Transient high temperature treatment: Impacts on taste-aroma formation and mechanisms in dried tuna

Abstract

Roasting is critical for flavor development in dried fish. This study investigated the influence of Transient High Temperature Technology (THTT) on the flavor profile of dried tuna products. Findings observed through electronic nose, Gas Chromatography–Ion Mobility Spectrometry, and Gas Chromatography-Olfactometry-Mass Spectrometry demonstrated THTT-mediated increased content of volatile compounds in dried tuna. Variable Importance in Projection and Relative Odor Activity Values screening revealed the identification of 29 key aroma substances. Regarding taste, 5′-monophosphate increased from 295.83 to 605.78 μg/g. Free Amino Acids (FAAs) contributing to sweet and umami tastes were increased from 829.01 to 2226.44 μg/g, from 321.78 μg/g to 1730.60 μg/g, respectively (P < 0.05). The nucleotides and FAAs synergistically improved the taste quality of dried tuna products, as speculated by the partial least squares regression analysis, accompanied by correlation analysis. Meanwhile, the THTT treatment promotes the formation of characteristic volatile compounds. Therefore, the proposed THTT provides a novel technological approach for enhancing the flavor of dried tuna products.

Graphical abstract

Highlights

• •Transient High Temperature Treatment (THTT) enhanced the flavor of dried Tuna samples.
• •61 and 89 volatile compounds were identified in dried tuna using GC–IMS and GC–O-MS, respectively.
• •29 key differential volatile compounds were screened by combined VIP and ROAV analysis.
• •Nucleotides and free amino acids (FAAs) improved the flavor of dried tuna samples in a synergistic manner.

1. Introduction

Tuna, a high-protein, low-fat premium marine fish, exhibits substantial nutritional and commercial value and is widely consumed across the global market (Mccluney et al., 2019; Bodin et al., 2022). Processing tuna into dried fish products represents a viable strategy that enables efficient resource utilization while meeting market demands. However, conventional tuna processing techniques, characterized by prolonged heating durations, often lead to reduced product quality stability, flavor deterioration, and non-uniform coloration, hindering its industry development (Pitarch et al., 2021), Therefore, developing novel processing technologies for dried tuna products constitutes a critical pathway for advancing the industry toward value-added deep processing.

Traditional hot-air roasting is the primary processing method in the dried fish industry. Wen et al. (Wen-ting et al., 2023) found that prolonged hot-air roasting leads to slow generation of volatile substances and requires extended heating time. Additionally, extended hot-air drying can cause crust charring, and excessive surface browning may produce burnt flavors, bitter substances, and harmful compounds such as heterocyclic amines (Portanguen et al., 2014). Furthermore, B. Mora et al. (2011) confirmed that steam heat treatment cannot adequately form a browning crust, making it difficult to generate a rich roasting aroma.Transient High Temperature Treatment (THTT) refers to a thermal process that raises food materials to an ultra-high temperature range of 130–160 °C within an extremely short time (0–10 s), and then carries out short-term, intermittent, and precisely temperature-controlled heat treatment at this temperature. Typical thermal processing methods that conform to the characteristics of THTT include ultra-high temperature instantaneous heating (UHT) and microwave-assisted heating. (Drotleff et al., 2026; K. Fan et al., 2024; J. Li et al., 2026). THTT has been extensively applied in sterilization and cooking processes for liquid foods and dairy products, primarily due to the frequent heating of food to high temperatures (H. Ding et al., 2023; Gao et al., 2024). Upon comparison with the conventional heating approaches, THTT can effectively eliminate microorganisms and prolong product shelf life while retaining nutrient levels and flavor, leveraging the greater heat sensitivity of microbes compared to food constituents (B. Deng et al., 2024), Moreover, it has been reported (Song et al., 2023) that THTT can activate the Maillard reaction, altering the color of the product and forming distinctive flavors. Chiu-Chu Hwang et al. (2024) confirmed the THTT-induced denaturation of myoglobin, myofibrillar protein, as well as muscle contraction, through rapid high-temperature treatment, leading to enhanced sensory attributes in chicken meat, including improved texture, juiciness, flavor profile, and overall acceptability. A further study (X. W. Wang et al., 2023) demonstrated that THTT treatment enhanced the degree of lipid hydrolysis, providing a biochemical basis for the development of characteristic aromas in pork belly.

Furthermore, FAAs and nucleotides are considered the major contributors to fish taste, demonstrating a close relationship with umami, sweetness, and bitterness (Moerdijk-Poortvliet et al., 2021). Qin et al. (Q. Dan et al., 2022) identified protein degradation as an essential pathway for taste development within aquatic products. A further study (Xiaoying et al., 2022) confirmed the presence of increased FAAs content in the THTT-treated fish, positively influencing the retention and enhancement of flavor components; however, underlying mechanisms still need to be clarified. Therefore, a comprehensive understanding of THTT's impact on food components and the underlying relationship between changes in nutritional quality and flavor development in tuna remains a critical research focus.

Currently, there is a scarcity of research focusing on the formation of flavor-contributing substances and taste-active substances, and their regulatory mechanisms during the processing of dried fish products using THTT. Therefore, the current study is based on evaluating the influence of THTT on flavor development in dried tuna products. Electronic nose (E-nose), Gas Chromatography-Olfactometry-Mass Spectrometry (GC-O-MS), and Gas Chromatography–Ion Mobility Spectrometry (GC-IMS) analyses were utilized for analyzing key compounds contributing to the aroma of THTT-treated dried tuna. Relative Odor Activity Value (ROAV) and Variable Importance in Projection (VIP) values then enabled the determination of the compounds most responsible for flavor, clarifying the odor generation. Moreover, flavor nucleotides and FAAs were determined, and taste activity values (TAV) were analyzed to assess the effects of THTT on the taste attributes of dried tuna products. Correlation analysis was employed to elucidate the mechanisms by which THTT influences the flavor and taste-active compound formations in dried tuna, offering a novel approach for enhancing the overall product's flavor.

2. Materials and methods

2.1. Materials

The Zhejiang Rongchuang Food Industry Co., Ltd provided the Thunnus obesus samples and compound seasonings, used as curing ingredients. Auxiliary materials, including salt, white sugar, ginger, light and dark soy sauces, and cooking wine, were procured from local supermarkets. Nucleotide standards, including adenosine 5′-monophosphate (AMP), inosine 5′-monophosphate (IMP), hypoxanthine ribonucleoside (HxR), guanosine 5′-monophosphate (GMP), and hypoxanthine (Hx) of 98% purity, were from Tianjin Alta Technology Co., Ltd. (Tianjin, China). The 3-nonanone standard (purity: 99%) was from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan), and other analytical-grade chemicals were from Sinopharm Chemical Reagent Beijing Co., Ltd.

2.2. Samples

The production process of dried tuna is shown in Fig. 1.

Fig. 1.

Dried tuna product production process.

The experimental material used pre-rigor bigeye tuna meat, stored at −20 °C, and thawed naturally at 4 °C for 2 h before use, followed by removal of surface fascia and other impurities. After that, the meat was sliced with the dimensions of 6 × 2.4 × 1.2 cm (L×W×H, 18-20 g per piece, 15 tuna fish pieces prepared), thoroughly cleaned, and labeled as RAW-FM1 for subsequent processing. The meat was uniformly mixed with salt, white sugar, ginger, light and dark soy sauces, cooking wine, and water, with each ingredient added proportionally to the meat's mass, followed by curing at 4 °C for 14 h. Both cured fish and uncured fish samples werethen placed in anelectric thermostatic drying oven (Memmert GmbH + Co KG, Schwabach, Germany) preheated to 45 °C for low-temperature baking for 6 h; the marinated and low-temperature baked fish meat is recorded as LTB-FM2. Subsequently, both uncured and cured dried tuna samples were placed in a hot air rotary baking oven (Model ML-AH-5-R, Meletino, Foshan Nimida Machinery Equipment Co., Ltd., Foshan, China), heated to 130 °C within 8 s, and treated for 11 minand labeled as UM-THTT-FM3 (uncured dried tuna samples) and P-THTT-FM4 (cured dried tuna samples), respectively. After heat treatment, transfer the samples to a sterile stainless steel cooling rack and cool them with chilled forced air at an ambient temperature of 22 ± 2 °C for 10 min until the core temperature of the tuna meat drops to 25 °C, followed by storage at −80 °C until required. This research employed a total of four dried Thunnus obesus samples.

2.3. E-nose

The flavor profile variation in dried tuna was determined using a PEN3 electronic nose (WinMuster Airsense Analytics, Schwerin, Germany). Briefly, 3 g of tuna was transferred to a centrifuge tube (50 mL) and sealed, followed by insertion of the injection needle for direct headspace sampling. Each detection cycle lasted for 100 s, including a 5 s pre-injection period and a 100 s formal detection phase, with a 400 mL/min rate of carrier gas flow. Data arising from 69 to 71 s were collected for summary and analysis.

2.4. GC-IMS

Analysis was performed as previously described (Jingyi et al., 2023) with slight modifications, using a Flavour Spec1H1-00053 gas chromatography-ion mobility spectrometer (GAS, Germany) for the determination of volatile organic compounds (VOCs). Two grams of dried tuna in a headspace vial (20 mL) were maintained for 15 min at 60 °C with shaking at 500 rpm. Five hundred microliters of headspace gas were then injected using a syringe at 85 °C. A WAX-type chromatographic column (15 m, ID: 0.53 mm) was used at 60 °C, with 99.999% pure nitrogen as the carrier gas and the following conditions: 2 mL/min initial flow rate, kept constant for 2 min, raised to 10 mL/min within 8 min, with further elevation to 100 mL/min within the next 10 min and to 150 mL/min within another 10 min. The IMS temperature was kept constant at 45 °C, with a total runtime of 35 min, after which the flow was stopped. Qualitative and quantitative analyses were performed through GC×IMS Library Search. Data visualization, including 3D response spectra, 2D fingerprint maps and differential comparison spectra, was undertaken using the Reporter and Gallery Plot modules in VOCal, enabling comparative analysis of the sample's profile.

2.5. HS-SPME-GC-O-MS

The VOCs in dried tuna were analyzed as described (Haiyan et al., 2023), with slight modifications. Five grams of tuna was mixed with 20 μL of 3-nonanone (50 μg/mL) as the internal standard in a 20 mL headspace vial. The material was extracted for 30 min at 40 °C using a 50/30 μm DVB/CAR/PDMS fiber (Supelco, USA). Briefly, the sealed headspace vial was placed in a tri-mode autosampler (PAL RSI, CTC Analytics AG, Switzerland) and subjected to extraction in SPME mode using a 5-min incubation period, agitation at 450 rpm, and extraction for 30 min at 40 °C. The fiber was inserted to a depth of 18 mm and desorbed for 5 min.

The carrier gas was high-purity (99.999%) helium, flowing through a DB-5 MS chromatographic column (30 m × 0.25 mm×0.25 μm, TG-5 MS, Thermo Fisher Scientific, USA) at a 1.6 mL/min flow rate. The injector was adjusted to 250 °C, while the initial temperature was 40 °C for 1 min, raised to 100 °C at 2 °C/min with a 2 min hold, a further raise to 180 °C at 5 °C/min, maintained for 5 min, and a final elevation to 250 °C with a 4 min hold. The MS was performed using an electron impact ionization (EI) mode, with an ion source maintained at 280 °C, and an ionization energy of 70 eV. Full-scan mode data acquisition was undertaken over a mass range of 45∼650 amu.

Qualitative methods, quantitative methods, and Gas Chromatography-Olfactometry (GC-O) analysis. Please refer to Supplementary Material Appendix 1.

2.6. Relative odor activity values (ROAV) calculation

ROAVs can be classified as the ratio of the relative proportions of VOCs to their odor thresholds (L. Yan et al., 2024), with ROAV> 1 indicating their contributions to the overall flavor. The odor threshold data were obtained from the Flavor Compound Database of Shanghai Jiao Tong University (https://mffi.sjtu.edu.cn/database).

Furthermore, ROAVs can be determined using Equation (1)(Ma et al., 2024).

$$\begin{array}{c}ROAV=\frac{C_i/{OT}_i}{C_{\mathit{max}}/{OT}_{\mathit{max}}}\end{array}$$ (1)

Where C_i denotes the VOC contents (μg/kg), while OT_i denotes their odor thresholds in water (μg/kg). $C_{\mathit{max}}$ denotes the content of the VOC that contributes the most to the overall flavor (μg/kg), while ${OT}_{\mathit{max}}$ denotes its corresponding odor threshold in water (μg/kg).

2.7. 5′-nucleotide contents

Nucleotides were analyzed as described (Mingyu et al., 2022) with slight modifications. Briefly, 2 g of sample homogenate was mixed with 20 mL of 10% (v/v) HClO₄, vortexed for 1 min, and centrifuged (8000 rpm, 10 min, 4 °C). The supernatant was re-extracted using 10 mL of 5% (v/v) HClO₄. The pH of the combined supernatants was adjusted to 6.0∼6.4 using NaOH (1 mol/L), while the solution was made up to 50 mL with ultrapure water in a pre-cooled volumetric flask at 4 °C and centrifuged (8000 rpm, 10 min, 4 °C). The supernatant was filtered (0.22 μm microporous membrane) and kept at 4 °C for further analysis.

High-Performance Liquid Chromatography (HPLC) was performed using a Venusil XBP C₁₈ column (250 × 4.6 mm, 5 μm) at 35 °C. The mobile phase comprised a 1:1 (v/v) ratio of 0.02 mol/L KH₂PO₄ and K₂HPO₄・3H₂O, with pH 6.0 adjusted using phosphoric acid. The analysis was conducted using a detection wavelength of 254 nm, injection volume of 20 μL, and flow rate of 1.0 mL/min.

2.8. FAAs analysis and calculation of taste activity value (TAV)

A slightly modified method of Geng et al. (Jieting et al., 2018) was adopted for the analysis. Briefly, 5 g of the sample was homogenized with 25 mL of 10% (m/v) trichloroacetic acid, followed by centrifugation at 10 000 rpm for 10 min at 4 °C. The supernatant was filtered (0.22 μm) and subsequently injected into an automatic amino acid analyzer for quantification.

The taste contribution of FAAs was evaluated by calculating their TAV using Equation (2) (K. Qin et al., 2025) using the threshold of TAV >1 as representative of contribution to food taste, with values of TAV <1 regarded as having no significant impact.

$$\begin{array}{c}TAV=\frac{C}{T}\end{array}$$ (2)

Where C denotes the FAA concentration in the sample (μg/g), and T represents their corresponding taste threshold (μg/g).

2.9. Statistical analysis

Three replicates were used in all analyses. Significant differences among the samples were determined through the utilization of one-way analysis of variance (ANOVA) using SPSS 17.0 software. Presentations of radar charts and heat maps were generated using OriginPro 2021 (OriginLab, Northampton, MA). Variable importance in projection (VIP) values were determined using SIMCA 14.1 (Sartorius AG, Germany), with variables having a VIP value > 1 considered as significantly differential variables. The partial least squares regression (PLSR) model was established by The Unscrambler X 10.4 (Camo, Asa, Norway). Bubble pool plots were generated through the CNSknowall online platform.

3. Results and discussion

3.1. E-nose assessments

According to the data (Fig. 2A), the response signals of dried tuna samples on W5S, W6S, W1S, W1W, W2S, and W3S sensors demonstrated significant variations (p < 0.05), with all response values exceeding 1. The distinctive flavor of dried tuna arises from characteristic chemical compounds generated during the roasting process, such as nitrogen oxides, hydrides, alkanes, sulfur-containing compounds, alcohols, aromatic, and aliphatic compounds. Furthermore, these substances result from the pyrolysis and degradation of amino acids and proteins during heating, accompanied by the Maillard reaction occurring between amino acids and reducing sugars (Hao et al., 2025; Maria et al., 2021). THTT revealed significantly increased responses of the W1W, W1S, and W2S sensors. According to the selective recognition characteristics of the sensors, this process appears to induce the formation of sulfur compounds, alcohols, certain aromatic substances, and alkanes, with alcohols, aromatics, and alkanes demonstrating significant contribution to the fruity and nutty flavors, while the sulfur compounds, such as thioethers, responsible to inducing fishy odors, and dimethyl trisulfide contribute to the unique aroma of seafood (Buettner, 2017; X.-H. Huang et al., 2019; Jensen et al., 2002). Furthermore, THTT applied after marination promotes the development of unique seafood flavor and simultaneously enhances pleasant flavors such as fresh and sweet fruity notes and nutty aromas.

Fig. 2.

Radar plot of e-nose volatile scores (A) and the corresponding principal component analysis (PCA) (B).

The radar plot variations were clarified through the PCA experiment. Fig. 2 (B) illustrates that the first and second principal components (PC1 and PC2, respectively) together contributed 98.4%, of the overall variance, indicating the significant potential of the model to fully characterize the flavor variations of the samples. Furthermore, the RAW-FM1 and LTB-FM2 demonstrated no significant variations, indicating that low-temperature baking results in dried tuna with odor characteristics largely similar to those of the raw sample. UM-THTT-FM3 and P-THTT-FM4 were spatially distinct from RAW-FM1 and LTB-FM2, demonstrating that THTT significantly altered the aromatic characteristics of dried tuna. However, the e-nose alone remained unable to completely identify and quantify volatile compounds due to the possible interference of sensor materials with the detection results. Hence, the current research employed more sensitive and accurate techniques for the comprehensive evaluation of volatile components.

3.2. GC–IMS

3.2.1. GC-IMS identification of VOCs

GC-IMS was utilized for identifying the variations in the VOCs of dried tuna during different processing stages. Data visualization included 3D topographic maps and 2D differential comparison maps, with 3D spectra employed for the comparison of the aroma profiles of the four samples, as shown in Fig. 3 (A). The 3D spectrum comprised the signal peak intensity (Z-axis), retention index (Y-axis), and drift time (X-axis) of the components. The background in the plot is shown in blue, and a more intense red coloration of the signal peaks indicates higher concentrations (Weller et al., 2022). The aroma components and corresponding concentration varied among the four samples, suggesting an alteration in the aroma profile during various processing stages. To enable further comparisons of the VOCs from different processing stages, the 2D spectra in Fig. 3 (B) and the differential spectra in Fig. 3 (C) were generated by dimensionality reduction of the original spectra (Z. He et al., 2025). The findings demonstrated marked differences in the types and amounts of VOCs among the samples. With the RAW-FM1 spectrum used as a reference, the appearance of small red spots indicated increased compound contents at certain peak positions. In addition, the processed dried tuna demonstrated higher flavor compounds compared with raw tuna meat, indicating the potential of THTT to facilitate the formation of flavor compounds. LTB-FM2, UM-THTT-FM3, and P-THTT-FM4 treatments demonstrated higher proportions of volatile compounds, primarily resulting from protein degradation, lipid oxidation, and the Maillard reaction during the processing.

Fig. 3.

GC-IMS analysis of VOCs in dried tuna. 3D spectrogram(A). Two-dimensional topographic plot(B). Differential comparison spectra (C). VOCs fingerprints at various processing stages (D).

A fingerprint of VOCs was constructed to identify changes in flavor compounds occurring during the dried tuna processing (Fig. 3 D), where individual rows correspond to selected signal peaks within a sample and columns represent the same VOC across various dried tuna samples (Lu et al., 2024). The color brightness indicates positive correlations with compound concentrations, with lower color intensity demonstrating lower concentration (S. Wang et al., 2020). A total of 61 volatile substances were identified by GC-IMS analysis, including 15 aldehydes, 5 alcohols, 12 ketones, 13 alkenes, 4 esters, 2 heterocyclic compounds, 4 acidic substances, 1 other, and 5 unknown substances. The fingerprint can be divided into nine distinct regions: A, B, C, D, E, F, G, H, and I, as illustrated in Fig. 3 (D). The regions A, B, C, and D represent the inherent flavor regulating substances of the dried tuna, while Region F indicates that the marination process significantly promoted the accumulation of acetic acid, (E)-2-hexenal, benzaldehyde, α-pinene, myrcene, ocimene, terpinolene, dipentene, 2-methyl-2-pentenal, L-β-pinene, tetrahydrofuran, nonanal, 2-butanone, and propanoic acid. As shown in Regions C and G, low-temperature baking significantly increased the proportions of propanoic acid, triethylamine, and ethyl formate, while reducing the contents of propanal, n-hexanal, 2-methyl-2-pentenal, and 2-butanone, altering the flavor of dried tuna from pungent fatty and aldehydic notes to milder acidic and ester-like characteristics (Hatairad et al., 2021; Jing et al., 2019; J. Zheng et al., 2025). However, the regions D and I illustrate that THTT reduced the contents of 3-methylbutyl 2-methylbutanoate, n-octanal, and n-butanol, while demonstrating increased content of ethanol, 3-hydroxy-2-butanone, and cyclohexanone. High temperatures facilitated the process of oxidation, Maillard reaction, and degradation of various substances of dried tuna, facilitating the formation of new compounds and thus transforming its flavor from fruity and fatty notes to mellow and creamy notes (Ninglong et al., 2021; Ran et al., 2022; Xing et al., 2025). As indicated in Regions E and H, THTT increased the proportions of n-butanol, 6-methyl-5-hepten-2-one, hydroxyacetone, 1-octen-3-one, 2,5-dimethylpyrazine, 2-pentanone and isovaleraldehyde, while reducing the concentration of compounds having unpleasant odor such as 2-methylbutyraldehyde and tert-butanol, which are by-products of lipid oxidation and hydroxy compounds of branched-chain hydrocarbons (Deng et al., 2024; Xiaoshan et al., 2022), The rapid thermal effect of THTT generates a localized high-temperature steam environment, which accelerates the volatilization and release of these compounds from the matrix (SeMyung et al., 2021), transforming the flavor of dried tuna from pungent off-flavors to a mellow profile with malty, ketonic, and nutty aromas derived from pyrazines (GuoTing et al., 2021; Mortzfeld et al., 2020), while preserving the characteristic seafood flavor imparted by the compound 1-octen-3-one (T. YAN et al., 2026).

The above analysis demonstrated the intensified THTT-induced Maillard and Strecker degradation reactions, facilitating the formation of complex flavor substances such as pyrazines and ketones (Pan et al., 2025; Yingchen et al., 2023). Moreover, the short processing time inhibits lipid oxidation, limiting the accumulation of undesirable aldehydes, while preserving key seafood flavor compounds, thus maintaining the characteristic seafood aroma of the product.

3.2.2. Partial least squares-discriminant analysis (PLS-DA)

PLS-DA is a supervised multivariate statistical method for regression modeling of multiple dependent variables on multiple independent variables (Li et al., 2025). The analysis employed an independent variable fitting index ($R_x^2$) of 0.991, a dependent variable fitting index ($R_y^2$) of 0.992, and the model prediction index ($Q^2$) of 0.98, indicating that the model had an excellent explanatory ability and predictive ability (Yuting et al., 2023). Furthermore, the results of the 200-permutation test showed that the y-intercept criteria for R2 (cum) and Q2 (cum) were 0.025 and −0.452, respectively, which indicated that the developed models were reliable (P. Sun et al., 2024). The PLS-DA score plot for GC-IMS data (Fig. 4C) demonstrated a variance explanation of 91.5%. This plot exhibited effective discrimination between different sample groups. Hence, the findings were determined to be appropriate for analyzing the volatile compounds in dried tuna.

Fig. 4.

Distribution of VIP values(A),permutation test plot of 200 permutations (B), and PLS-DA score scatter plot (C).

The VIP is frequently used for evaluating the key variables in the PLS-DA model (Y. Wang et al., 2025), with compounds having VIP value > 1 identified as characteristic aroma components (Q. Chen et al., 2024; Yujiao et al., 2022), The observed VIP values (Fig. 4A) show the identification of 13 significant volatile compounds, including 6 ketones, 1 aldehyde, 1 ester, 2 acids, and 1 alkene. Propanoic acid-D, propanoic acid-M, and 2-methyl-2-hepten-6-one, having VIP values exceeding 2.0, were identified as the major VOCs, followed by β-ocimene-M, 2-methyl butanal, and Acetone, having VIP values higher than 1.4.

3.3. GC-O-MS analysis

3.3.1. GC-O-MS analysis of volatile substances

HS-SPME-GC-O-MS analysis identified 89 compounds (Fig. 5 and Table 1), including 6 aldehydes, 5 ketones, 11 alcohols, 3 sulfur-containing compounds, 14 benzene and related derivatives, 11 nitrogen-containing compounds, 4 esters, 18 alkanes, 19 alkenes, 2 acidic substances, 4 ethers and 1 other compounds, as shown in Fig. 5 (A). The P-THTT-FM4 group demonstrated abundant volatile compounds, dominated by aldehydes, benzene and its derivatives, alcohols, alkanes, nitrogen-containing compounds, and alkenes, with significantly higher proportions compared with other experimental groups, identifying the P-THTT-FM4 group as possessing the richest and most complex flavor profile. Furthermore, Fig. 5 (B) and (C) illustrate the variations in the levels of VOCs among the samples. The P-THTT-FM4 group demonstrated the highest contents of aldehydes, ketones, benzene and its derivatives along with alcohols, relative to the other groups. The UM-THTT-FM3 group showed the optimal concentration of alkanes and nitrogen-containing compounds, while the LTB-FM2 group demonstrated the highest alkenes content.

Fig. 5.

Heatmaps of VOCs in dried tuna samples measured by HS-SPME-GC–O-MS (A), intense red regions represent higher concentrations, while intense blue regions denote lower concentrations of VOCs in dried tuna. Bubble pool plots of VOCs in dried tuna measured by GC-O-MS (B), where R stands for RAW-FM1, L for LTB-FM2, U for UM-THTT-FM3, and P for P-THTT-FM4. Stacked jade block plots of VOCs in dried tuna samples by GC-O-MS (C).

Table 1.

Analysis of VOCs in dried tuna associated with various processing stages based on gc-O-ms

Number	Name of compound	CAS	Formula	RI	Content (μg/kg)
					RAW-FM1	LTB-FM2	UM-THTT-FM3	P-THTT-FM4
1	2-Pentanamine, 2,4,4-trimethyl-	107-45-9	C8H19N	879	/	1.614	/	/
2	Methylamine, N,N-dimethyl-	75-50-3	C3H9N	502	/	/	33.051	83.826
3	2-Propanamine	75-31-0	C3H9N	477	/	/	0.477	/
4	Formamide, N-butyl-	871-71-6	C5H11NO	1067	/	0.340	0.514	0.949
5	Cystine	56-89-3	C6H12N2O4S2	/	/	0.475	/	0.380
6	Methanaminium, 1-carboxy-N,N,N-trimethyl-, hydroxide, inner salt	107-43-7	C5H11NO2	/	/	/	0.385	/
7	Ethanamine, N-methyl-	624-78-2	C3H9N	482	/	/	0.413	0.627
8	Ethanol, 2-(2-propenyloxy)-	111-45-5	C5H10O2	769	/	/	/	0.429
9	Propanoic acid	1979/9/4	C3H6O2	700	/	/	/	47.103
10	Propanoic acid, ethyl ester	105-37-3	C5H10O2	692	/	45.547	/	/
11	1,3,5-Cycloheptatriene	544-25-2	C7H8	772	0.210	/	/	/
12	Thiophene, 2-methyl-	554-14-3	C5H6S	775	0.171	0.646	0.740	1.477
13	Benzenamine, 3-nitro-N-(2-thienylmethyl)-	18210-24-7	C11H10N2O2S	/	/	/	/	0.997
14	2-Hexanone, 6-hydroxy-	21856-89-3	C6H12O2	/	/	/	/	0.852
15	Oxirane, hexyl-	2984-50-1	C8H16O	/	0.139	/	/	/
16	2-Octen-1-ol	22104-78-5	C8H16O	1066	/	/	/	0.405
17	1,4-Benzenediol, 2,6-bis(1,1-dimethylethyl)-	2444-28-2	C14H22O2	/	/	/	0.457	/
18	Hydrazinecarboxylic acid, phenylmethyl ester	5331-43-1	C8H10N2O2	/	/	/	0.472	/
19	Glycine benzyl ester	1738-68-7	C9H11NO2	/	/	/	0.453	/
20	Benzeneethanol, à,á-dimethyl-	52089-32-4	C10H14O	/	1.528	/	/	/
21	Benzene, 1,3-dimethyl-	108-38-3	C8H10	866	0.749	0.939	/	0.645
22	Styrene	100-42-5	C8H8	893	/	0.626	/	/
23	1,3,7-Octatrien-5-yne	16607-77-5	C8H8	825	/	/	/	1.556
24	Ethylbenzene	100-41-4	C8H10	855	1.040	0.312	2.887	0.640
25	1,3,5,7-Cyclooctatetraene	629-20-9	C8H8	850	/	/	1.350	0.815
26	Benzeneethanamine, á-methyl-	582-22-9	C9H13N	1130	/	/	/	0.519
27	Indan, epoxide	768-22-9	C9H8O	/	/	0.326	/	1.130
28	3H-2-Benzopyran-3-one, 1,4-dihydro-	4385-35-7	C9H8O2	/	/	/	/	0.447
29	à-Pinene	80-56-8	C10H16	937	/	5.946	/	16.777
30	1,2,4-Trioxolane, 3-phenyl-	23253-30-7	C8H8O3	/	/	/	/	1.143
31	Benzaldehyde	100-52-7	C7H6O	962	/	8.570	/	2.289
32	1,2-Benzenedicarboxaldehyde	643-79-8	C8H6O2	/	/	/	/	1.429
33	Benzene, 1-ethyl-4-methyl-	622-96-8	C9H12	954	1.035	/	1.443	/
34	Benzene, (1-nitroethyl)-	7214-61-1	C8H9NO2	/	/	/	0.803	/
35	á-Pinene	127-91-3	C10H16	993	/	59.256	/	90.068
36	5-Hepten-2-one, 6-methyl-	110-93-0	C8H14O	990	/	/	/	10.019
37	Cyclopentaneacetaldehyde, 2-formyl-3-methyl-à-methylene-	5951-57-5	C10H14O2	1314	/	/	2.082	/
38	cis-á-Farnesene	28973-97-9	C15H24	1444	/	/	1.038	/
39	Decane	124-18-5	C10H22	1000	0.144	0.508	1.660	7.708
40	Bicyclo [3.1.0]hexan-2-ol, 2-methyl-5-(1-methylethyl)-, (1à,2à,5à)-	17699-16-0	C10H18O	1070	/	/	/	3.771
41	o-Cymene	527-84-4	C10H14	1022	0.761	13.782	1.287	12.245
42	Limonene	138-86-3	C10H16	1030	1.245	793.245	0.401	1326.871
43	Cyclohexene, 1-methyl-5-(1-methylethenyl)-, (R)-	1461-27-4	C10H16	1033	/	0.885	28.860	/
44	Propane, 1-(1,1-dimethylethoxy)-2,2-dimethyl-	32970-46-0	C9H20O	/	12.209	15.218	34.699	12.752
45	4-Carene, (1S,3R,6R)-(−)-	5208-49-1	C10H16	1060	/	/	/	0.685
46	p-Mentha-1(7),8(10)-dien-9-ol	29548-13-8	C10H16O	1287	/	0.968	/	0.571
47	ç-Terpinene	99-85-4	C10H16	1060	1.086	106.785	10.596	174.444
48	Decane, 4-methyl-	2847-72-5	C11H24	1060	/	/	0.610	/
49	Decane, 3-methyl-	13151-34-3	C11H24	1071	/	1.001	4.413	4.309
50	Isopinocarveol	6712-79-4	C10H16O	1180	0.136	/	/	0.826
51	3-Nonanone	925-78-0	C9H18O	1090	200.000	200.000	200.000	200.000
52	1,6-Octadien-3-ol, 3,7-dimethyl-	78-70-6	C10H18O	1099	/	1.256	0.373	1.784
53	Nonanal	124-19-6	C9H18O	1099	1.142	2.380	4.016	1.037
54	2,6-Dimethyldecane	13150-81-7	C12H26	1112	0.278	/	1.333	/
55	Decane, 2,5,6-trimethyl-	62108-23-0	C13H28	/	/	/	1.516	3.862
56	2-Methyl-1-undecanol	10522-26-6	C12H26O	1422	/	0.514	/	0.896
57	2,2-Dimethyl-3-hydroxypropionaldehyde	597-31-9	C5H10O2	/	/	/	/	0.638
58	Tetracontane, 3,5,24-trimethyl-	55162-61-3	C43H88	/	/	/	0.483	/
59	2(3H)-Furanone, dihydro-4,4-dimethyl-	13861-97-7	C6H10O2	/	0.107	/	/	0.444
60	Terpinen-4-ol	562-74-3	C10H18O	1177	/	4.000	/	4.244
61	1H-Indene, 1-methylene-	2471-84-3	C10H8	1097	0.100	/	/	/
62	à-Terpineol	98-55-5	C10H18O	1189	/	1.513	/	1.454
63	Heptadecane, 2,6-dimethyl-	54105-67-8	C19H40	/	0.818	2.175	0.419	1.879
64	1-Hexene, 3,5-dimethyl-	7423-69-0	C8H16	700	/	/	0.386	/
65	Ethanol, 2-phenoxy-	122-99-6	C8H10O2	1226	4.048	6.781	10.558	11.898
66	2,4,6-Octatrienoic acid	5205-32-3	C8H10O2	/	/	/	/	0.422
67	Diborane (6), æ-mercapto-	22548-43-2	B2H6S	/	/	0.196	/	/
68	2,2,7,7-Tetramethyloctane	1071-31-4	C12H26	1036	0.591	0.646	1.040	/
69	Dodecane, 2,6,11-trimethyl-	31295-56-4	C15H32	1275	/	0.845	8.177	/
70	Dodecane, 4,6-dimethyl-	61141-72-8	C14H30	1325	/	/	2.547	/
71	2-Dodecenal	4826-62-4	C12H22O	1465	/	0.225	/	/
72	Tridecane	629-50-5	C13H28	1284	/	1.045	4.104	4.654
73	Dodecane, 2,7,10-trimethyl-	74645-98-0	C15H32	/	/	0.951	/	1.561
74	Decane, 2,6,6-trimethyl-	62108-24-1	C13H28	/	0.511	/	/	1.042
75	Nonane, 2,2,4,4,6,8,8-heptamethyl-	4390/4/9	C16H34	1322	12.990	18.759	26.329	30.593
76	Hexane, 3,3-dimethyl-	563-16-6	C8H18	744	0.295	/	1.432	1.967
77	2,4,4-Trimethyl-1-hexene	51174-12-0	C9H18	/	/	0.271	/	/
78	ç-Dodecalactone	2305/5/7	C12H22O2	1678	/	/	/	0.595
79	Propane, 2-methyl-2-nitro-	594-70-7	C4H9NO2	/	/	0.485	0.533	/
80	Dodecane, 5,8-diethyl-	24251-86-3	C16H34		0.385	0.622	/	0.613
81	Octane, 2,4,6-trimethyl-	62016-37-9	C11H24		0.212	/	1.078	/
82	Nonane, 3-methyl-5-propyl-	31081-18-2	C13H28		0.462	1.534	/	/
83	cis-à-Bisabolene	29837-07-8	C15H24		/	/	/	1.724
84	Bicyclo [7.2.0]undec-4-ene, 4,11,11-trimethyl-8-methylene-, [1 R-(1R*,4Z,9S*)]-	118-65-0	C15H24		/	1.165	/	1.135
85	Cyclohexene, 1,5,5-trimethyl-3-methylene-	16609-28-2	C10H16		/	0.329	0.557	/
86	trans-à-Bergamotene	13474-59-4	C15H24		/	1.781	/	/
87	Cyclohexane, 1-methylene-4-(1-methylethenyl)-	499-97-8	C10H16		/	0.209	0.382	/
88	Decane, 2-methyl-	6975-98-0	C11H24		0.836	6.856	2.622	27.067
89	Bicyclo [3.1.1]hept-2-ene, 2,6-dimethyl-6-(4-methyl-3-pentenyl)-	17699-05-7	C15H24		/	0.728	1.340	1.858

Aldehydes are recognized as the major odor contributing compounds in fish muscle during thermal processing (J. Li et al., 2025). Table 1 shows the six identified aldehydes, namely, Benzaldehyde, 1,2-Benzenedicarboxaldehyde, Cyclopentaneacetaldehyde (2-formyl-3-methyl-methylene), Nonanal, 2,2-Dimethyl-3-hydroxypropionaldehyde, and 2-Dodecenal, with Benzaldehyde exhibiting nutty flavor and derived from the Strecker degradation of phenylalanine (Na et al., 2022), In contrast, 1,2-Benzenedicarboxaldehyde was formed by the further oxidation of Benzaldehyde, with extensive detection in the P-THTT-FM4 group, demonstrating synergistic action with Benzaldehyde to enrich the layering of fishy aromatic notes. Fish muscle is enriched with unsaturated fatty acids and nonanal, formed by the oxidative cleavage of linoleic acid (Luo et al., 2026), which contributes a rich, fatty aroma and fresh sweetness. However, excessive amounts can lead to rancid and greasy off-flavors (Zeng et al., 2025). The UM-THTT-FM3 group demonstrated the highest concentration of Nonanal, reaching 4.016 μg/kg (Table 1, Table 2), with ROAV analysis identifying Nonanal (ROAV>1) as the key compound contributing to the prominent cured flavor in this group. The formation of 2-Dodecenal is induced by the oxidation of C₂₀ and higher unsaturated fatty acids (Lara et al., 2022), and contributes significantly to the typical fatty aroma of marine fish. However, its excessive accumulation can result in a strong fishy odor and greasiness, demonstrating significant detection in the LTB-FM2 group. Among the 6 aldehyde flavor compounds detected in the current study, four were present in the P-THTT-FM4 group, aligning well with the higher aldehyde content observed in this group by E-nose and confirming its richer and more diverse flavor composition.

Table 2.

Analysis of FN content in dried tuna samples.

FNs	Content (μg/g)
	RAW-FM1	LTB-FM2	UM-THTT-FM3	P-THTT-FM4
GMP	17.36 ± 2.77d	22.33 ± 2.24c	40.35 ± 2.46b	28.51 ± 3.45d
IMP	1833.42 ± 43.73a	989.37 ± 32.55a	1235.13 ± 36.45a	748.477 ± 20.65b
AMP	295.83 ± 23.53c	473.07 ± 39.14b	179.61 ± 15.91b	605.78 ± 18.70c
Hx	6.64 ± 0.93d	5.53 ± 0.11c	102.34 ± 0.11b	26.77 ± 1.98d
HxR	580.24 ± 43.81b	1092.56 ± 113.45a	1131.80 ± 113.45a	1088.85 ± 29.61a

Alcohols are formed by the oxidative degradation of unsaturated fatty acids (Zhang et al., 2018). The transient high-temperature treatment process facilitated the oxidation of monounsaturated fatty acids into peroxides, with further degradation of intermediates into aldehydes (Siliang et al., 2020), with subsequent reduction to their corresponding alcohols by aldehyde reductase. As seen in Table 1, Table 2, a total of 11 alcoholic flavor compounds were detected through GC-O-MS, with the most abundant and extensive distribution of 2-phenoxyethanol. However, it demonstrated a negligible effect on the tuna flavor due to the high aroma threshold. Oxidative degradation of unsaturated fatty acids led to the formation of 1,6-Octadien-3-ol, 3,7-dimethyl-, which has a pronounced impact on the dried tuna flavor, contributing fresh and vibrant notes to the product, and is excessively identified in the P-THTT-FM4 and LTB-FM2 groups, with respective values of 1.784 and 1.256 μg/kg. Its content was negligible in the UM-THTT-FM3 group, thus validating its formation during the marination process and the potential of THTT to enhance its corresponding flavor. Enzymatic hydration of pinene results in the formation of α-Terpineol (ZhongLei et al., 2022), which, upon thermal rearrangements, provides Terpinen-4-ol (R.R. P., 2017), with both terpenolic alcohols contributing a fresh citrus and a mild woody aroma, respectively (F. He et al., 2018; Yujiao et al., 2022), enriching the flavor layering of dried tuna. Both terpenolic alcohols were detected in the LTB-FM2 and P-THTT-FM4 groups. Furthermore, among the 11 detected alcohol substances, the P-THTT-FM4 group demonstrated the highest proportions, indicating its abundant alcoholic flavor profile.

Nitrogen-containing compounds are essential in developing the flavor profile of dried tuna. GC-O-MS identified 14 nitrogen-containing compounds, with N,N-dimethylmethylamine primarily originating from trimethylamine N-oxide (TMAO) degradation by microbial enzymes, which are naturally present in tuna muscle (Lesley et al., 2018; Yamagata et al., 1969). The content of trimethylamine, as the core product of this conversion pathway, is closely associated with the dried tuna flavor, with excessive accumulation inducing flavor deterioration (Cobb and Vanderzant, 1971; Zheng et al., 2025) (and generating a fishy odor, while an appropriate content serves as the core volatile compound contributing the characteristic seafood flavor. As shown in Table 1, the UM-THTT-FM3 and P-THTT-FM4 groups demonstrated higher content of trimethylamine compared with other groups, but remained within a limit of 1900 μg/kg (Z. ZHAI et al., 2025). The observed finding indicated that transient high-temperature treatment inactivated microorganisms, inhibiting the excessive trimethylamine production (Heising et al., 2014). Meanwhile, high temperatures facilitate the decomposition of TMAO into trimethylamine, supporting previous results (Y. Li et al., 2020). In summary, THTT could enhance the seafood flavor of dried tuna.

Esters are a significant aroma contributor in dried tuna products, exerting considerable effect on the overall flavor characteristics of the food because of low odor thresholds (Huaixiang et al., 2023). These compounds have been reported to induce fruity, floral, and creamy aromas even at low concentrations (Jie et al., 2020). Lipolysis and esterification result in the formation of ester: enzymes hydrolyze triglycerides to release free fatty acids (FFAs) and glycerol (Cho et al., 2023), followed by esterase-mediated esterification of FFAs with alcohols (Ping et al., 2025). Four esters were identified through GC-O-MS in dried tuna, including Propanoic acid, ethyl ester, γ-Dodecalactone, Hydrazinecarboxylic acid, phenylmethyl ester, and Glycine benzyl ester. Propanoic acid, ethyl ester, is formed through esterification and provides a fruity aroma (Wan-Hou and Zhi-Shen, 2012). In contrast, ç -Dodecalactone is formed from unsaturated fatty acids like linoleic and linolenic acids through oxidative hydroxylation, followed by hydrolysis and lactonization to generate the cyclic lactone (Shuji et al., 2022), contributing to the fatty and fruity aromas (C. Chen et al., 2022), with respective identification in LTB-FM2 and P-THTT-FM4 groups.

Alkenes are key volatile compounds in dried tuna, occurring in an abundant variety and quantity, and can be classified into alkene hydrocarbons (non-terpenoid) and terpenoid compounds. Terpenoids are a major class of natural products imparting distinct fresh, floral, and woody aromas to dried tuna (Li et al., 2025). Table 1, Table 2 summarize 19 identified alkene volatile compounds through the GC-O-MS, including 16 terpenoid alkenes and 3 alkene hydrocarbons. The LTB-FM2 and P-THTT-FM4 groups demonstrated the presence of relatively high levels of α-Pinene, Limonene, and ç –Terpinene, with the P-THTT-FM4 group demonstrating their highest ROAV values and concentrations. The observed findings suggested that although these compounds were produced during the low-temperature baking process, THTT has the potential to enhance their flavor contribution, further confirming its role in promoting the formation of characteristic flavor compounds in dried tuna.

3.3.2. Analysis of key volatile flavor compounds

ROAV can be used to assess the food flavor contribution of each compound (Yifan et al., 2020); therefore, this study utilized this approach to examine the major contributors to dried tuna flavor. Compounds having ROAV ≥1 were recognized as the primary flavor contributors, while 0.1 ≤ ROAV <1 values were indicative of secondary contributors (Dan et al., 2022). As shown in Fig. 6, 16 VOCs were found, including 4 alcohols, 1 ketone, 2 aldehydes, 1 ester, 2 acids, and 5 alkenes. Among them, Methylamine, N,N-dimethyl-, Propanoic acid, ethyl ester, Limonene, ç -Terpinene, Nonanal, and 1,6-Octadien-3-ol, 3,7-dimethyl- exhibited ROAV ≥1, while Propanoic acid, Styrene, à-Pinene, Benzaldehyde, á-Pinene, à-Terpineol, and ç -Dodecalactone demonstrated ROAV $\in [0,1)$. The following compounds contribute distinct characteristics: ç-Terpinene and Limonene impart a unique floral and woody aroma to P-THTT-FM4, while Methylamine, N,N-dimethyl- gives UM-THTT-FM3 a pronounced fishy odor.

Fig. 6.

ROAVs of key aroma compounds in dried tuna. Red and blue colors indicate higher and lower levels, respectively, of VOCs.

3.4. Analysis of 5′-nucleotides

Free nucleotides contribute substantially to the flavor of aquatic products (H. Fan et al., 2014). The effect of THTT on the nucleotide content in dried tuna is illustrated in Table 2. AMP, IMP, and GMP are key nucleotides contributing to umami and sweet characteristics in marine products (Zhao et al., 2021). RAW-FM1 demonstrated the presence of the highest IMP content (1833.42 μg/g), indicating that marination and heat treatment decreased the IMP content, consistent with the findings of Xu (Pengfei et al., 2023). The P-THTT-FM4 group revealed the identification of the highest AMP content (605.78 μg/g), primarily due to the inactivation of deaminase by THTT, therefore preventing its conversion into IMP. Meanwhile, the high temperature facilitated the protein hydrolysis to generate more abundant umami and sweet nucleotides, with significant participation in the conversion of IMP to AMP (Bao et al., 2025), thereby resulting in lower levels of IMP and higher levels of AMP. In contrast to the P-THTT-FM4 group, the UM-THTT-FM3 group demonstrated an opposite trend, primarily due to the inhibited deaminase activity mediated by the marination (S. Li et al., 2017), preventing the frequent conversion of AMP to IMP during the low-temperature baking stage of dried tuna. This effect resulted in decreased AMP content (179.61 μg/g) in the UM-THTT-FM3 group, while promoting the accumulation of IMP as a precursor of GMP (Acosta et al., 2018), ultimately resulting in a significantly higher GMP content in this group (40.35 μg/g).

3.5. Free amino acid analysis

FAAs are recognized as the major flavor components in aquatic products, contributing to umami, sweet, and bitter tastes (Akitomi et al., 2013). As shown in Table 3, all dried tuna samples demonstrated significantly increased content of FAAs compared with that in RAW-FM1 (P < 0.05), primarily due to the formation of FAAs from thermal degradation of proteins (Rehlund et al., 2025). The P-THTT-FM4 group demonstrated markedly higher levels of umami, sweet, and bitter FAAs relative to the other groups (P < 0.05), thus indicating that THTT-treated dried tuna possessed a richer flavor hierarchy along with significantly enhanced flavor intensity.

Table 3.

Analysis of FAA content in dried tuna samples.

FAAs	Content (μg/g)
	RAW-FM1	LTB-FM2	UM-THTT-FM3	P-THTT-FM4
Asp	2.73 ± 0.36d	170.70 ± 16.44b	25.18 ± 4.54c	223.94 ± 8.27a
Glu	319.05 ± 78.40b	1519.19 ± 375.05a	508.53 ± 40.06b	1506.66 ± 409.44a
Umami FAAs	321.78 ± 78.75b	1689.89 ± 367.20a	533.71 ± 36.07b	1730.60 ± 404.46a
Ser	74.74 ± 4.50c	306.58 ± 22.53a	203.81 ± 29.71b	203.81 ± 29.71a
Thr	78.00 ± 9.06d	255.078 ± 7.94b	178.52 ± 7.90c	397.43 ± 12.96a
Ala	455.76 ± 6.04d	985.92 ± 19.06b	717.95 ± 18.80c	1165.97 ± 73.73a
Gly	220.50 ± 16.95b	329.78 ± 19.99a	245.17 ± 10.10b	363.8 ± 13.88a
Sweet FAAs	829.01 ± 34.35d	1877.36 ± 53.52b	1345.45 ± 50.67c	2226.44 ± 156.19a
Val	231.032 ± 84.65b	677.02 ± 103.83a	759.42 ± 75.51a	593.31 ± 174.34a
Met	64.59 ± 6.14c	9.70 ± 0.83d	233.01 ± 28.89a	142.06 ± 14.54a
Ile	75.48 ± 10.87b	494.98 ± 109.06a	451.62 ± 122.93a	433.30 ± 126.69a
Leu	119.50 ± 27.39b	736.24 ± 148.84a	833.57 ± 192.33a	733.41 ± 211.99a
Tyr	136.13 ± 46.47b	354.96 ± 74.68a	342.84 ± 92.17a	352.69 ± 56.19a
Lys	380.53 ± 56.08c	1205.14 ± 183.39b	2355.57 ± 196.09a	1261.65 ± 305.05b
His	9370.94 ± 148.31b	8463.38 ± 140.56c	7183.49 ± 77.64d	10 097.57 ± 529.68a
Arg	29.05 ± 11.53b	467.72 ± 146.56a	306.23 ± 112.00a	403.73 ± 146.17a
Bitter FAAs	10 407.26 ± 384.94a	12 409.14 ± 893.53ab	12 465.77 ± 886.64ab	14 017.73 ± 1514.16a
Cys	29.54 ± 1.80c	172.26 ± 11.83a	2.44 ± 0.50d	154.19 ± 8.56a
Phe	67.49 ± 25.08b	431.28 ± 116.39a	237.32 ± 88.64ab	370.48 ± 95.71a
Tasteless FAAs	97.04 ± 26.25b	603.55 ± 124.11a	239.76 ± 88.80b	524.66 ± 101.84a
Total FAAs	11 655.09 ± 521.44c	16 579.94 ± 1415.83ab	14 584.69 ± 1060.31bc	18 499.43 ± 2170.05a

The taste thresholds and the calculation formula for the sample TAV were obtained based on the previously published data (HU et al., 2021). The flavor intensity of FAAs was evaluated using TAV. In general, TAV >1 compounds are regarded as key substances affecting food flavor (Liu et al., 2020), As presented in Fig. 7, Glu was considered as a representative umami FAAs (Yu et al., 2024), with a significant increase in its content (1506.66 μg/g, TAV>1) after THTT (P < 0.05). Ala was the primary taste-active FAAs that contributed to the sweetness in dried tuna. Being the main source of sweetness in the P-THTT-FM4 group, Ala demonstrated a content of 1165.97 μg/g (TAV >1). Umami FAAs and nucleotides can typically synergize with sweet FAAs to improve the umami taste (Yoshii et al., 1986). The transamination reaction of Ala with α-ketoglutarate resulted in the formation of Glu (Rao and Adams, 1978).

Fig. 7.

TAV of FAAs in dried tuna samples. Blue color represents higherTAV, while red indicates lower TAV.

After THTT treatment, the contents of bitter FAAs (Val, His, and Lys) were significantly increased (TAV >1), consistent with the results of Li et al. (Rui et al., 2022). Although His is categorized as a bitter amino acid, Fang et al. indicated its association with the meaty aroma of certain aquatic products. (FANG et al., 2025). Moreover, bitter FAAs imparted distinct flavor qualities, improved the umami taste (D.-W. Chen and Zhang, 2007), and ultimately enriched the flavor hierarchy.

3.6. Correlation analysis and mechanism discussion

The relationship between producing taste and odor-generating compounds in dried tuna products after THTT was analyzed. The results demonstrated that THTT significantly enhanced the synergistic enhancement effects of umami nucleotides, umami FAAs and sweet FAAs in dried tuna products. Meanwhile, such thermal processing facilitated the specific formation of 29 important aroma compounds, thereby providing a theoretical foundation for the targeted regulation of flavor quality in dried tuna products. As illustrated in Fig. 8, AMP and IMP exhibited marked positive and negative correlations with umami FAAs and sweet FAAs, respectively. This result is further confirmed by the correlation between AMP and IMP (P < 0.001,r = −0.691), indicating a negative correlation of AMP (as the IMP precursor) with the content of IMP. Excluding His, most bitter FAAs showed a strong positive correlation with HxR. Excessive accumulation of HxR contributed to the bitter flavor of fish products (J. Qin et al., 2020), which can be effectively suppressed by inhibiting IMP. A PLSR analysis was conducted to confirm the interaction between taste nucleotides and FAAs (Fig. 9). The loading plot featured two ellipses, with the outer and inner ellipses signifying 100 and 50% of the explained variance, respectively (Danni et al., 2022; Lv et al., 2024; D. Sun et al., 2025).

Fig. 8.

Correlation analysis between free FAAs, nucleotides, and key aroma compounds (ROAV >1, VIP >1) in dried tuna samples. Red represents a strong correlation and blue represents a weak correlation.

Fig. 9.

PLSR of taste nucleotides and FAAs in dried tuna samples.

IMP was projected on the left, whereas AMP and FAAs were on the right side of the loading plot. IMP and FAAs were negatively and AMP and FAAs were positively correlated. AMP was positioned in the vicinity of umami and sweet FAAs, which enhanced the umami intensity of THTT-treated dried tuna products because of the combined effects of AMP, umami, and sweet FAAs. Therefore, THTT had a distinctive effect on the differences in nucleotide and FAAS levels in dried tuna products. The potential taste production pathway of dried tuna products was proposed according to the differences in taste qualities (Fig. 10). The death of aquatic animals decomposed muscle glycogen and produced large amount of ATP (Zhang et al., 2025; Zhang et al., 2024), with ATPase being enzymatically active during the low-temperature baking stage, resulting in a significant accumulation of AMP in dried tuna (GU et al., 2020). THTT was then inactivated by 5′-adenylate deaminase to inhibit the accumulation of IMP. Furthermore, THTT considerably inhibited the spontaneous exothermic process of converting AMP to IMP (Danni et al., 2022), decreasing the IMP content.Moreover, IMP exhibited comparatively lower Gibbs free energy than HxR, suggesting a relatively challenging conversion of IMP to HxR (Y. Zhai et al., 2024), and relying on the enzymatic reaction of 5′-adenylate deaminase. THTT inactivated 5′-adenylate deaminase, consequently preventing the accumulation of HxR According to the results of correlation and thermodynamic analysis, the nucleotides and FAAs synergistically improved the taste of dried tuna products.

Fig. 10.

Schematic illustrating the influence of THTT on producing flavor (odor and taste) and potential pathways for the dried tuna. Transformation pathway for taste nucleotides (A), generation pathway for taste-demonstrating FAAs (B), and production pathway for volatile components (C) in THTT-treated dried tuna sample.

Fig. 8, Fig. 10 illustrate the correlations and hypothetical underlying mechanisms of the characteristic volatile compounds. During thermal transformation, á-Pinene showed a positive association with 3,7-dimethyl-1,6-octadien-3-ol (r = 0.98217, P < 0.01) and 6-methyl-5-hepten-2-one (r = 0.93656, P < 0.01), transformed into β-pinene via thermal isomerization, and then underwent further thermal decomposition to produce myrcene (Y. HUANG et al., 2025; Kolicheski et al., 2007). Subsequently, myrcene was hydrolyzed to generate 3,7-dimethyl-1,6-octadien-3-ol (Xun et al., 2023) and oxidized to form geranial or neral (Agnieszka et al., 2022), which, when combined, formed citral. Citral was hydrolyzed to generate an intermediate, 3,7-dimethyl-7-hydroxyoctanal, which was then converted into 6-methyl-5-hepten-2-one and acetaldehyde via retro-aldol condensation (Jun et al., 2022; Valla et al., 2006). Acetic acid ethyl ester exhibited a significant negative correlation with Ala (r = −0.61242, P < 0.01) and α-Pinene (r = −0.8258, P < 0.01). Transamination converted Ala into pyruvate (LIU et al., 2025), which subsequently decarboxylated to form acetaldehyde. Along with previously described acetaldehyde, acetaldehyde resulting from the retro-aldol condensation of 3,7-dimethyl-7-hydroxyoctanal was reduced to ethanol by the release of free radicals from heated fish muscle (Andrés et al., 2023; Y. CHEN et al., 2026). Meanwhile, some acetaldehyde was oxidized to acetic acid upon heating, which subsequently underwent an esterification reaction with ethanol to produce ethyl acetate. Furthermore, acetone was oxidized to produce propionic acid, which was then esterified with ethanol to yield ethyl propionate. ç -Terpinene showed a significant positive association with Limonene (r = 0.99897, P < 0.01) and (+)-limonene-M (r = 0.9234, P < 0.01), whereas 3,7-dimethyl-1,6-octadien-3-ol with limonene (r = 0.97976, P < 0.01) and γ-Terpinene (r = 0.98734, P < 0.01). Limonene cannot be prepared de novo in most animals and is instead derived from dietary sources. Limonene and (+)-limonene-M can be isomerized to ç-Terpinene, illustrating a strong correlation among these three compounds. Moreover, 3,7-dimethyl-1,6-octadien-3-ol displayed a positive association with limonene (r = 0.97976, P < 0.01) and ç-terpinene (r = 0.98734, P < 0.01), most likely as a result of oxidation and rearrangement of monoterpene precursors yielding monoterpene alcohols (Ren et al., 2020). Nonanal showed a correlation coefficient with most of the volatile compounds, where $|r|$ <0.9, indicating a relatively independent biosynthetic pathway. According to Zhai et al. (Y. Zhai et al., 2024), the formation of nonanal in dried fish products is primarily linked to the degradation of oleic acid through two pathways: (a) formation of alkoxyl radicals by cleaving 9-hydroperoxide of oleic acid, followed by the keto-enol tautomerization, and (b) by the β-homolytic cleavage of 10-hydroperoxide of oleic acid (A. Ding et al., 2020; Wu et al., 2026). TMAO was decomposed to prepare a methylamine, N, N-dimethyl-. TMAO showed a strong correlation ($|r|$ >0.9) with various distinctive flavor compounds, mostly due to a combination of TMAO decomposition and lipid oxidation during processing (PARKIN & HULTIN, 2008), resulting in highly consistent variations in the contents of these substances, thus demonstrating a significant positive correlation. In the above changes, terpenoids such as limonene, α-pinene, β-pinene, γ-terpinene, and terpinen-4-ol were significantly enriched in the P-THTT-FM4 group.These terpenoids are mainly derived from the spice adjuncts added during the marination stage. THTT treatment promotes their release from the spice matrix and penetration into the fish tissue, imparting citrus, pine needle, and herbal aromas to the finished productShort-chain esters such as ethyl propanoate and ethyl acetate impart fruity and sweet aroma notes to the product. In addition, lipid oxidation products such as 1,6-octadien-3-ol, 3,7-dimethyl- (linalool), nonanal, and 2-methyl-2-hepten-6-one are generated from polyunsaturated fatty acids in tuna muscle through thermal degradation. The high content of such substances in the P-THTT-FM4 group indicates that THTT can accelerate the oxidative decomposition of lipids while avoiding significant volatile losses caused by prolonged heating, thereby imparting the characteristic oily, fatty, and mild nutty and milky aromas of cooked fish products to the product. All these products contributed to the overall flavor of dried tuna.

4. Conclusions

The current research study systematically examined the influence of THTT on the dried tuna flavor, essential metabolic pathways, and regulatory mechanisms. GC-O-MS and GC-IMS identified 89 and 61 VOCs, respectively. THTT was demonstrated to cause protein hydrolysis, lipid oxidation, Maillard reaction, and other biochemical alterations in dried tuna, resulting in an increase in the contents of various volatile compounds, including aldehydes, benzene and its derivatives, alcohols, alkanes, nitrogen-containing compounds, and alkenes, therefore contributing to the improved flavor expression of THTT-treated dried tuna. Furthermore, ROAV and VIP analyses identified 16 and 13 important aroma compounds, respectively. In addition, THTT stimulated the accumulation of umami nucleotide AMP (increased from 295.83 μg/g to 605.78 μg/g) and increased the contents of bitter substances, such as Hx (increased from 6.64 μg/g to 26.77 μg/g) and HxR (increased from 580.24 μg/g to 1088.85 μg/g), in dried tuna. The dried tuna showed a significant increase in the contents of umami (increased from 321.78 μg/g to 1730.60 μg/g), sweet (increased from 829.01 μg/g to 2226.44 μg/g), and bitter (increased from 10 407.26 μg/g to14017.73 μg/g) FAAs, along with distinctive volatile compounds, after THTT treatment, leading to a more varied hierarchy of tastes. Based on the correlation analysis, it is speculated that the flavor quality of dried tuna products was synergistically improved by the THTT treatment-induced nucleotide degradation and protein hydrolysis. It is further speculated that controlling enzyme activities and thermochemical reactions promoted the enrichment of flavor substances, improved flavor characteristics. This study offers a novel heating treatment to enhance the flavor characteristics of dried tuna products. Further studies should focus on the analysis of sensory evaluation experiments, closely monitor the role of lipid oxidation, and combine flavoromics and lipidomics to elucidate the comprehensive mechanism of lipid conversion to flavor during THTT treatment of dried tuna.

CRediT authorship contribution statement

XU Long: Writing– review & editing, Writing– original draft, Investigation, Conceptualization, Data curation, Methodology, Visualization, Formal analysis, Validation. YU Haixia: Writing– review & editing, Supervision, Resources, Project administration, Funding acquisition. ZHANG Xiaojun: Writing– review & editing, Writing– original draft, Supervision, Investigation, Funding acquisition, Conceptualization. ZHOU Peng: Writing– review & editing,. Yu Haitao: Resources. Wang Wenhao: Software, Visualization. XU Dan: Writing– review & editing. CHEN Xiaxia: Writing– review & editing.

Data availability

Data will be made available on request.

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.

Appendix A. Supplementary data

The following is the Supplementary data to this article.