanal and nonanal exhibited grassy, fatty, and citrus notes at low concentrations; however, octanal functioned as a fishy off-odor compound, manifesting greasy notes at elevated concentrations (Dai et al., 2024). During initial storage, octanal and nonanal displayed OAVs of 11.60 and 46.60 respectively in UFF. By mid-storage, octanal progressively decreased in UFF, while PEE showed an initial increase followed by decline. Notably, OAV peak of nonanal in PEE was delayed until day 9, indicating that PLEE effectively inhibited lipid oxidation-induced accumulation of these aldehydes (P < 0.05). Dodecanal (Maria Emília Araújo & de Oliveira, 2019), characterized by fatty/waxy notes, transiently emerged in UFF during early storage before disappearing, whereas PEE only detected this compound at storage termination, a progression pattern analogous to that of octanal. In contrast, acetophenone generated negative sensory attributes: its metallic/rancid odor correlated with “metallic” off-flavors in fishiness and contributed to musty/grassy off-odors characteristic of freshwater fish (Liu, Zhao, et al., 2024; Zhang et al., 2023). This compound originated from phenylalanine degradation or lipid oxidation byproducts (Dang et al., 2024; Liu, Wen, et al., 2024), identified exclusively in UFF and PEE groups. It reached its peak OAV (0.07) in UFF on day 6. (E)-2-Octen-1-ol (grassy/plant-like odor), a typical fishiness component, was only detected on day 3 in both groups. Its generation linked to polyunsaturated fatty acid oxidation or Pseudomonas metabolism. dl-Perillaldehyde emerged as a compound unique to PEE, imparting woody-citrus notes. (E)-Linalool oxide (furanoid), exclusive to PEE, contributed fresh floral nuances while inhibiting fishiness perception. Its content declined progressively during storage, resulting from linalool oxidation or microbial metabolism (Li et al., 2025), indicating positive contributions to the overall flavor of refrigerated snakehead fillets. Indane attained its maximum OAV (0.51) on day 6 in UFF and PEE. While its faint aromaticity minimally impacted flavor individually, it may synergistically intensify off-odors when combined with other fishiness compounds (Peng et al., 2024). Formation pathways included aromatic amino acid degradation and lipid oxidation. 2-Methylnaphthalene (coal tar/camphor odor) showed an initial decline followed by an increase in UFF, peaking at 0.65 on day 12, 21.07% higher than PEE, and identified as a critical spoilage off-flavor compound (Liu et al., 2024). Microbiological analysis Microbial enumeration of snakehead fillets The dynamic changes in TVC are illustrated in Fig. 1(E). Across three sample groups, TVC exhibited a significant increasing trend (P < 0.05) during refrigerated storage. In the initial storage phase (from day 0 to day 3), the rapid proliferation pattern of TVC corresponded with concurrent accumulation of TMA and TVB-N, validating microbial metabolism as the primary driver for generating spoilage products like TMA and TVB-N (Zhou et al., 2024). It is worth noting that at the initial time point, the TVC levels in UFF and SDW had reached 3.2–3.3 lg CFU/g, which was significantly higher than that in PEE (2.74 lg CFU/g; P < 0.05), indicating that PLEE effectively inactivated surface-adherent microorganisms (Erkan et al., 2008). During storage, the TVC in PEE remained significantly lower (P < 0.05) than in UFF at all time points. This antimicrobial efficacy stemmed from multi-target inhibitory effects of PLEE: bioactive compounds disrupted microbial membrane integrity, inducing cytoplasmic leakage, while simultaneously inhibiting key energy-metabolizing enzymes such as ATP synthase (Zhou et al., 2024). Consequently, PLEE substantially inhibited microbial proliferation in snakehead fillets, thereby delaying flavor deterioration during refrigerated storage. Microbiota diversity test According to high-throughput sequencing results at the phylum level (Fig. 3A), Pseudomonadota abundance exhibited a declining trend during refrigerated storage, yet remained the dominant microbiota. Pseudomonas species (Pseudomonas fluorescens, P. fragi), as psychrotrophs (adapting to −15-20 °C), accelerated spoilage by proteolytic and lipolytic activities via secreted proteases/lipases, generating ammonia, sulfides (H₂S), trimethylamine (fishy off-odor), and volatile off-flavor aldehydes/ketones. Concurrently, Bacillota abundance increased significantly over storage time, with members like Brochothrix producing hydrogen sulfide and amines via protein degradation. This phylum, alongside Lactobacillus, collectively contributed to slime formation and spoilage off-odors. Remarkably, PEE displayed the highest microbial diversity during initial storage (day 0 to day 3), where Chloroflexota and Acidobacteriota abundances ranked second only to Pseudomonadota, indicating PEE specifically inhibited Bacillota proliferation.Fig. 3The change in Relative abundance on Phylum level (A), Community heatmap analysis on Genus level (B) and Relative abundance on genus level (C) of fillets during 4°C cold storage; LEfSe multi-level species hierarchical tree diagram (D); LDA discrimination results on genus level (E); The top 10 significant inter-group difference bacterial genera (F).Fig. 3 The change in Relative abundance on Phylum level (A), Community heatmap analysis on Genus level (B) and Relative abundance on genus level (C) of fillets during 4°C cold storage; LEfSe multi-level species hierarchical tree diagram (D); LDA discrimination results on genus level (E); The top 10 significant inter-group difference bacterial genera (F). Analysis of the genus-level microbial heatmap (Fig. 3B) and relative abundance plot (Fig. 3C) revealed that microbial diversity declined in both UFF and PEE during refrigerated storage. In fresh fillets (UFF_0d), the dominant genera included Sphingomonas (32.60%), Acinetobacter (23.68%), and the Burkholderia-Caballeronia-Paraburkholderia (BCP-complex) (8.14%). As for day 3, the relative abundance of these genera exhibited dramatic reductions: Sphingomonas decreased to 0.27% (UFF) and 3.88% (PEE), while Acinetobacter declined to 1.06% (UFF) and 5.34% (PEE). Concurrently, Pseudomonas emerged as the predominant spoilage genus, its relative abundance surging from 2.25% to 80.56% (UFF) and 70.56% (PEE). At storage endpoint, Pseudomonas abundance in UFF remained 11.23% higher than in PEE. Lactococcus showed significant growth in both groups, though its proliferation rate was markedly lower in PEE. Conversely, Carnobacterium proliferated preferentially in PEE, exceeding UFF by 7.82% at endpoint. These findings confirmed that PLEE effectively inhibited key spoilage genera (Pseudomonas, Lactococcus, Rouxiella). The marked inhibition of Pseudomonas and Lactococcus by PLEE directly correlated with the reduced accumulation of TMA (Section 3.1.3) and key off-flavor aldehydes (octanal, nonanal) observed in the PEE group (Section 3.2.2). These findings indicate that PLEE not only lowers the absolute abundance of spoilage bacteria but also reshapes the microbial community structure in a way that suppresses the metabolic production of volatile off-flavors. LEfSe analysis identified 41 differential microbial taxa (Fig. 3 DE), including 12 genera: Sphingomonas, BCP-complex, Methylobacterium, Pseudomonas, Chryseobacterium, Bacillus, Enterococcus, Rouxiella, Carnobacterium, Aeromonas, Lactococcus, and Janthinobacterium. Kruskal-Wallis H tests (P < 0.05) further confirmed significant inter-group differences for the top 10 genera (Fig. 3F): Pseudomonas > Lactococcus > Sphingomonas > Carnobacterium > Rouxiella > BCP-complex > Brochothrix > Aeromonas > Bacillus > Brevundimonas, of which eight genera were differential taxa identified by LEfSe, and these eight bacterial genera were defined as key differentially abundant microorganisms. Pseudomonas, Lactococcus, Carnobacterium, Rouxiella, and Aeromonas functioned as core spoilage genera. These microbiota influenced food spoilage rates and sensory attributes through diverse metabolic pathways, contributing to spoilage complexity and heterogeneity. PLEE inhibited growth of key spoilage genera (Pseudomonas, Lactococcus, Rouxiella), modulated microbial diversity, and restructured dominant flora by promoting competitive of Carnobacterium occupation, thereby delaying spoilage in refrigerated snakehead fillets. Lipidomic profile Non-targeted lipidomics analysis of refrigerated snakehead fillets revealed the process of lipid deterioration, identifying 1913 lipid molecules across all samples, including glycerophospholipids (GP), sphingolipids (SP), fatty acyls (FA), glycerolipids (GL), and sterol lipids (ST) (Table 6). Clustering heatmaps of the top 50 most abundant lipids (Fig. 4A) indicated triglycerides (TG) and phosphatidylcholine (PC) as the predominant lipid classes. TG, the primary storage lipid in fish muscle, exhibited the highest initial abundance (Xu et al., 2025). For day 12, TG abundance in UFF decreased relative to fresh fillets due to hydrolytic and oxidative losses during refrigeration, whereas PEE showed significantly elevated TG levels. This preservation correlated with the radical-scavenging activity of perilla polyphenols (rosmarinic acid), which inhibited lipid oxidation and maintained TG stability (Lu et al., 2017).Table 6The variation in lipid content in snakehead fillets with different treatments during 4 °C cold storage.CategoryClassUFF-0dSDW-0dPEE-0dUFF-3dSDW-3dPEE-3dUFF-6dSDW-6dPEE-6dUFF-12dSDW-12dPEE-12dGPBiotinylPE12.21 ± 0.28Ba12.21 ± 0.28ABa12.21 ± 0.28Aa12.22 ± 0.13Ba12.83 ± 0.15Ba12.93 ± 1.03Aa11.08 ± 0.18Aa12.05 ± 0.43Ab12.56 ± 0.33Ab11.44 ± 0.63ABa12.00 ± 0.04Aa12.34 ± 0.51AaBisMePA21.4 ± 0.44Aa21.4 ± 0.44Aa21.4 ± 0.44Aa21.9 ± 0.31Aa21.81 ± 0.21Aa22.47 ± 0.49Ba23.75 ± 0.74Bb22.59 ± 0.08Ba22.61 ± 0.18Ba25.5 ± 0.22Cb25.33 ± 0.29Cb24.25 ± 0.32CaCL1363.33 ± 4.57Ba1363.33 ± 4.57Ba1363.33 ± 4.57Ba1364.94 ± 5.6Ba1359.16 ± 5.69Ba1370.59 ± 2.27Ba1310.49 ± 23.9Aa1353.32 ± 14.91ABab1355.61 ± 10.75ABb1290.63 ± 23.67Aa1325.17 ± 14.94Aab1344.55 ± 7.62AbDLCL49.4 ± 0.57Aa49.4 ± 0.57Aa49.4 ± 0.57Aa49.47 ± 0.59Aa47.64 ± 0.54Aa53 ± 0.98Ab71.3 ± 3.08Ba55.61 ± 9.25ABa64.57 ± 7.02Ba69.15 ± 3.02Ba67.19 ± 0.99Ba66.73 ± 0.43BadMePE353.68 ± 0.64Ca353.68 ± 0.64Ba353.68 ± 0.64BCa350.73 ± 0.17BCa351.26 ± 0.65Ba355.95 ± 0.52Cb340.11 ± 7.26ABa351.68 ± 1.6Ba349.01 ± 3.38Aa338.45 ± 5.14Aa342.78 ± 2.85Aab350.83 ± 0.73ABbLdMePE92.57 ± 4.06Aa92.57 ± 4.06Aa92.57 ± 4.06Aa91.23 ± 4.21Aa87.73 ± 0.66Aa97.97 ± 1.5Ab116.07 ± 2.76Ba105.45 ± 7.89Ba110.6 ± 3.77Ba118.96 ± 1.68Bb114.09 ± 1.24Ba113.89 ± 1.32BaLPC268.69 ± 4.35Aa268.69 ± 4.35Aa268.69 ± 4.35Aa268.75 ± 3.38Aa267.65 ± 4.8Aa277.06 ± 4.12Aa308.75 ± 11.46Ba288.83 ± 12.78Ba295.23 ± 3.05Ba326.52 ± 4.62Cb316.82 ± 3.91Cab308.66 ± 4.54CaLPE173.23 ± 3.54Aa173.23 ± 3.54Aa173.23 ± 3.54Aa175.34 ± 1.83Aa172.82 ± 0.69Aa185.23 ± 3.56Bb201.88 ± 4.19Bb187.25 ± 6.59Ba189.03 ± 5.02Bab210.52 ± 2.29Cb207.36 ± 2.23Cab201.37 ± 3.86CaLPG11.64 ± 3.35Aa11.64 ± 3.35Aa11.64 ± 3.35Aa13.75 ± 2.8Aa9.94 ± 0.51Aa22.22 ± 1.62Bb27.25 ± 1.2Bb21.38 ± 2.92Ba25.15 ± 1.74Bab25.92 ± 1.27Ba24.38 ± 0.8Ba22.82 ± 2.29BaLPI19.47 ± 1.12Aa19.47 ± 1.12Aa19.47 ± 1.12Aa23.47 ± 2.39Aa24.96 ± 0.35ABa26.6 ± 5.56Aa43.44 ± 2.08Ba33.38 ± 8.5BCa40.48 ± 2.99Ba42.6 ± 2.47Ba39.47 ± 2.01Ca39.16 ± 1.09BaLPS5.52 ± 0.27Aa5.52 ± 0.27Aa5.52 ± 0.27Aa5.55 ± 0.12Aa5.87 ± 0.28Aab6.28 ± 0.25ABb14.02 ± 1.22Ba10.4 ± 4.37Aa13.17 ± 1.14Ca13.28 ± 1.22Ba8.62 ± 3.28Aa10.64 ± 3.3BCaMePC523.46 ± 7.08Aa523.46 ± 7.08Aa523.46 ± 7.08Aa514.65 ± 5.78Aa519.94 ± 2.81Aa550.09 ± 1.93Bb518.28 ± 13.7Aa523.41 ± 6.17Aab544.87 ± 8.98Bb517.46 ± 16.82Aa523.55 ± 18.29Aa542.99 ± 3.76BaMLCL177.97 ± 2.35Aa177.97 ± 2.35Aa177.97 ± 2.35Aa178.2 ± 1.44Ab170.97 ± 0.44Aa182.61 ± 2.79Ab212.43 ± 4.69Ba197.63 ± 12.87Ba205.01 ± 4.97Ba217.22 ± 2.9Ba212.57 ± 5.33Ba212.19 ± 2.15BaPA84.62 ± 0.89Aa84.62 ± 0.89Aa84.62 ± 0.89Aa84.92 ± 1.19Aa86.36 ± 3.68Aab91.27 ± 1.09Bb82.53 ± 1.38Aa84.38 ± 0.85Aa91.08 ± 1.28Bb82.8 ± 1.4Aa84.6 ± 0.86Aa90.93 ± 1.43BbPC2136.16 ± 18.49Aa2136.16 ± 18.49Aa2136.16 ± 18.49Aa2128.06 ± 3.5Aa2137.49 ± 11.58Aa2166.48 ± 9.94Ab2107.64 ± 42.87Aa2139.02 ± 12.17Aa2133.35 ± 18.35Aa2138.26 ± 19.43Aa2140.46 ± 36.61Aa2150.96 ± 2.1AaPE2244.22 ± 10.07Aa2244.22 ± 10.07Aa2244.22 ± 10.07Aa2244.91 ± 5.69Aa2244.88 ± 11.79Aa2268.02 ± 5.16Bb2254.29 ± 11.55Aa2248.98 ± 11.65Aa2269.74 ± 0.96Ba2251.65 ± 4.5Aa2255.94 ± 13.13Aa2274.82 ± 10.2BaPEt82.92 ± 0.86Aa82.92 ± 0.86Aa82.92 ± 0.86Aa84.42 ± 0.39Ab83.96 ± 1.3Aab82.28 ± 0.45Aa86.36 ± 0.94Ba85.06 ± 0.51Aa86.49 ± 1.48Ba88.65 ± 0.3Ca90.15 ± 0.84Ba89.21 ± 1.08CaPG175.98 ± 5.72Aa175.98 ± 5.72Aa175.98 ± 5.72Aa177.12 ± 3.32Aa176.11 ± 1.81Aa179.17 ± 1.24ABa182.51 ± 4.22Aa178.72 ± 7.58Aa185.68 ± 2.92Ba180.43 ± 6.7Aa176.68 ± 6.45Aa185.42 ± 3.08ABaphSM22.7 ± 0.1Ba22.7 ± 0.1Ba22.7 ± 0.1ABa22.67 ± 0.12Ba22.72 ± 0.13Ba22.77 ± 0.07Ba21.71 ± 0.29Aa22.51 ± 0.36ABb22.4 ± 0.22Aab21.79 ± 0.11Aa22.1 ± 0.18Aa22.49 ± 0.1ABbPI236.27 ± 0.79ABa236.27 ± 0.79Aa236.27 ± 0.79Aa238.92 ± 1.2Ba240.61 ± 2.53Ba238.7 ± 0.25Aa236.12 ± 2.45ABa236.52 ± 0.53Aa238.27 ± 1.5Aa233.98 ± 2.15Aa235.93 ± 1.19Aa235.8 ± 2.76AaPIP220.16 ± 0.55Aa20.16 ± 0.55Aa20.16 ± 0.55Aa20.5 ± 0.11Aa20.9 ± 0.42Aa20.35 ± 0.09Aa20.1 ± 0.43Aa20.34 ± 0.18Aa20.15 ± 0.36Aa19.85 ± 0.31Aa20.23 ± 0.44Aa19.78 ± 0.04AaPMe66.32 ± 0.09ABa66.32 ± 0.09Aa66.32 ± 0.09Aa65.52 ± 0.54Aa66.02 ± 0.44Aa67.5 ± 0.72Ab67.25 ± 0.54BCa67.77 ± 0.32Ba72.5 ± 2.34Bb67.68 ± 0.18Ca68.32 ± 0.5Ba69.32 ± 2.09ABaPS539.98 ± 1.67Ba539.98 ± 1.67Ba539.98 ± 1.67Ba540.16 ± 1.92Ba544.15 ± 2.65Ba540.99 ± 0.33Ba521.73 ± 5.24Aa536.09 ± 6.29ABb534.06 ± 4.32ABab517.05 ± 1.84Aa525.05 ± 5.26Aab531.55 ± 3.9AbSPCer729.37 ± 4.73Aa729.37 ± 4.73Aa729.37 ± 4.73Aa734.46 ± 0.49ABa736.41 ± 9.27Aa752.52 ± 4.38Bb735.98 ± 4.74ABa733.08 ± 4.87Aa751.81 ± 7.87Bb744.93 ± 8.35ABa744.57 ± 11.2Aa756.4 ± 3.44BaCerG2GNAc111.63 ± 0.09Aa11.63 ± 0.09Aa11.63 ± 0.09Aa11.68 ± 0.07Aa11.73 ± 0.09Aa11.65 ± 0.03Aa14.88 ± 0.37Ca12.66 ± 1.66Aa13.56 ± 1.67Ba13.9 ± 0.47Ba12.86 ± 0.71Aa12.95 ± 0.31ABaCerP6.79 ± 0.59Aa6.79 ± 0.59Aa6.79 ± 0.59Aa7.03 ± 0.11Aa7.06 ± 0.08Aa7.1 ± 0.02Aa6.74 ± 0.12Aa6.99 ± 0.12Aa6.9 ± 0.14Aa6.69 ± 0.63Aa6.77 ± 0.05Aa6.95 ± 0.05AaGD1a6.94 ± 0.09Ba6.94 ± 0.09Ba6.94 ± 0.09Aa6.86 ± 0.02Ba6.8 ± 0.06ABa6.59 ± 0.5Aa6.82 ± 0.03Ba6.84 ± 0.04ABa6.8 ± 0.03Aa6.51 ± 0.16Aa6.74 ± 0.12Aa6.63 ± 0.09AaGM317.37 ± 1.66Aa17.37 ± 1.66Aa17.37 ± 1.66Aa17.54 ± 0.64Aa17.27 ± 0.93Aa16.71 ± 0.76Aa17.2 ± 0.82Aa18.47 ± 0.89Aa17.68 ± 0.98Aa17.19 ± 0.45Aa17.73 ± 0.87Aa17.59 ± 0.72AaGT37.81 ± 0.08Ba7.81 ± 0.08ABa7.81 ± 0.08Ba7.82 ± 0.1Ba7.88 ± 0.09Ba7.74 ± 0.04ABa7.62 ± 0.06Aa7.82 ± 0.12ABb7.68 ± 0.03Aab7.54 ± 0.08Aa7.69 ± 0.03Ab7.67 ± 0.06AbHex1Cer266.63 ± 3.29Aa266.63 ± 3.29Aa266.63 ± 3.29Aa267.06 ± 1.03Aa267.2 ± 2.31Aa272.79 ± 0.63Bb262.04 ± 2.36Aa266.55 ± 2.48Aa266.03 ± 2.41Aa265.58 ± 3.65Aab262.12 ± 3.19Aa270.98 ± 0.76BbHex2Cer229.61 ± 1.65Ba229.61 ± 1.65Aa229.61 ± 1.65Aa230.74 ± 1.93Ba230.79 ± 3.61Aa235.68 ± 1.69Ba223.8 ± 2.07Aa228.62 ± 3.9Aab230.58 ± 3.5Ab222.53 ± 4.5Aa225.36 ± 3.88Aab231.66 ± 1.76ABbSM491.18 ± 7.08ABa491.18 ± 7.08Aa491.18 ± 7.08Aa491.4 ± 1.28ABab488.33 ± 5.31Aa496.77 ± 2.87Ab486.38 ± 3.1Aa488.37 ± 2.44Aa490.19 ± 3.67Aa494.9 ± 2.57Ba489.92 ± 3.31Aa491.23 ± 2.75AaSPH87.95 ± 0.22Ca87.95 ± 0.22Aa87.95 ± 0.22Ba87.55 ± 0.26Cab86.28 ± 1.28Aa88.75 ± 0.73Bb85.56 ± 0.29Aa86.06 ± 1.47Aa86.08 ± 0.91Aa86.5 ± 0.59Ba85.77 ± 1.41Aa85.64 ± 0.35AaSPHP7.93 ± 0.13Aa7.93 ± 0.13Aa7.93 ± 0.13Aa7.88 ± 0.02Aa7.86 ± 0.03Aa7.96 ± 0.03Ab7.86 ± 0.14Aa7.9 ± 0.14Aa7.85 ± 0.14Aa7.77 ± 0.29Aa7.82 ± 0.2Aa7.79 ± 0.14AaST15.53 ± 0.38Aa15.53 ± 0.38Aa15.53 ± 0.38ABa15.61 ± 0.33Aa15.85 ± 0.6Aa15.74 ± 0.39ABa14.92 ± 1.5Aa15.81 ± 0.54Aa15.06 ± 0.58Aa16.43 ± 0.62Ab15.43 ± 0.4Aa16.05 ± 0.19BabFAAcCa90.2 ± 2.62Aa90.2 ± 2.62Aa90.2 ± 2.62Aa88.44 ± 3.35Aa89.34 ± 3.66Aa88.86 ± 1.83Aa85.2 ± 3.85Aa87.76 ± 5.56Aa88.26 ± 1.79Aa90.94 ± 3.81Aa89.31 ± 4.29Aa91.41 ± 2.52AaAEA14.48 ± 0.05Aa14.48 ± 0.05Aa14.48 ± 0.05Aa14.81 ± 0.18ABa14.76 ± 0.11Aa14.83 ± 0.18Ba15.24 ± 0.16Bb14.91 ± 0.18Aab14.74 ± 0.2ABa15.91 ± 0.42Ca15.58 ± 0.63Ba14.99 ± 0.15BaCo24.81 ± 0.49Aa24.81 ± 0.49Aa24.81 ± 0.49Aa24.46 ± 0.49Aa24.65 ± 0.44Aa24.61 ± 0.16Aa24.11 ± 0.92Aa24.83 ± 0.43Aa24.29 ± 0.38Aa24.74 ± 1.32Aa24.86 ± 0.68Aa24.2 ± 0.52AaFA45.73 ± 0.75Aa45.73 ± 0.75Aa45.73 ± 0.75Aa48.79 ± 0.52Bab47.74 ± 1.8ABa50.14 ± 0.53Bb51.84 ± 1.6Cb48.39 ± 1.18Ba52.36 ± 1.4Cb53.51 ± 0.83Ca53.09 ± 0.97Ca54.5 ± 0.42DaOAHFA72.28 ± 2.46Aa72.28 ± 2.46Aa72.28 ± 2.46Aa80.06 ± 1.01Ba75.85 ± 5.34Aa75.5 ± 0.72Aa87.57 ± 2.51Cb79.58 ± 4.89Aa83.39 ± 2.86Bab90.08 ± 0.72Cb88.48 ± 1.48Bab87.57 ± 1.13CaWE8.49 ± 0.31Aa8.49 ± 0.31Aa8.49 ± 0.31ABa8.39 ± 0.16Aa8.51 ± 0.28Aa8.63 ± 0.17Ba8.08 ± 0.21Aa8.44 ± 0.28Aa8.11 ± 0.22Aa8.42 ± 0.06Aa8.41 ± 0.37Aa8.28 ± 0.19ABaGLDG962.36 ± 10.41Aa962.36 ± 10.41Aa962.36 ± 10.41ABa969.31 ± 9Aa962.19 ± 8.21Aa953.62 ± 9.2Aa978.09 ± 16.92Aa962.03 ± 16.53Aa982.88 ± 17.1BCa971.07 ± 28.58Aa978.94 ± 30.36Aa987.38 ± 6.08CaDGMG6.33 ± 0.05Ba6.33 ± 0.05Ba6.33 ± 0.05Aa6.36 ± 0.04Ba6.39 ± 0.05Ba14.45 ± 0.39Cb6.11 ± 0.07Aa6.32 ± 0.08ABa14.18 ± 0.42Cb6.03 ± 0.04Aa6.2 ± 0.07Aa13.3 ± 0.62BbMGDG100.59 ± 0.68Ba100.59 ± 0.68Aa100.59 ± 0.68Aa101.77 ± 0.42Ba101.15 ± 1.25Aa100.68 ± 1.14Aa97.68 ± 2.06Aa100.36 ± 1.12Aa100.65 ± 1.47Aa97.89 ± 1.33Aa99.19 ± 1.05Aab101.1 ± 0.59AbMGMG0.03 ± 0Aa0.03 ± 0ABa0.03 ± 0Aa0.03 ± 0Aa0.03 ± 0Ba6.8 ± 0.21Bb0.02 ± 0Aa0.03 ± 0ABa6.69 ± 0.24Bb0.03 ± 0.02Aa0.02 ± 0Aa6.53 ± 0.29BbTG2740.01 ± 29.29Aa2740.01 ± 29.29Aa2740.01 ± 29.29ABa2789.81 ± 28.68Ab2766.22 ± 23.9Aab2709.21 ± 53.15Aa2805.64 ± 25.37Aa2728.58 ± 68.32Aa2761.05 ± 61.74ABa2736.81 ± 77.32Aa2717.06 ± 92.98Aa2801.49 ± 3.83BaMG62.06 ± 0.67Aa62.06 ± 0.67Aa62.06 ± 0.67Aa62.23 ± 0.88ABab61.48 ± 0.66Aa63.61 ± 1.05Ab64.04 ± 1.04ABab62.74 ± 0.97Aa65.88 ± 1.41Bb64.63 ± 1.92Ba65.82 ± 0.97Ba66.22 ± 0.96BaSTChE40.28 ± 2.49Aa40.28 ± 2.49Aa40.28 ± 2.49Aa40.42 ± 3.01Aa41.69 ± 0.48Aab45.23 ± 1.38BCb44.99 ± 7.08Aa42.67 ± 2.88Aa43.19 ± 2.82ABa50.24 ± 8.15Aa50.91 ± 4.81Ba47.84 ± 0.7CaCmE22.09 ± 0.7Aa22.09 ± 0.7Aa22.09 ± 0.7Aa22.25 ± 0.11Aa22.31 ± 0.28Aa22.99 ± 0.3ABb22.08 ± 0.42Aa22.5 ± 0.38Aa22.89 ± 0.58ABa22.67 ± 0.49Aa22.71 ± 0.46Aa23.21 ± 0.46BaZyE25.25 ± 0.38Aa25.25 ± 0.38Aa25.25 ± 0.38Aa24.66 ± 0.78Aa24.93 ± 0.31Aa24.96 ± 0.23Aa24.34 ± 1.11Aa25.14 ± 0.35Aa25.14 ± 0.21Aa25.54 ± 0.56Aa25.58 ± 0.4Aa25.49 ± 0.48AaNote: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05).Fig. 4The change in Lipid heatmap analysis (A) and Relative abundance on Lipid (B) and VIP value and heatmap of lipid (C) of fillets during 4 °C cold storage.Fig. 4 Table 6The variation in lipid content in snakehead fillets with different treatments during 4 °C cold storage.CategoryClassUFF-0dSDW-0dPEE-0dUFF-3dSDW-3dPEE-3dUFF-6dSDW-6dPEE-6dUFF-12dSDW-12dPEE-12dGPBiotinylPE12.21 ± 0.28Ba12.21 ± 0.28ABa12.21 ± 0.28Aa12.22 ± 0.13Ba12.83 ± 0.15Ba12.93 ± 1.03Aa11.08 ± 0.18Aa12.05 ± 0.43Ab12.56 ± 0.33Ab11.44 ± 0.63ABa12.00 ± 0.04Aa12.34 ± 0.51AaBisMePA21.4 ± 0.44Aa21.4 ± 0.44Aa21.4 ± 0.44Aa21.9 ± 0.31Aa21.81 ± 0.21Aa22.47 ± 0.49Ba23.75 ± 0.74Bb22.59 ± 0.08Ba22.61 ± 0.18Ba25.5 ± 0.22Cb25.33 ± 0.29Cb24.25 ± 0.32CaCL1363.33 ± 4.57Ba1363.33 ± 4.57Ba1363.33 ± 4.57Ba1364.94 ± 5.6Ba1359.16 ± 5.69Ba1370.59 ± 2.27Ba1310.49 ± 23.9Aa1353.32 ± 14.91ABab1355.61 ± 10.75ABb1290.63 ± 23.67Aa1325.17 ± 14.94Aab1344.55 ± 7.62AbDLCL49.4 ± 0.57Aa49.4 ± 0.57Aa49.4 ± 0.57Aa49.47 ± 0.59Aa47.64 ± 0.54Aa53 ± 0.98Ab71.3 ± 3.08Ba55.61 ± 9.25ABa64.57 ± 7.02Ba69.15 ± 3.02Ba67.19 ± 0.99Ba66.73 ± 0.43BadMePE353.68 ± 0.64Ca353.68 ± 0.64Ba353.68 ± 0.64BCa350.73 ± 0.17BCa351.26 ± 0.65Ba355.95 ± 0.52Cb340.11 ± 7.26ABa351.68 ± 1.6Ba349.01 ± 3.38Aa338.45 ± 5.14Aa342.78 ± 2.85Aab350.83 ± 0.73ABbLdMePE92.57 ± 4.06Aa92.57 ± 4.06Aa92.57 ± 4.06Aa91.23 ± 4.21Aa87.73 ± 0.66Aa97.97 ± 1.5Ab116.07 ± 2.76Ba105.45 ± 7.89Ba110.6 ± 3.77Ba118.96 ± 1.68Bb114.09 ± 1.24Ba113.89 ± 1.32BaLPC268.69 ± 4.35Aa268.69 ± 4.35Aa268.69 ± 4.35Aa268.75 ± 3.38Aa267.65 ± 4.8Aa277.06 ± 4.12Aa308.75 ± 11.46Ba288.83 ± 12.78Ba295.23 ± 3.05Ba326.52 ± 4.62Cb316.82 ± 3.91Cab308.66 ± 4.54CaLPE173.23 ± 3.54Aa173.23 ± 3.54Aa173.23 ± 3.54Aa175.34 ± 1.83Aa172.82 ± 0.69Aa185.23 ± 3.56Bb201.88 ± 4.19Bb187.25 ± 6.59Ba189.03 ± 5.02Bab210.52 ± 2.29Cb207.36 ± 2.23Cab201.37 ± 3.86CaLPG11.64 ± 3.35Aa11.64 ± 3.35Aa11.64 ± 3.35Aa13.75 ± 2.8Aa9.94 ± 0.51Aa22.22 ± 1.62Bb27.25 ± 1.2Bb21.38 ± 2.92Ba25.15 ± 1.74Bab25.92 ± 1.27Ba24.38 ± 0.8Ba22.82 ± 2.29BaLPI19.47 ± 1.12Aa19.47 ± 1.12Aa19.47 ± 1.12Aa23.47 ± 2.39Aa24.96 ± 0.35ABa26.6 ± 5.56Aa43.44 ± 2.08Ba33.38 ± 8.5BCa40.48 ± 2.99Ba42.6 ± 2.47Ba39.47 ± 2.01Ca39.16 ± 1.09BaLPS5.52 ± 0.27Aa5.52 ± 0.27Aa5.52 ± 0.27Aa5.55 ± 0.12Aa5.87 ± 0.28Aab6.28 ± 0.25ABb14.02 ± 1.22Ba10.4 ± 4.37Aa13.17 ± 1.14Ca13.28 ± 1.22Ba8.62 ± 3.28Aa10.64 ± 3.3BCaMePC523.46 ± 7.08Aa523.46 ± 7.08Aa523.46 ± 7.08Aa514.65 ± 5.78Aa519.94 ± 2.81Aa550.09 ± 1.93Bb518.28 ± 13.7Aa523.41 ± 6.17Aab544.87 ± 8.98Bb517.46 ± 16.82Aa523.55 ± 18.29Aa542.99 ± 3.76BaMLCL177.97 ± 2.35Aa177.97 ± 2.35Aa177.97 ± 2.35Aa178.2 ± 1.44Ab170.97 ± 0.44Aa182.61 ± 2.79Ab212.43 ± 4.69Ba197.63 ± 12.87Ba205.01 ± 4.97Ba217.22 ± 2.9Ba212.57 ± 5.33Ba212.19 ± 2.15BaPA84.62 ± 0.89Aa84.62 ± 0.89Aa84.62 ± 0.89Aa84.92 ± 1.19Aa86.36 ± 3.68Aab91.27 ± 1.09Bb82.53 ± 1.38Aa84.38 ± 0.85Aa91.08 ± 1.28Bb82.8 ± 1.4Aa84.6 ± 0.86Aa90.93 ± 1.43BbPC2136.16 ± 18.49Aa2136.16 ± 18.49Aa2136.16 ± 18.49Aa2128.06 ± 3.5Aa2137.49 ± 11.58Aa2166.48 ± 9.94Ab2107.64 ± 42.87Aa2139.02 ± 12.17Aa2133.35 ± 18.35Aa2138.26 ± 19.43Aa2140.46 ± 36.61Aa2150.96 ± 2.1AaPE2244.22 ± 10.07Aa2244.22 ± 10.07Aa2244.22 ± 10.07Aa2244.91 ± 5.69Aa2244.88 ± 11.79Aa2268.02 ± 5.16Bb2254.29 ± 11.55Aa2248.98 ± 11.65Aa2269.74 ± 0.96Ba2251.65 ± 4.5Aa2255.94 ± 13.13Aa2274.82 ± 10.2BaPEt82.92 ± 0.86Aa82.92 ± 0.86Aa82.92 ± 0.86Aa84.42 ± 0.39Ab83.96 ± 1.3Aab82.28 ± 0.45Aa86.36 ± 0.94Ba85.06 ± 0.51Aa86.49 ± 1.48Ba88.65 ± 0.3Ca90.15 ± 0.84Ba89.21 ± 1.08CaPG175.98 ± 5.72Aa175.98 ± 5.72Aa175.98 ± 5.72Aa177.12 ± 3.32Aa176.11 ± 1.81Aa179.17 ± 1.24ABa182.51 ± 4.22Aa178.72 ± 7.58Aa185.68 ± 2.92Ba180.43 ± 6.7Aa176.68 ± 6.45Aa185.42 ± 3.08ABaphSM22.7 ± 0.1Ba22.7 ± 0.1Ba22.7 ± 0.1ABa22.67 ± 0.12Ba22.72 ± 0.13Ba22.77 ± 0.07Ba21.71 ± 0.29Aa22.51 ± 0.36ABb22.4 ± 0.22Aab21.79 ± 0.11Aa22.1 ± 0.18Aa22.49 ± 0.1ABbPI236.27 ± 0.79ABa236.27 ± 0.79Aa236.27 ± 0.79Aa238.92 ± 1.2Ba240.61 ± 2.53Ba238.7 ± 0.25Aa236.12 ± 2.45ABa236.52 ± 0.53Aa238.27 ± 1.5Aa233.98 ± 2.15Aa235.93 ± 1.19Aa235.8 ± 2.76AaPIP220.16 ± 0.55Aa20.16 ± 0.55Aa20.16 ± 0.55Aa20.5 ± 0.11Aa20.9 ± 0.42Aa20.35 ± 0.09Aa20.1 ± 0.43Aa20.34 ± 0.18Aa20.15 ± 0.36Aa19.85 ± 0.31Aa20.23 ± 0.44Aa19.78 ± 0.04AaPMe66.32 ± 0.09ABa66.32 ± 0.09Aa66.32 ± 0.09Aa65.52 ± 0.54Aa66.02 ± 0.44Aa67.5 ± 0.72Ab67.25 ± 0.54BCa67.77 ± 0.32Ba72.5 ± 2.34Bb67.68 ± 0.18Ca68.32 ± 0.5Ba69.32 ± 2.09ABaPS539.98 ± 1.67Ba539.98 ± 1.67Ba539.98 ± 1.67Ba540.16 ± 1.92Ba544.15 ± 2.65Ba540.99 ± 0.33Ba521.73 ± 5.24Aa536.09 ± 6.29ABb534.06 ± 4.32ABab517.05 ± 1.84Aa525.05 ± 5.26Aab531.55 ± 3.9AbSPCer729.37 ± 4.73Aa729.37 ± 4.73Aa729.37 ± 4.73Aa734.46 ± 0.49ABa736.41 ± 9.27Aa752.52 ± 4.38Bb735.98 ± 4.74ABa733.08 ± 4.87Aa751.81 ± 7.87Bb744.93 ± 8.35ABa744.57 ± 11.2Aa756.4 ± 3.44BaCerG2GNAc111.63 ± 0.09Aa11.63 ± 0.09Aa11.63 ± 0.09Aa11.68 ± 0.07Aa11.73 ± 0.09Aa11.65 ± 0.03Aa14.88 ± 0.37Ca12.66 ± 1.66Aa13.56 ± 1.67Ba13.9 ± 0.47Ba12.86 ± 0.71Aa12.95 ± 0.31ABaCerP6.79 ± 0.59Aa6.79 ± 0.59Aa6.79 ± 0.59Aa7.03 ± 0.11Aa7.06 ± 0.08Aa7.1 ± 0.02Aa6.74 ± 0.12Aa6.99 ± 0.12Aa6.9 ± 0.14Aa6.69 ± 0.63Aa6.77 ± 0.05Aa6.95 ± 0.05AaGD1a6.94 ± 0.09Ba6.94 ± 0.09Ba6.94 ± 0.09Aa6.86 ± 0.02Ba6.8 ± 0.06ABa6.59 ± 0.5Aa6.82 ± 0.03Ba6.84 ± 0.04ABa6.8 ± 0.03Aa6.51 ± 0.16Aa6.74 ± 0.12Aa6.63 ± 0.09AaGM317.37 ± 1.66Aa17.37 ± 1.66Aa17.37 ± 1.66Aa17.54 ± 0.64Aa17.27 ± 0.93Aa16.71 ± 0.76Aa17.2 ± 0.82Aa18.47 ± 0.89Aa17.68 ± 0.98Aa17.19 ± 0.45Aa17.73 ± 0.87Aa17.59 ± 0.72AaGT37.81 ± 0.08Ba7.81 ± 0.08ABa7.81 ± 0.08Ba7.82 ± 0.1Ba7.88 ± 0.09Ba7.74 ± 0.04ABa7.62 ± 0.06Aa7.82 ± 0.12ABb7.68 ± 0.03Aab7.54 ± 0.08Aa7.69 ± 0.03Ab7.67 ± 0.06AbHex1Cer266.63 ± 3.29Aa266.63 ± 3.29Aa266.63 ± 3.29Aa267.06 ± 1.03Aa267.2 ± 2.31Aa272.79 ± 0.63Bb262.04 ± 2.36Aa266.55 ± 2.48Aa266.03 ± 2.41Aa265.58 ± 3.65Aab262.12 ± 3.19Aa270.98 ± 0.76BbHex2Cer229.61 ± 1.65Ba229.61 ± 1.65Aa229.61 ± 1.65Aa230.74 ± 1.93Ba230.79 ± 3.61Aa235.68 ± 1.69Ba223.8 ± 2.07Aa228.62 ± 3.9Aab230.58 ± 3.5Ab222.53 ± 4.5Aa225.36 ± 3.88Aab231.66 ± 1.76ABbSM491.18 ± 7.08ABa491.18 ± 7.08Aa491.18 ± 7.08Aa491.4 ± 1.28ABab488.33 ± 5.31Aa496.77 ± 2.87Ab486.38 ± 3.1Aa488.37 ± 2.44Aa490.19 ± 3.67Aa494.9 ± 2.57Ba489.92 ± 3.31Aa491.23 ± 2.75AaSPH87.95 ± 0.22Ca87.95 ± 0.22Aa87.95 ± 0.22Ba87.55 ± 0.26Cab86.28 ± 1.28Aa88.75 ± 0.73Bb85.56 ± 0.29Aa86.06 ± 1.47Aa86.08 ± 0.91Aa86.5 ± 0.59Ba85.77 ± 1.41Aa85.64 ± 0.35AaSPHP7.93 ± 0.13Aa7.93 ± 0.13Aa7.93 ± 0.13Aa7.88 ± 0.02Aa7.86 ± 0.03Aa7.96 ± 0.03Ab7.86 ± 0.14Aa7.9 ± 0.14Aa7.85 ± 0.14Aa7.77 ± 0.29Aa7.82 ± 0.2Aa7.79 ± 0.14AaST15.53 ± 0.38Aa15.53 ± 0.38Aa15.53 ± 0.38ABa15.61 ± 0.33Aa15.85 ± 0.6Aa15.74 ± 0.39ABa14.92 ± 1.5Aa15.81 ± 0.54Aa15.06 ± 0.58Aa16.43 ± 0.62Ab15.43 ± 0.4Aa16.05 ± 0.19BabFAAcCa90.2 ± 2.62Aa90.2 ± 2.62Aa90.2 ± 2.62Aa88.44 ± 3.35Aa89.34 ± 3.66Aa88.86 ± 1.83Aa85.2 ± 3.85Aa87.76 ± 5.56Aa88.26 ± 1.79Aa90.94 ± 3.81Aa89.31 ± 4.29Aa91.41 ± 2.52AaAEA14.48 ± 0.05Aa14.48 ± 0.05Aa14.48 ± 0.05Aa14.81 ± 0.18ABa14.76 ± 0.11Aa14.83 ± 0.18Ba15.24 ± 0.16Bb14.91 ± 0.18Aab14.74 ± 0.2ABa15.91 ± 0.42Ca15.58 ± 0.63Ba14.99 ± 0.15BaCo24.81 ± 0.49Aa24.81 ± 0.49Aa24.81 ± 0.49Aa24.46 ± 0.49Aa24.65 ± 0.44Aa24.61 ± 0.16Aa24.11 ± 0.92Aa24.83 ± 0.43Aa24.29 ± 0.38Aa24.74 ± 1.32Aa24.86 ± 0.68Aa24.2 ± 0.52AaFA45.73 ± 0.75Aa45.73 ± 0.75Aa45.73 ± 0.75Aa48.79 ± 0.52Bab47.74 ± 1.8ABa50.14 ± 0.53Bb51.84 ± 1.6Cb48.39 ± 1.18Ba52.36 ± 1.4Cb53.51 ± 0.83Ca53.09 ± 0.97Ca54.5 ± 0.42DaOAHFA72.28 ± 2.46Aa72.28 ± 2.46Aa72.28 ± 2.46Aa80.06 ± 1.01Ba75.85 ± 5.34Aa75.5 ± 0.72Aa87.57 ± 2.51Cb79.58 ± 4.89Aa83.39 ± 2.86Bab90.08 ± 0.72Cb88.48 ± 1.48Bab87.57 ± 1.13CaWE8.49 ± 0.31Aa8.49 ± 0.31Aa8.49 ± 0.31ABa8.39 ± 0.16Aa8.51 ± 0.28Aa8.63 ± 0.17Ba8.08 ± 0.21Aa8.44 ± 0.28Aa8.11 ± 0.22Aa8.42 ± 0.06Aa8.41 ± 0.37Aa8.28 ± 0.19ABaGLDG962.36 ± 10.41Aa962.36 ± 10.41Aa962.36 ± 10.41ABa969.31 ± 9Aa962.19 ± 8.21Aa953.62 ± 9.2Aa978.09 ± 16.92Aa962.03 ± 16.53Aa982.88 ± 17.1BCa971.07 ± 28.58Aa978.94 ± 30.36Aa987.38 ± 6.08CaDGMG6.33 ± 0.05Ba6.33 ± 0.05Ba6.33 ± 0.05Aa6.36 ± 0.04Ba6.39 ± 0.05Ba14.45 ± 0.39Cb6.11 ± 0.07Aa6.32 ± 0.08ABa14.18 ± 0.42Cb6.03 ± 0.04Aa6.2 ± 0.07Aa13.3 ± 0.62BbMGDG100.59 ± 0.68Ba100.59 ± 0.68Aa100.59 ± 0.68Aa101.77 ± 0.42Ba101.15 ± 1.25Aa100.68 ± 1.14Aa97.68 ± 2.06Aa100.36 ± 1.12Aa100.65 ± 1.47Aa97.89 ± 1.33Aa99.19 ± 1.05Aab101.1 ± 0.59AbMGMG0.03 ± 0Aa0.03 ± 0ABa0.03 ± 0Aa0.03 ± 0Aa0.03 ± 0Ba6.8 ± 0.21Bb0.02 ± 0Aa0.03 ± 0ABa6.69 ± 0.24Bb0.03 ± 0.02Aa0.02 ± 0Aa6.53 ± 0.29BbTG2740.01 ± 29.29Aa2740.01 ± 29.29Aa2740.01 ± 29.29ABa2789.81 ± 28.68Ab2766.22 ± 23.9Aab2709.21 ± 53.15Aa2805.64 ± 25.37Aa2728.58 ± 68.32Aa2761.05 ± 61.74ABa2736.81 ± 77.32Aa2717.06 ± 92.98Aa2801.49 ± 3.83BaMG62.06 ± 0.67Aa62.06 ± 0.67Aa62.06 ± 0.67Aa62.23 ± 0.88ABab61.48 ± 0.66Aa63.61 ± 1.05Ab64.04 ± 1.04ABab62.74 ± 0.97Aa65.88 ± 1.41Bb64.63 ± 1.92Ba65.82 ± 0.97Ba66.22 ± 0.96BaSTChE40.28 ± 2.49Aa40.28 ± 2.49Aa40.28 ± 2.49Aa40.42 ± 3.01Aa41.69 ± 0.48Aab45.23 ± 1.38BCb44.99 ± 7.08Aa42.67 ± 2.88Aa43.19 ± 2.82ABa50.24 ± 8.15Aa50.91 ± 4.81Ba47.84 ± 0.7CaCmE22.09 ± 0.7Aa22.09 ± 0.7Aa22.09 ± 0.7Aa22.25 ± 0.11Aa22.31 ± 0.28Aa22.99 ± 0.3ABb22.08 ± 0.42Aa22.5 ± 0.38Aa22.89 ± 0.58ABa22.67 ± 0.49Aa22.71 ± 0.46Aa23.21 ± 0.46BaZyE25.25 ± 0.38Aa25.25 ± 0.38Aa25.25 ± 0.38Aa24.66 ± 0.78Aa24.93 ± 0.31Aa24.96 ± 0.23Aa24.34 ± 1.11Aa25.14 ± 0.35Aa25.14 ± 0.21Aa25.54 ± 0.56Aa25.58 ± 0.4Aa25.49 ± 0.48AaNote: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05). The variation in lipid content in snakehead fillets with different treatments during 4 °C cold storage. Note: The capital letters marked on each data point represent the differences at the 0.05 level for different storage times in the same processing group, and the lowercase letters represent the differences at the 0.05 level for the same storage time point in different processing groups. Different letters in the same column or row indicate significant difference (P < 0.05). The change in Lipid heatmap analysis (A) and Relative abundance on Lipid (B) and VIP value and heatmap of lipid (C) of fillets during 4 °C cold storage. One-way ANOVA identified ten differential lipids: LdMePE (20:1), MePC (11:0/20:3), PA (22:5/23:1), MePC (15:0/18:3), MLCL (14:2/18:2/18:2), LPG (16:0), DGMG (16:0), DLCL (23:1/22:6), LPG (18:1) and PG (16:1/18:3). Relative abundance plots (Fig. 4B) demonstrated lower fluctuations in PEE and elevated levels of MePC (11:0/20:3), PA (22:5/23:1), MePC (15:0/18:3), DGMG (16:0) and LPG (18:1) in PEE during early storage. This stabilization arose from PLEE antioxidants (rosmarinic acid, perillaldehyde) quenching free radicals to inhibit unsaturated fatty acid oxidation, thereby protecting PUFA-rich lipids (PA (22:5/23:1)). Concurrently, phospholipids containing medium/short-chain unsaturated fatty acids like MePC (11:0/20:3) were shielded from oxidation into hexanal and 2-nonenal (key fishy odorants), preventing early spoilage off-flavors. Screening with VIP > 1 and P < 0.05 yielded 17 key differential lipids, six of which overlapped with the ANOVA-derived list: LPG (18:1), PG (16:1/18:3), PA (22:5/23:1), DGMG (16:0), MePC (11:0/20:3) and MePC (15:0/18:3). VIP heatmaps (Fig. 4C) confirmed PLEE significantly enhanced PG (16:1/18:3), PA (22:5/23:1), DGMG (16:0), MePC (11:0/20:3) and MePC (15:0/18:3), evidencing its capacity to attenuate PUFA oxidation and preserve phospholipids (MePC) via radical scavenging. The preservation of these polyunsaturated-acyl-chain-containing lipids (PG, PA, MePC) by PLEE had direct consequences on flavor quality. As shown in Section 3.2.2, the PEE group exhibited significantly lower levels of fishy odorants such as 1-octen-3-ol, octanal, and dodecanal. These results collectively suggest that the antioxidant activity of PLEE protects susceptible phospholipids from oxidative degradation, thereby limiting the formation of volatile secondary oxidation products that cause off-flavors. Correlation analysis Correlation analysis of key microbial communities, lipids, and differential flavor compounds revealed significant associations among eight microbial genera, six lipid molecules, and eleven flavor substances (|ρ| > 0.4, Fig. 5A). Specifically, 1,4-diethylbenzene exhibited positive correlations with spoilage-related genera (Carnobacterium, Aeromonas, Lactococcus, and Rouxiella), indicating synergistic roles in flavor deterioration. Unique flavor markers in PEE, trans-furan linalool oxide and perillaldehyde, and the genus Bacillus, they were significantly positively correlated with lipid molecules containing polyunsaturated acyl chains, such as PG (16:1/18:3), PA (22:5/23:1), DGMG (16:0), MePC (11:0/20:3), MePC (15:0/18:3). As precursors to aromatic compounds, the preserved levels of these lipids indicated that PLEE effectively retarded unsaturated fatty acid degradation, thereby inhibiting flavor deterioration in snakehead fillets (Cao et al., 2025). Additionally, acetophenone (a secondary lipid oxidation product) negatively correlated with these lipids, further confirming the antagonistic relationship between lipid structural integrity and oxidation byproduct formation. Concurrently, LPG (18:1) exhibited negative correlations with trans-2-octen-1-ol, 1-octen-3-ol, heptanol, octanal, and dodecanal (all characteristic fishy odorants). The notably higher LPG (18:1) content in the PEE group demonstrated that PLEE inhibited phospholipid hydrolysis and subsequent oxidation pathways, reducing the generation of fishy aldehydes and alcohols.Fig. 5Correlation analysis and mechanism interpretation. Correlation network plot of differential lipids, Microbiota and key flavor substances (A); The inhibitory pathway of PLEE on flavor deterioration of snakehead fillets during cold storage (B).Fig. 5 Correlation analysis and mechanism interpretation. Correlation network plot of differential lipids, Microbiota and key flavor substances (A); The inhibitory pathway of PLEE on flavor deterioration of snakehead fillets during cold storage (B). Based on the established correlations and the inhibitory pattern of PLEE against flavor deterioration in refrigerated snakehead fillets, a triple synergistic inhibition pathway was elucidated (Fig. 5B). Primarily, PLEE modulated microbial communities by significantly inhibiting the growth and metabolic activities of spoilage bacteria (Aeromonas, Lactococcus, and Rouxiella), thereby reducing the accumulation of total volatile basic nitrogen (TVB-N) and trimethylamine (TMA). Concurrently, PLEE decelerated polyunsaturated fatty acid degradation while substantially diminishing the accumulation of lipid peroxides and secondary oxidation products (1-octen-3-ol), mitigating rancidity development. Additionally, characteristic terpenoids in PLEE, specifically trans-furan linalool oxide and perillaldehyde, enhanced sensory acceptance through olfactory interference. These pathway constituted a coordinated regulatory network: Key bioactive constituents (perillaldehyde) exhibited dual antimicrobial and antioxidant functions, disrupting the lipid oxidation-microbial spoilage positive feedback loop (notably lipase-mediated fatty acid release fueling microbial metabolism). Combined with odor masking and off-flavor inhibition, this network maintained flavor homeostasis in refrigerated fillets. Conclusions This study investigated the effects of PLEE on snakehead fillets utilizing microbial diversity and lipidomics analyses. Results indicated that PLEE treatment effectively delayed the quality deterioration in fillets, particularly mitigating flavor degradation. During storage, 95 volatile compounds were identified, with 11 key differential flavor compounds exhibiting significant correlations (|ρ| > 0.4) with specific microorganisms (Aeromonas, Lactococcus, Rouxiella) and lipid molecules. Characteristic flavor compounds in the PLEE group positively correlated with lipids containing polyunsaturated acyl chains, whereas typical spoilage genera showed positive correlations with the off-flavor marker 1,4-diethylbenzene. Additionally, the lipid oxidation product acetophenone inversely correlated with these lipids, and the lysophospholipid LPG (18:1) demonstrated significant negative correlations with characteristic fishy odorants. In summary, PLEE maintains flavor homeostasis through three pathway: (1) inhibiting proliferation and metabolic activities of spoilage microorganisms to retard off-flavor compound accumulation; (2) protection of lipids (PG (16:1/18:3), PA (22:5/23:1)) to inhibit phospholipid hydrolysis-derived fishy odorants and attenuate lipid oxidation; (3) masking undesirable flavors via characteristic terpenoids (trans-furan linalool oxide, perillaldehyde) in PLEE. These findings can provide a scientific guidance for mitigating flavor deterioration of refrigerated snakehead fillets. CRediT authorship contribution statement Jinlin Li: Supervision, Funding acquisition, Conceptualization. Yuyan Jin: Writing – original draft, Visualization, Conceptualization. Huimin Ma: Visualization. Bin Peng: Data curation. Bizheng Zhong: Data curation. Zongcai Tu: Supervision. Cheng-wei Yu: Writing – review & editing, Funding acquisition, Conceptualization. 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.