Skip to main content
NCBI home page
Current Research in Food Science logoLink to Current Research in Food Science
. 2026 Jan 27;12:101328. doi: 10.1016/j.crfs.2026.101328

Molecular sensory science elucidates the contribution of methyl ketones to the creamy aroma of Monascus-fermented cheese

Yadong Wang 1, Ying Wang 1, Ting Meng 1, Bei Wang 1,⁎, Yanbo Wang 1
PMCID: PMC12876307  PMID: 41657719

Abstract

Monascus-fermented cheese (MC) is a novel Chinese delicacy characterized by distinctive fruity, creamy, and fermented aromas. This study systematically characterized, using integrated sensory-directed analytical approaches, the aroma compounds during the 90-d ripening stage of MC using molecular sensory science techniques to elucidate the key methyl ketones responsible for its creamy flavor profile. Quantitative descriptive analysis (QDA) revealed a significant increase in the creamy intensity as the ripening process progressed (p < 0.05). Solid-phase microextraction (SPME)-Arrow-gas chromatography-mass spectrometry (GC-MS) identified 115 volatile compounds, of which 36, including 13 methyl ketones, were selected (FD ≥ 81) for further analysis. Further odor activity value (OAV) analysis and aroma omission/addition experiments identified 2-heptanone, 2-nonanone, and 2-undecanone as the key methyl ketone compounds in MC. Stable isotope dilution analysis (SIDA) was employed to quantify these three compounds and to validate their content dynamics across different ripening stages. This study identifies the key compounds behind MC's creamy aroma, providing a scientific basis for future optimization studies.

Keywords: Monascus, Cheese, Methyl ketones, Molecular sensory science

Graphical abstract

Image 1

Open in a new tab

Highlights

  • •

    Creamy increased significantly during the 90-day ripening of MC.

  • •

    Three key methyl ketones (C7, C9, C11) contribute to the creamy aroma.

  • •

    SIDA was applied for the accurate quantification of three key methyl ketones.

1. Introduction

Monascus-fermented cheese (MC) is a novel Chinese cheese product produced through the secondary fermentation of cheese using Monascus spp. Not only does this process provide the cheese with a distinctive pink hue, but it also creates a complex flavor profile consisting of fruity, fermented, and creamy aromas, highlighting its significant research value and market potential (Kumura et al., 2018; Y. Wang et al., 2024; Wu et al., 2019; Zeng et al., 2022). Monascus-associated enzymatic systems contribute to lipid and protein transformation during ripening, providing rich precursors for the formation of flavor compounds. Its metabolic activities can also directly produce various flavor-active compounds (Yang et al., 2025; Zhang et al., 2024, Zhang et al., 2024). However, research on the sensory qualities of MC remains limited, and further investigation is needed to clarify the key compounds responsible for its creamy aroma.

The flavor profile of cheese is a crucial sensory attribute that directly determines consumer acceptance and commercial value (Luo et al., 2024; Martin et al., 2024). The flavor profile of cheese is a highly complex system, formed via the combined action of hundreds of volatile compounds. Its composition and concentration are influenced by multiple factors, including milk sources, starter cultures, production techniques, and ripening conditions (Weimer, 2007). Although over 600 aroma compounds have been identified in cheese to date, only a handful represent the primary characteristic components that contribute to its flavor profile. Of these, only a small number denote key aroma compounds that significantly contribute to the distinctive sensory qualities of a particular cheese type (Ning et al., 2025; J. Wang et al., 2021; W. Zhang et al., 2024). For instance, the characteristic compounds in blue cheese include 2-heptanone, 2-nonanone, methyl butyrate, methyl hexanoate, methyl octanoate, and ethyl formate. However, only 2-heptanone and 2-nonanone are the key aroma compounds that contribute significantly to the distinctive creamy flavor of blue cheese (High et al., 2020). The accurate identification of key aroma compounds is fundamental to understanding, controlling, and ultimately optimizing cheese sensory quality.

Methyl ketone compounds, particularly medium-chain methyl ketones such as 2-pentanone, 2-heptanone, and 2-octanone, contribute significantly to the creamy and fruity aromas in several cheese varieties. These compounds originate primarily from microbial metabolism and the β-oxidation pathway of fatty acids (S. Xu et al., 2025). During cheese ripening, lipases secreted by microorganisms such as Penicillium and Aspergillus molds hydrolyze triglycerides into free fatty acids. Subsequently, these fatty acids undergo reactions, including decarboxylation under the action of the microorganisms' β-oxidation enzyme system, yielding methyl ketones with one fewer carbon atom than the original fatty acids (Hawke J., 1966). For instance, caprylic acid undergoes β-oxidation to yield 2-heptanone, whilst decanoic acid produces 2-nonanone. These methyl ketones not only exhibit strong aromatic activity and distinctive odor characteristics, but they also often form the foundational “milky” or “creamy” flavor profile of cheese (Moio et al., 2020). As in blue cheese, 2-heptanone and 2-nonanone have been identified as key contributors to the characteristic creamy flavor, exhibiting significantly higher odor activity values (OAVs) than other aroma compounds (Caron et al., 2021). Similarly, in other mold-ripened cheeses such as Camembert, methyl ketones are essential for the overall flavor profile (Palermo et al., 2024). Therefore, since MC also involves the fermentation by molds such as Monascus, systematically assessing the composition, content, and flavor contribution of methyl ketones is vital for evaluating the creaminess of the cheese when analyzing its flavor profile.

Molecular sensory science plays a critical role in pinpointing the key aroma-active compounds within complex food matrices. This technological framework transcends the limitations of conventional chemical identification by systematically integrating instrumental analysis, sensory evaluation, and statistical analysis, enabling the precise identification of compounds that influence sensory attributes (Buyukkurt, 2024). This method involves non-targeted screening and quantification via gas chromatography-mass spectrometry (GC-MS), combined with OAV calculations for the preliminary identification of potential key compounds, followed by final confirmation through sensory omission, reconstruction, and addition experiments (J. Li et al., 2025; Yu et al., 2025; Zhu et al., 2025). This method has been extensively employed to assess the flavor of cheese and other food products. For instance, molecular sensory science has identified butyric acid, methyl mercaptan, ethyl acetoacetate, δ-lactone, furanone, and ethyl hexanoate, etc., as the principal compounds that distinguish between cheddar cheese samples at different ripening stages. This type of research provides crucial insights for the development and refinement of Cheddar cheese-related products (J. Wang et al., 2021). Furthermore, molecular sensory science has identified 15 key compounds in coffee, indicating that the aromatic profile undergoes significant transformation as the roast intensity increases. As the beans roast, the flavors shift from the grassy, cereal-like notes of green beans to the caramel, nutty, and toasted aromas characteristic of roasted coffee (He et al., 2025). Similarly, the primary aroma compounds commonly found in Sichuan hotpot oil include acetic acid, formaldehyde, (E)-2-octenal, (E, E)-2,4-dodecenal, and (E)-2-heptanal (Li et al., 2025, Li et al., 2025, Li et al., 2025). Moreover, stable isotope dilution analysis (SIDA) is widely recognized as the benchmark for precise quantification in these studies (Creek et al., 2012; Gallardo-Fernandez et al., 2022).

This study aims to investigate the systematic characterization of methyl ketone compounds in MC using a molecular sensory science approach. First, quantitative descriptive analysis (QDA) was used to track the changes in the creamy intensity throughout the cheese ripening process, while SPME-Arrow-GC-MS was employed for comprehensive spectral identification and preliminary quantification of its volatile constituents. Furthermore, Pearson's correlation analysis was used to quantify relationships between sensory attributes and flavor compounds. Next, the OAVs were calculated to assess the flavor contribution, followed by sensory validation via aroma reconstruction, deletion, and addition experiments. Finally, SIDA was performed to measure the key methyl ketone compounds accurately. These findings will provide essential information for an MC flavor chemistry database and a theoretical foundation for guiding the development of preferred flavor profiles by regulating fermentation and ripening processes.

2. Materials and methods

2.1. Chemicals and reagents

The n-alkanes (C7−C30, 98 %), 2-methyl-3-heptanone (99 %), 2-pentanone (99 %), 2,3-butanedione (99 %), 2-hexanone (98 %), 2-heptanone (≥98 %), 3-octanone (≥98 %), 2-octanone (≥98 %), 2-nonanone (≥99 %), 2-undecanone (99.5 %), acetophenone (99 %), 2-dodecanone (98 %), 2-tridecanone (99.9 %), 2-pentadecanone (95 %), and pulegone (≥90 %) were purchased from Sigma-Aldrich (St. Louis, MO). The [2H5]-2-heptanone (≥95 %), [2H5]-2-nonanone (≥95 %), and [2H5]-2-undecanone (≥95 %) were supplied by SHANGHAI ZZBIO Co. Ltd. (Shanghai, China), while the methanol (>99 %) and n-hexane were purchased from MERDA Technology Co. Ltd. (Beijing, China).

2.2. Cheese manufacturing

The MC was prepared according to a standardized protocol established in our previously published research (Y. Wang et al., 2025). To minimize experimental variation, strict control was maintained over the raw milk source, starter culture dosage, and processing conditions. The milk was first pasteurized at 65 °C for 30 min. It was then inoculated at 31 °C with 0.02 % (w/v) of a commercial freeze-dried starter culture (CHOOZIT MM100 Mesophilic Starter Culture, Danisco, Denmark) and M. purpureus M1 spore liquid (1.0 × 108 spores/mL), followed by acidification for 30 min. Subsequently, commercial rennet (CHR Hansen, Hoersholm, Denmark) was added at a concentration of 0.2 g/L, and coagulation proceeded at 31 °C for 45 min. The resulting curd was cut into cubes, whey was drained, and NaCl was added at 2 g per 100 g of curd. The cheese was then placed in a constant temperature and humidity incubator (LHS-80-II, Shanghai Yiheng Scientific Instrument Co., Ltd., China) at 26 °C and 90 % relative humidity for 5 d, followed by storage at 10 °C and 90 % relative humidity for 90 d. Sampling was conducted at four distinct ripening stages (0, 30, 60, and 90 d). Samples were taken in triplicate.

2.3. Quantitative descriptive analysis (QDA)

The overall flavor of the different MC samples was evaluated via QDA. Ethical approval for the involvement of human subjects in this study was granted by Beijing Technology and Business University Research Ethics Committee (reference no. 63 of 2024). Ten assessors (two males, eight females, average age 25) conducted descriptive sensory testing on the MC samples. Six of these were experienced in olfactory analysis, while the remaining four possessed expertise in the sensory attributes of dairy products. All assessors received training in the sensory evaluation of MC. Seven relevant aroma attributes were selected to describe the overall aroma profile of the cheese, including yogurt, creamy, milky-off, musty, sour, fruity, and fermented. Table S1 presents the definitions of the terms used. The assessors were instructed to rate the intensity of each attribute on a 10-point scale (0 = absent to 9 = very strong). The average results obtained from the sensory panel were ultimately plotted onto a sensory profile chart. Experiments were conducted at room temperature, approximately 23 °C.

2.4. Headspace solid-phase microextraction (SPME)-Arrow

A 6 g cheese sample was transferred to a 20 mL headspace vial. The sample was then spiked with 1 μL of an internal standard solution, which consisted of 2-methyl-3-heptanone at a concentration of 0.816 μg/μL (in n-hexane). The MC sample was equilibrated at 50 °C for 30 min, followed by insertion of the extraction head (DVB/CWR/PDMS, Supelco, Bellefonte, PA) into the headspace vial for adsorption for an additional 30 min.

2.5. Gas chromatography-mass spectrometry analysis (GC-MS)

The volatile aroma compounds were identified using a GC-MS-TQ8040 NX GC-MS system (GC-MS-TQ8040 NX, Shimadzu Corporation, Japan) coupled with an AOC-6000 multi-function injector (Shimadzu Corporation, Japan), while an SH-Polar Wax capillary column (60 m × 0.25 mm × 0.25 μm, Shimadzu Corporation, Japan) was employed for separation. Highly pure helium (purity ≥99.999 %) was used as a carrier gas at a flow rate of 1 mL/min. The qualitative data were acquired in full-scan mode (Q3Scan), while the chromatographic oven temperature started at 40 °C for 5 min, which was increased to 250 °C at 3 °C/min, where it was maintained for 15 min, with a total run time of 90 min. The other parameters included a solvent delay time of 4 min, an electron impact ionization source at 70 eV, an ion-source temperature of 200 °C, an inlet temperature of 250 °C, and a mass scan range of 33–550 m/z. Multiple reaction monitoring (MRM) was employed for quantitative data acquisition. The chromatographic oven temperature was set to 40 °C for 5 min, then increased to 250 °C at 3 °C/min, where it was maintained for 3 min, with a total run time of 78 min. The other parameters included solvent delay times of 6 min (hexane as solvent) and 12 min (methanol as solvent), and an electron impact ionization source at 70 eV, an ion-source temperature of 200 °C, an interface temperature of 250 °C, and a mass scan range of 35–400 m/z, utilizing the Shimadzu Aroma Compounds Database and NIST20 database. Each sample was analyzed in triplicate.

2.6. Gas chromatography-olfactometry (GC-O)

The GC-O system combined an Agilent GC7890B/MS5977A (Agilent Technologies, USA) and an ODP-3 (Gerstel GmbH, Germany), with sniff-port and transfer-line temperatures maintained at 230 °C and 250 °C, respectively. Moist air was introduced at the detection port to prevent the nasal passages of the evaluators from drying out and to mitigate olfactory fatigue. Three trained panel members conducted the olfactory assessment to ensure accurate identification of the odor-active compounds (Li et al., 2025, Li et al., 2025, Li et al., 2025). Each assessor participated in two 90-min evaluation sessions to prevent fatigue. During GC-O analysis, participants recorded aroma descriptors and the times at which they first appeared. The presence of an aroma was confirmed if two or more panel members detected the odor. Ultra-high purity nitrogen (99.999 %) served as the carrier gas in the GC-O analysis.

2.7. Qualitative analysis

Four methods were employed for compound identification. First, a search was performed against the Shimadzu Aroma Compound Database and the NIST 20 spectral library. Second, the retention indexes (RI) of the target compounds were calculated by comparing their GC retention times with those of n-alkanes C7-C30 in identical chromatographic conditions. Third, the results were compared with the sensory evaluation findings and, finally, verified against reference standards.

2.8. Quantitative analysis

An internal standard method was used for the preliminary quantitative analysis of the volatile compounds identified in the MC. The concentration of each identified substance was calculated as the ratio of its peak area to that of the internal standard (2-methyl-3-heptanone).

The 36 characteristic compounds were quantified again using external standard curves comprising at least five data points (with internal standard correction). Concentration gradients were established for each compound based on the semi-quantitative methyl ketone results and sequentially added to the MC simulation system, yielding standard curves comprising no fewer than five data points. Next, 2-methyl-3-heptanone was added to the simulated MC system as an internal standard. All standard curves were constructed by plotting the standard compound-to-internal standard peak area ratio against the corresponding concentration ratio. All analyses were performed in triplicate.

SIDA was used to reassess three key methyl ketone compounds: 2-heptanone, 2-nonanone, and 2-undecanone. Isotopically labeled mixed standard solutions were prepared using methanol as the isotopic internal standard, containing [2H5]-2-heptanone, [2H5]-2-nonanone, and [2H5]-2-undecanone, which were added to the MC matrix, along with the gradient-diluted solutions of the three compound standards. Standard curves for all compounds were constructed by plotting the peak area ratio of the standard to the stable-isotope internal standard against the concentration ratio. All analyses were performed in triplicate.

2.9. Threshold analysis

The simulated MC matrix was prepared by blending milk protein concentrate powder (MPC 70, Saishang Dairy Company, Ningxia, China) with ultrapure water, followed by pasteurization. This substrate was then coagulated via rennet incubation at 31 °C. After draining the whey, the resulting fresh curd was freeze-dried, ground into powder, and set aside. Subsequently, the powdered curd (protein phase, 25 %), deodorized sunflower oil (fat phase, 35 %), citrate sodium (emulsifier, 1.5 %), and ultrapure water (water phase, 38.5 %) were combined in proportions matching those of actual MC. To ensure homogeneous distribution of the internal standard, the final simulated MPC/water/oil mixture was homogenized by two 30-s cycles in a mixer.

The existing literature was consulted to ascertain the detection thresholds of the 13 methyl ketone compounds (Gemert, 2013; J. Wang et al., 2021). A concentration gradient range was established and used to prepare the standard samples, which were added to the base solution. These mixtures were equilibrated at ambient temperature for 10 min and used for olfactory analysis by 20 assessors with experience in evaluating dairy sensory attributes. All the sensory tests were performed at 23 ± 1 °C (room temperature) using a three-point test method. The samples were assigned three-digit codes, with each assessment employing a 10 g sample. Each group contained three samples, including two zero-group samples. The evaluators were required to identify the sample that differed from the other two. The final accuracy rate was calculated, and the statistical results were cross-referenced with a three-point test table to determine the final detection threshold.

2.10. Odor activity value (OAV) analysis

The OAVs of the methyl ketone compounds were calculated to determine their role in the cheese aroma. Compounds with OAV ≥1 were considered likely contributors to the aromatic profile of the sample. For each compound, the OAV is defined as the quotient of its concentration and its respective odor detection threshold. (Trabue et al., 2006).

2.11. Aroma recombination, omission, and addition experiments

Characteristic flavor compounds were selected and added to the simulated matrix at specified concentrations for the recombination test. The samples were equilibrated at ambient temperature for 10 min and prepared for sensory evaluation, as described in Section 2.3.

The omission experiments involved adding 36 characteristic flavor compounds at specified concentrations to simulated matrices. The samples were equilibrated at ambient temperature for 10 min and prepared for olfactory analysis by 20 assessors with experience in the sensory evaluation of dairy attributes. All sensory tests were conducted at 23 ± 1 °C (room temperature). Thirteen aroma-deficiency models were created, each missing one methyl ketone compound. The sensory evaluation employed a three-point test method where the samples were assigned three-digit codes, with each assessment using a 10 g sample. Each group comprised three samples, including two control samples. The evaluators were required to identify the sample that differed from the other two. The accuracy rates were calculated, and the results were cross-referenced with a three-point test table.

The addition experiment incorporated key compounds into the recombinant matrix and authentic samples at concentrations of 1 × , 10 × , 100 × , 1000 × , and 10000 × relative to the threshold. The overall aroma profile was characterized using eight attributes: yogurt, creamy, milky-off, musty, sour, fruity, fermented, and preference. The assessors were instructed to rate the intensity of these attributes on a scale from 0 (absent) to 9 (very strong). The average results of ten assessors were ultimately plotted as a sensory profile. All sensory tests were conducted at 23 ± 1 °C (room temperature).

2.12. Statistical analysis

Significant differences among the key aroma compounds were assessed using one-way analysis of variance (ANOVA), with post-hoc Duncan's multiple range test conducted at a 95 % confidence level. These analyses were performed using the Statistical Package for the Social Sciences (SPSS, version 23.0; SPSS Inc., Chicago, IL, USA). The OriginPro 2021 64-bit software (OriginPro Lab Corp., Northampton, America) was used for the sensory radar figures, while R (version 4.5.0, R Foundation for Statistical Computing, Vienna, Austria) and Gephi (version 0.10.1, The Gephi Consortium, Paris, France) were employed to plot the Pearson correlation and visualization results for sensory attributes and characteristic compounds.

3. Results and discussion

3.1. Sensory evaluation

The QDA values of the MC at different ripening stages indicated that its sensory attributes changed significantly over time. The radar chart in Fig. 1 highlighted this dynamic transformation process, in which fermented, fruity, creamy, milky-off, and sour aromas exhibited highly significant differences (p < 0.001), while musty aromas also reached a substantial level. This shift in sensory characteristics strongly suggests the generation and accumulation of one or more key aroma compounds in specific metabolic pathways during cheese ripening. According to existing research on cheese flavor chemistry, the sour in cheese primarily results from fatty acid compounds (Filho et al., 2021; Hassan and Gawad, 2012), while the fermented and fruity aromas are mainly derived from ethyl ester compounds (Y. Wang et al., 2024). The evolution of the characteristic creamy aroma is also typically closely associated with lipid-degradation metabolites, particularly methyl ketones, which possess a distinctive creamy flavor. Their potential formation and transformation may serve as the fundamental material basis for the development of the core flavor profile in the MC examined in this study.

Fig. 1.

Fig. 1

Open in a new tab

Sensory radar chart of Monascus-fermented cheese (MC) at different ripening stages (∗∗∗, highly significant (p < 0.001); ∗∗, extremely significant (p < 0.01); ∗, significant (p < 0.05)).

3.2. Qualitative analysis of the volatile compounds

To investigate the evolution of the aromatic profile during ripening, the volatile compounds in MC at 0, 30, 60, and 90 days were analyzed using SPME-Arrow-GC-MS. As shown in Fig. 2A, a total of 115 volatile compounds were detected, including 30 esters, 23 acids, 18 alcohols, 15 ketones (including 13 methyl ketones), 10 lactones, seven alkenes, five aromatics, four aldehydes, and three sulfur-containing compounds. Fig. 2B illustrates the semi-quantitative changes in the volatile compound concentrations. Overall, significant accumulation of acids, esters, and ketones was evident during the latter stages of cheese ripening, corroborating the pronounced sensory attribute alterations described above.

Fig. 2.

Fig. 2

Open in a new tab

Volatile compounds in Monascus-fermented cheese (MC). (A) Quantitative analysis of volatile compounds during the ripening of MC. (B) Semi-quantitative heatmap of volatile compounds during the ripening of MC.

3.3. Qualitative analysis of the characteristic compounds

To further evaluate the contribution of characteristic volatile compounds to the MC ripening process, we employed aroma extract dilution analysis (AEDA), determined flavor dilution (FD) factors, and described their odor characteristics. It is generally accepted that the FD factor of a volatile compound indicates its odor potency or detection threshold within the extract, serving as a primary criterion for screening potential key aroma compounds (Booth et al., 2024; Schulze et al., 2024). However, it should be noted that a higher FD factor reflects greater detectability upon dilution rather than directly quantifying its perceptual contribution in the final food matrix, which is further influenced by synergistic and masking effects. Table 1 presents the AEDA results for the MC at different ripening stages. A total of 36 volatile odor compounds with FD factors ≥81 (where the compound exhibited FD ≥ 81 at one or more stages) were identified as characteristic MC compounds. Acids and esters dominated during the onset of ripening (MC-0). Among the 12 detected compounds with FD ≥ 81, 2-heptanone (FD = 81) represented the only identified characteristic methyl ketone odor compound. By 30 d of ripening (MC-30), both the diversity and flavor contribution of the methyl ketones increased significantly. The FD factors of 2-nonanone (FD = 81) and 2-tridecanone (FD = 243) increased substantially, while 2-pentanone (FD = 81) also started to gain prominence. This may be attributed to the accelerated production of medium-to long-chain methyl ketones by Aspergillus rubrum. These ketones, as well as acids such as octanoic acid (FD = 729) and decanoic acid (FD = 729), exhibited extremely high FD values during this stage, contributing to a complex, rich flavor profile. However, methyl ketones are essential for creamy, fruity flavor characteristics. The flavor contribution of the methyl ketones reached a peak at 60 d of ripening (MC-60). The FD factor of 2-tridecanone surged to 729, establishing it as one of the most significant odor compounds during this stage. Concurrently, 2-hexanone (FD = 81), 2-heptanone (FD = 81), 3-octanone (FD = 81), 2-octanone (FD = 81), 2-nonanone (FD = 81), 2-undecanone (FD = 81), 2-dodecanone (FD = 81), and 2-tridecanone (FD = 729) formed a highly active methyl ketone flavor cluster. Notably, the FD values of the acids generally declined during this phase, while those for methyl ketones and esters displayed a notable rise. This negative correlation is consistent with possible conversion pathways reported in previous studies (Huang et al., 2025; Liao et al., 2024), suggesting that free fatty acids, which accumulated earlier, were converted into corresponding methyl ketones. These compounds serve as key indicators of lipid degradation and metabolism during cheese ripening. Although the FD values of certain methyl ketones, including 2-nonanone, declined by the late ripening stage (MC-90), compounds such as 2-heptanone (FD = 81), 2-octanone (FD = 81), and 2-undecanone (FD = 81) continued to display strong flavor activity, demonstrating sustained stability. During this phase, ketones emerged as key flavor contributors alongside acids, collectively defining the final flavor profile of MC at the end of the ripening process.

Table 1.

Identification of the characteristic compounds in Monascus-fermented cheese and the evolution of their flavor FD across the 90-day ripening stage.

No. Compounds CAS Odor FD
 RI Identification1
MC-0 MC-30 MC-60 MC-90
1 2-pentanone 107-87-9 fruity, creamy, woody – 81 27 27 974 MS/RI/O/S
2 2,3-butanedione 431-03-8 yogurt 81 – – – 999 MS/RI/O/S
3 2-hexanone 591-78-6 fruity – 3 81 – 1060 MS/RI/O/S
4 2-heptanone 110-43-0 fruity, creamy, herbal 81 81 81 81 1184 MS/RI/O/S
5 3-octanone 106-68-3 herbal, creamy, resin 3 – 81 3 1272 MS/RI/O/S
6 2-octanone 111-13-7 fruity, creamy, herbal – 9 81 81 1283 MS/RI/O/S
7 2-nonanone 821-55-6 fruity, creamy, herbal 27 81 81 27 1386 MS/RI/O/S
8 2-undecanone 112-12-9 orange, butter, green plant 27 9 81 81 1615 MS/RI/O/S
9 Pulegone 89-82-7 mint flavor – 81 – 3 1644 MS/RI/O/S
10 acetophenone 98-86-2 musty, floral, almond 81 9 27 27 1652 MS/RI/O/S
11 2-dodecanone 6175-49-1 fruity – – 81 – 1692 MS/RI/O/S
12 2-tetradecanone 593-08-8 creamy, earthy – 243 729 – 1817 MS/RI/O/S
13 2-pentadecanone 2345-28-0 jasmine flavor 81 9 3 9 2041 MS/RI/O/S
14 ethyl hexanoate 123-66-0 apple, peach 1 3 81 27 1613 MS/RI/O/S
15 ethyl octanoate 106-32-1 fruity, fermented – 81 27 81 1438 MS/RI/O/S
16 ethyl nonanoate 123-29-5 fruity, rose, nut – 81 1 27 1539 MS/RI/O/S
17 hexyl hexanoate 6378-65-0 fresh vegetable, fruity – – 27 9 1236 MS/RI/O/S
18 ethyl decanoate 110-38-3 grape, apple – 81 81 81 1641 MS/RI/O/S
19 ethyl dodecate 106-33-2 green plant – – 729 243 1846 MS/RI/O/S
20 acetic acid 64-19-7 sour 27 243 81 243 1461 MS/RI/O/S
21 butyric acid 107-92-6 rancid, pungent, sweat 729 729 729 729 1635 MS/RI/O/S
22 valeric acid 109-52-4 sweat, sour, cheesy – 81 243 81 1743 MS/RI/O/S
23 hexanoic acid 142-62-1 sweat, sour, cheesy 81 27 81 27 1851 MS/RI/O/S
24 octanoic acid 124-07-2 sweat, cheesy 243 729 729 243 2062 MS/RI/O/S
25 nonanoic acid 112-05-0 green plant, cheesy 81 729 243 27 2169 MS/RI/O/S
26 decanoic acid 334-48-5 sour 243 243 729 243 2276 MS/RI/O/S
27 undecanoic acid 112-37-8 petroleum, sour 27 27 9 81 2382 MS/RI/O/S
28 dodecanoic acid 143-07-7 Metallic, lemon 27 729 81 243 2487 MS/RI/O/S
29 3-methyl-butyric acid 503-74-2 sweat, sour, rancid – 243 27 81 1676 MS/RI/O/S
30 benzoic acid 65-85-0 sour, sulfur 81 81 9 81 2443 MS/RI/O/S
31 styrene 100-42-5 fragrant, petrol – 81 27 81 1261 MS/RI/O/S
32 benzaldehyde 100-52-7 almond, caramel – 1 81 3 1533 MS/RI/O/S
33 phenylethanol 60-12-8 Honey, rose, clove 81 243 27 81 1920 MS/RI/O/S
34 2,3-butanediol 513-85-9 fruity, onion 27 – – – 1578 MS/RI/O/S
35 δ-decalactone 705-86-2 coconut 81 243 729 243 2211 MS/RI/O/S
36 δ-dodecalactone 713-95-1 fruity, sweet 27 243 81 243 2441 MS/RI/O/S
Open in a new tab

Notes:1MS = Mass Spectrometry; RI=Retention Index; O = odor; S = standard compound.

“-” means not detected.

3.4. Quantification of the characteristic compounds

Although AEDA can effectively identify key volatile compounds, the perceived odor intensity shows a nonlinear relationship with compound concentration. The sample matrix can cause significant discrepancies between the perceived quality of a sample odor in real-world conditions and the odor attributes detected via GC-O (Mahmoud and Zhang, 2024; Plutowska and Wardencki, 2008). Therefore, reliable quantitative analytical results are required to further evaluate the significance of key volatile compounds.

After qualitative analysis, the 36 characteristic odor compounds were quantified using an external standard method with internal standard correction in MRM mode. Table S2 presents the corresponding calibration curves and quantitative results. The findings highlighted the dynamic patterns of variation in these characteristic compounds throughout the 90-d MC ripening stage. Most compound concentrations increased initially, then declined as ripening progressed, reaching peak levels at 60 d. This pattern was particularly pronounced among methyl ketone compounds. For instance, the 2-heptanone and 2-nonanone concentrations increased approximately 36- and 30-fold, respectively, from 0 d to 60 d, followed by a decline by 90 d. This closely paralleled the evolution pattern of their FD factors observed during AEDA analysis. Furthermore, esters such as ethyl decanoate and ethyl hexanoate, as well as certain acids, including butyric acid and valeric acid, exhibited significant accumulation. Collectively, these compounds contributed to the rich, foundational flavor of MC during its late ripening phase. These quantitative findings provide an essential basis for subsequent research by establishing correlations with sensory attributes, calculating OAVs to accurately assess individual compound contributions to the overall flavor profile, and ultimately identifying key methyl ketones.

3.5. Association analysis between key compounds and sensory attributes

Pearson correlation analysis was employed to examine the relationships between the specific MC sensory attributes and the 36 characteristic compounds to determine their contribution to the flavor profile. As shown in Fig. 3, 85 significant pairwise correlations were identified (|r|>0.8, p < 0.05), comprising 55 pairs exhibiting significantly positive correlations and 30 pairs displaying substantially negative correlations. It is important to note that correlation analysis indicates statistical co-variation rather than direct causal contribution, particularly in a complex, multicollinear aroma system where compounds may change concurrently due to shared ripening dynamics or metabolic pathways.

Fig. 3.

Fig. 3

Open in a new tab

Network visualization of relationships between characteristic volatile compounds and sensory attributes. Positive and negative correlations are depicted in red and blue, respectively. Nodes are coloured according to node attributes: sensory attributes appear purple, while characteristic compounds are blue.

Sour exhibited a significant positive correlation with acetic acid, 3-methylbutyric acid, nonanoic acid, pulegone, benzenemethanol, 2-pentanone, and 2-octanone (r > 0.8, p < 0.05), while showing a highly negative association with 2-pentadecanone, 2,3-butanedione, 2,3-butanediol, and benzoic acid (r < −0.8, p < 0.05). Short-chain fatty acids, such as acetic acid, 3-methylbutanoic acid, and nonanoic acid, inherently possess pronounced rancid and perspiration-like odors, which are sensorially linked to sourness. Conversely, the long-chain methyl ketone 2-pentadecanone often carries waxy and floral notes, while 2,3-butanedione, which imparts rich, creamy aromas, shows a negative correlation with sour. This inverse statistical relationship suggests that their sensory profiles may be perceptually distinct within the cheese matrix. Increased concentrations of compounds providing sweet and milky aromas may co-occur with a reduction in the perceived intensity of sour, highlighting the complexity of the flavor system.

The fermented aromas exhibited a significant positive correlation with several ethyl esters and acids (r > 0.8, p < 0.05). These ethyl esters, particularly those from medium-chain fatty acids, are known for pleasant fermented and fruity aromas. Their accumulation in the MC primarily results from esterification reactions during enzymatic catalysis (Y. Xu et al., 2020). The strong positive correlation observed is consistent with the established role of esters in fermented aromas and reflects the metabolic activity during Monascus fermentation.

Fruity exhibited a significant positive correlation with compounds such as 2-tridecanone, pulegone, and δ-decalactone (r > 0.8, p < 0.05), while showing a substantial negative association with 2,3-butanediol, benzoic acid, nonanoic acid, and 2,3-butanedione (r < −0.8, p < 0.05). This pattern of association aligns with the reported fruity characteristics of certain lactones (Chen et al., 2022; Syed et al., 2023) and methyl ketones. The yogurt flavor displayed significant positive correlations (r > 0.8, p < 0.05) with 2,3-butanedione, 2,3-butanediol, 2-pentadecanone, acetophenone, nonanoic acid, and benzoic acid. These compounds are commonly associated with rich, fermented dairy flavors in food products. Furthermore, the yogurt flavor exhibited a highly negative association with 2-pentanone, 2-octanone, 3-methyl-butyric acid, pulegone, benzenemethanol, and 2-tridecanone (r < −0.8, p < 0.05).

The milky-off odor exhibited a significant positive association (r > 0.8, p < 0.05) with hexanoic acid, butyric acid, ethyl hexanoate, ethyl octanoate, benzylethanol, ethyl nonanoate, benzaldehyde, ethyl decanoate, pulegone, 3-methylbutyric acid, valeric acid, and 2-tridecanone. These compounds are sensorially associated with milky aroma at low concentrations but can present unpleasant sweat-like or rancid odors at elevated levels. Furthermore, the milky-off odor exhibited a significant negative correlation with benzoic acid, 2,3-butanedione, and 2,3-butanediol (r < −0.8, p < 0.05). The inverse statistical relationship suggests that higher levels of these negatively correlated compounds co-occur with lower perceived intensity of milky-off odors within the tested samples. The musty odor exhibited a significant positive correlation with pulegone, 2-tridecanone, 3-methylbutyric acid, and ethyl decanoate (r > 0.8, p < 0.05), while showing a substantial negative association with δ-decalactone, nonanoic acid, acetophenone, 2,3-butanedione, 2,3-butanediol, and benzoic acid (r < −0.8, p < 0.05). These positively correlated compounds are known to impart moldy, earthy, and other aged flavor characteristics, and their accumulation co-occurs with the manifestation of musty notes, consistent with the reported metabolic activity of Monascus purpureus.

Creamy exhibited a significant positive correlation with several methyl ketones (2-undecanone, 2-pentanone, 2-nonanone, 2-heptanone, 2-octanone, 2-dodecanone) as well as ethyl dodecanoate and benzylethanol (r > 0.8, p < 0.05). This statistical association identifies methyl ketones as key compounds of interest for further investigation regarding the creamy aroma. These medium-chain methyl ketones primarily originate from lipid β-oxidation pathways. Their distinctive dairy and fruity notes are characteristic of creamy flavor profiles. The co-correlation of non-methyl ketone compounds such as ethyl dodecanoate and phenethyl alcohol suggests that the overall creamy perception may arise from a combination of multiple compound classes.

Correlation analysis indicated that multiple aroma compounds co-varied with the sensory characteristics of the MC in complex patterns of positive and negative associations. Among these, the strong correlation between methyl ketones and creaminess was particularly notable. To move beyond correlation and establish causal contribution, subsequent research in this study focused on methyl ketones through threshold analysis, OAV calculations, and systematic sensory validation experiments (including aroma recombination, omission, and addition experiments). This approach aimed to identify key drivers of creaminess from a flavor-chemistry perspective.

3.6. OAV analysis of the characteristic methyl ketone compounds

The impact of aroma compounds on the overall fragrance of cheese depends on their concentrations and sensory thresholds. It is generally accepted that a compound is considered potentially contributory when OAV ≥1, although synergistic effects may occur below threshold. Therefore, this study calculated the OAVs of 13 methyl ketones in the MC. The thresholds listed in Table 2 were derived from determinations conducted using a simulated cheese matrix. The relevant data are provided in Table S3.

Table 2.

Threshold and OAV values of 13 methyl ketone compounds in Monascus-fermented cheese.

No. Compounds CAS Threshold (μg/kg) OAV
MC-0 MC-30 MC-60 MC-90
1 2-pentanone 107-87-9 800.00 – 0.95 2.52 1.74
2 2,3-butanedione 431-03-8 41.00 0.43 – – –
3 2-hexanone 591-78-6 0.82 – 1.00 1.63 –
4 2-heptanone 110-43-0 164.00 2.87 49.85 106.36 33.58
5 3-octanone 106-68-3 6.20 3.28 – 3.26 3.27
6 2-octanone 111-13-7 0.82 – 9.4 21.03 5.38
7 2-nonanone 821-55-6 0.164 1.34 16.44 40.84 11.20
8 2-undecanone 112-12-9 0.082 20.72 53.56 213.09 73.52
9 acetophenone 98-86-2 3.28 7.71 7.11 7.40 7.36
10 2-dodecanone 6175-49-1 8.20 – – 7.06 –
11 2-tridecanone 593-08-8 0.82 – 36.51 36.35 31.84
12 2-pentadecanone 2345-28-0 8.20 6.01 2.31 2.41 5.28
13 pulegone 89-82-7 0.82 – 1.51 1.50 1.50
Open in a new tab

Analysis of Table 2 indicated that the flavor contribution of the methyl ketone compounds increased with the cheese ripening. At the beginning of the ripening phase (MC-0), six methyl ketones already exhibited OAV≥1, including 2-heptanone, 3-octanone, 2-nonanone, 2-undecanone, acetophenone, and 2-pentadecanone. Of these, 2-undecanone (OAV = 20.72) and 2-heptanone (OAV = 2.87) exhibited significant flavor activity. Nine methyl ketone compounds exhibited OAV≥1 at 30 d of cheese ripening (MC-30), including 2-hexanone, 2-heptanone, 2-octanone, 2-nonanone, 2-undecanone, acetophenone, 2-tridecanone, 2-pentadecanone, and pulegone. The 60-d ripening cheese (MC-60) contained twelve methyl ketone compounds with OAV≥1, namely 2-pentanone, 2-hexanone, 2-heptanone, 3-octanone, 2-octanone, 2-nonanone, 2-undecanone, acetophenone, 2-dodecanone, 2-tridecanone, 2-pentadecanone, and pulegone. By the final stage of ripening (MC-90), the number of active compounds decreased slightly to 10 and included 2-pentanone, 2-heptanone, 3-octanone, 2-octanone, 2-nonanone, 2-undecanone, acetophenone, 2-tridecanone, 2-pentadecanone, and pulegone. However, the OAV values of key compounds remained elevated, indicating their sustained contribution to the flavor profile. Throughout ripening, 2-heptanone, 2-nonanone, and 2-undecanone consistently exhibited high OAV values, peaking at 106.36, 40.84, and 213.09, respectively, at MC-60. This indicated that these three medium-chain methyl ketones make a prominent and stable contribution to the overall aroma of MC. Their OAV values initially increased, then declined as ripening progressed, suggesting a potential association with the metabolic activity of Aspergillus rubrum and dynamic changes in fatty acid precursors. Vigorous microbial β-oxidation generated substantial methyl ketone formation during the mid-ripening phase, while the possible concentration decline in later stages could be ascribed to further conversion or volatilization.

3.7. Sensory verification of the aroma contributions by the characteristic methyl ketone compounds

This study conducted a series of aroma reconstruction, omission, and addition experiments to effectively determine the role of methyl ketone compounds in shaping the MC flavor profile, particularly the key compounds identified via OAV analysis.

First, 36 previously identified characteristic odor compounds were incorporated into the simulated MPC/water/oil matrix, successfully establishing aroma recombination models at different MC ripening stages (0 d, 30 d, 60 d, and 90 d). The sensory evaluation results (Fig. 4) indicated that each recombination model closely mirrored the overall aroma profile of its corresponding authentic MC sample. This confirmed that the 36 compounds identified and quantified in the preliminary phase of this study via SPME-Arrow-GC-MS and GC-O analysis could reproduce the complex primary aromas of MC, establishing a reliable foundation for subsequent research into the aroma contributions of specific methyl ketones.

Fig. 4.

Fig. 4

Open in a new tab

Sensory comparison between Monascus-fermented cheese (MC) samples and reconstituted samples (0, 10, 40, 90 denote MC samples aged for 0, 10, 40, and 90 days respectively; 0-CZ, 10-CZ, 40-CZ, 90-CZ denote the corresponding reconstituted samples. ∗∗∗, highly significant (p < 0.001); ∗∗, extremely significant (p < 0.01); ∗, significant (p < 0.05)).

Omission experiments were conducted in the recombination model to accurately evaluate the flavor contribution of individual methyl ketones. This approach employed a three-point test to determine whether omitting specific methyl ketone compounds resulted in sensory-detectable differences between the basic and complete recombination models. The results are presented in Table 3. Statistical analysis indicated that omitting 2-heptanone, 2-nonanone, and 2-undecanone produced extremely significant (p ≤ 0.01) or highly significant (p ≤ 0.001) sensory differences during most ripening stages (three or more out of four stages). Notably, omitting both 2-heptanone and 2-nonanone at 60 d and 90 d of ripening yielded highly significant differences. This demonstrated their indispensable role in shaping the overall flavor profile of MC, particularly during the mid-to late-ripening stages. The sustained and significant contribution of these three methyl ketone compounds, coupled with their consistently high OAVs, highlighted their considerable influence on the MC flavor profile.

Table 3.

Omission results of 13 methyl ketones in Monascus-fermented cheese.

No. Compounds Samples
 Significant
MC-0 MC-30 MC-60 MC-90 MC-0 MC-30 MC-60 MC-90
1 2-pentanone – 8 11 11 – – ∗ ∗
2 2,3-butanedione 11 – – – ∗ – – –
3 2-hexanone – 7 11 – – – ∗ –
4 2-heptanone 11 12 13 15 ∗ ∗∗ ∗∗ ∗∗∗
5 3-octanone 4 – 10 12 – – – ∗
6 2-octanone – 12 11 11 – ∗ ∗ ∗
7 2-nonanone 12 12 14 13 ∗ ∗ ∗∗∗ ∗∗
8 2-undecanone 11 14 12 11 ∗ ∗∗∗ ∗ ∗
9 acetophenone – 10 12 10 – – ∗ –
10 2-dodecanone – 10 11 10 – – ∗ –
11 2-tridecanone – – 6 – – – – –
12 2-pentadecanone – 12 10 11 – ∗ – ∗
13 pulegone 9 12 11 10 – ∗ ∗ –
Open in a new tab

Notes.

∗∗∗, highly significant (p < 0.001); ∗∗, extremely significant (p < 0.01); ∗, significant (p < 0.05)). “-“Not significant.

To further investigate the dose-response relationship of these three methyl ketones and their impact on sensory preference, this study conducted additional experiments in the MC-60 simulation system, which exhibited the highest methyl ketone diversity. As illustrated in Fig. 5, variations in the 2-heptanone, 2-nonanone, and 2-undecanone concentrations had a significant regulatory effect on the flavor profile. When the added concentration reached 100 times its threshold level, the intensity of creaminess and fruitiness peaked, resulting in the highest overall preference. However, as the concentrations increased further, the intensity of these pleasant aromas declined, accompanied by the emergence of an unpleasant pungent note. The addition test results clearly revealed the dual role of these three methyl ketones in MC flavor. At appropriate concentrations, they are key contributors to the ideal creamy flavor profile. However, at excessive levels, they may lead to flavor defects.

Fig. 5.

Fig. 5

Open in a new tab

Sensory evaluation results for 2-heptanone (A, D), 2-nonanone (B, E) and 2-undecanone (C, F) added at threshold multiples to the original sample (MC-60) and matrix (CZ-60). (∗∗∗, highly significant (p < 0.001); ∗∗, extremely significant (p < 0.01); ∗, significant (p < 0.05)).

In summary, preliminary screening through OAV analysis, followed by thorough validation of the model system reliability employing aroma reconstruction experiments, allowed the confirmation of key contributions and dose effects via sensory evaluation using omission and addition experiments. This study conclusively identified 2-heptanone, 2-nonanone, and 2-undecanone as the most critical characteristic methyl ketones during the ripening process of MC.

3.8. SIDA of the key methyl ketone compounds

The matrix effect may affect the quantification accuracy of the external standard method (Nedele et al., 2021). SIDA is a quantitative technique prized for its speed, sensitivity, high accuracy, and precision. These attributes, coupled with its capacity to mitigate matrix effects and improve analyte recovery, make it particularly suitable for analyzing complex food matrices and conducting trace-level determinations. This effectiveness stems from the use of isotopically labeled internal standards, which share nearly identical physicochemical properties with the target analytes (San et al., 2016). The three key methyl ketone compounds were re-quantified using stable isotopes. Table 4 shows the calibration curves and precise quantitative results. Similar to the results obtained via the external standard method, 2-nonanone was the most abundant compound among all the analytes in the MC, followed by 2-heptanone and 2-undecanone. However, consistent with previous studies, the specific concentrations of each compound varied between the results obtained via the external standard method and those derived from SIDA calculations (Armas et al., 2022; Li et al., 2023). Notably, the 2-nonanone concentration identified via SIDA in the quantitative MC matrix was approximately three times higher than that determined using the external standard method. These variations can be attributed to several factors, a key one being the inherent limitations of the external standard method. Unlike SIDA, its quantification relies on a standard matrix that may not fully replicate the specific binding affinity between volatiles and the actual MC matrix, potentially compromising precision.

Table 4.

Concentration of key methyl ketone compounds in Monascus-fermented cheese based on stable isotope quantification.

Compounds CAS Standard curve R2 Spike recovery rate % Concentration (μg/kg)a
 Selected ions (m/z)b
MC-0 MC-30 MC-60 MC-90
2-Heptanone 110-43-0 y = 0.631x+3.3977 0.9974 100.97 969.12 ± 17.68 4836.71±
887.13		 8985.04±
320.59		 2419.99 ± 52.88 63.00 > 46.00–119 0.00 > 46.00
2-Nonanone 821-55-6 y = 0.9769x+1.7468 0.9977 112.31 563.46 ± 29.13 16923.05±
7183.37		 19981.15 ± 12789.96 1240.34 ± 143.85 63.00 > 46.00–103.00 > 75.00
2-Undecanone 112-12-9 y = 0.578x+0.2253 0.9973 117.51 114.15 ± 16.31 550.35±
143.41		 2018.08±
100.15		 599.96±
22.03		 58.00 > 43.00–147.00 > 58.00
Open in a new tab

Notes.

a

Average value from three parallel experiments.

b

The ion pair preceding ‘-’ is quantitative, ‘-’ denotes qualitative ion pairs, the ion before ‘>’ is the parent ion, and the ion after ‘>’ is the daughter ion.

4. Conclusion

This study employed molecular sensory science techniques to systematically analyze the composition, content changes, and flavor contribution of methyl ketone compounds in MC during its 90-d aging process. First, QDA was employed to reveal the notable changes in the creamy aroma intensity of the MC during ripening. SPME-Arrow-GC-MS was used to select 36 characteristic compounds from a total of 115 volatile compounds, which included 13 methyl ketones. Aroma reconstruction experiments confirmed that these selected compounds accurately represented the overall aroma profile. Concurrent sensory omission experiments and OAV analysis indicated that 2-heptanone, 2-nonanone, and 2-undecanone were essential for establishing the characteristic creamy aroma of the cheese. Additional experiments revealed the dose-response relationship of these three compounds. At optimal concentrations, they significantly enhanced the creamy aroma, while excessive quantities produced off-flavors. The precise quantification of the three key methyl ketones via SIDA provided a reliable method for accurately determining aroma compounds in MC. This study elucidates the key chemical basis of the creamy aroma in MC and provides mechanistic insights that may inform future flavor regulation strategies. It should be noted that the cheese samples analyzed in this study were obtained from a single manufacturing batch sampled over time. While the manufacturing process was strictly standardized to ensure stability, and analyses were performed in triplicate, the lack of independent biological replicates across multiple batches remains a limitation. Future studies should include multiple independent batches to further validate the consistency of Monascus metabolism and enzyme activity variations reported here.

CRediT authorship contribution statement

Yadong Wang: Formal analysis, Investigation, Methodology, Writing—original draft, and editing. Ying Wang: Writing—review and editing. Ting Meng: Writing—review and editing. Bei Wang: Supervision, Writing—review and editing, Funding. Yanbo Wang, Supervision.

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.

Acknowledgment

This work was supported by the National Natural Science Foundation of China (project No.32572740) Beijing, and the Excellent Doctoral Dissertation Cultivation Plan Project of Beijing Technology and Business University in 2025.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.crfs.2026.101328.

Abbreviations

MC, Monascus-fermented cheese; QDA, Quantitative descriptive analysis; GC-MS, Gas chromatograph-mass spectrometer; AEDA, Aroma extract dilution analysis; OAV, Odor activity values; SPME arrow, Solid-phase microextraction Arrow; SIDA, Stable isotope dilution analysis; GC-O, Gas chromatography-olfactometry.

Appendix A. Supplementary data

The following is the supplementary data to this article:

Multimedia component 1
mmc1.docx (36.2KB, docx)

References

  1. Armas Y.C.d., Herrera U.M.G., Serrano I.R., Buhay W.M., Skelton S., Suárez R.R., Roksandic M., Laffoon J. Multiproxy paleodietary reconstruction using stable isotopes and starch analysis: the case of the archaeological site of Playa del Mango, Granma, Cuba. J. Archaeol. Sci.: Report. 2022;46 doi: 10.1016/j.jasrep.2022.103671. [DOI] [Google Scholar]
  2. Booth M., Tello E., Peterson D.G. Characterization and profiling of key odorants in roasted hybrid american-european hazelnuts. J. Agric. Food Chem. 2024;72(38):21144–21151. doi: 10.1021/acs.jafc.4c05302. [DOI] [PubMed] [Google Scholar]
  3. Buyukkurt O.K. Application of molecular sensory analysis in determining food flavor: a review. J. Raw Mater. Processed Foods. 2024;5(1):1–10. doi: 10.57252/jrpfoods.2024.1/20. [DOI] [Google Scholar]
  4. Caron T., Piver M.L., Péron A.-C., Lieben P., Lavigne R., Brunel S., Roueyre D., Place M., Bonnarme P., Giraud T. Strong effect of Penicillium roqueforti populations on volatile and metabolic compounds responsible for aromas, flavor and texture in blue cheeses. Int. J. Food Microbiol. 2021;354 doi: 10.1016/j.ijfoodmicro.2021.109174. [DOI] [PubMed] [Google Scholar]
  5. Chen C., Liu Z., Yu H., Xu Z., Tian H. Flavoromic determination of lactones in cheddar cheese by GC-MS-olfactometry, aroma extract dilution analysis, aroma recombination and omission analysis. Food Chem. 2022;368 doi: 10.1016/j.foodchem.2021.130736. [DOI] [PubMed] [Google Scholar]
  6. Creek D.J., Chokkathukalam A., Jankevics A., Burgess K.E.V., Breitling R., Barrett M.P. Stable isotope-assisted metabolomics for network-wide metabolic pathway elucidation. Anal. Chem. 2012;84(20):8442–8447. doi: 10.1021/ac3018795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Filho M.D.J., Klein B., Wagner R., Godoy H. Key aroma compounds of canastra cheese: HS-SPME optimization assisted by olfactometry and chemometrics. Food Res. Int. 2021;150(Pt A) doi: 10.1016/j.foodres.2021.110788. [DOI] [PubMed] [Google Scholar]
  8. Gallardo-Fernandez M., Valls-Fonayet J., Valero E., Hornedo-Ortega R., Richard T., Troncoso A.M., Garcia-Parrilla M.C. Isotopic labelling-based analysis elucidates biosynthesis pathways in Saccharomyces cerevisiae for melatonin, serotonin and hydroxytyrosol formation. Food Chem. 2022;374(16) doi: 10.1016/j.foodchem.2021.131742. [DOI] [PubMed] [Google Scholar]
  9. Gemert L.J.V. 2013. Compilations of Odour Threshold Values in Air, Water and Other Media. [Google Scholar]
  10. Hassan F.A.M., Gawad M.A.M.A.E.-. Flavour compounds in cheese. Int. J. Acad. Res. 2012;4(5):169–181. doi: 10.7813/2075-4124.2012/4-5/A.20. (review) [DOI] [Google Scholar]
  11. Hawke J.C. Section D. Dairy chemistry. The formation and metabolism of methyl ketones and related compounds. J. Dairy Res. 1966;33(2):225–243. [Google Scholar]
  12. He W., Chen X., Sun J., Chen H., Sun B. Unveiling key aroma compounds formation during the roasting process of Yunnan arabica coffee using molecular sensory science approaches. Food Chem. 2025;494 doi: 10.1016/j.foodchem.2025.145853. [DOI] [PubMed] [Google Scholar]
  13. High R., Eyres G.T., Bremer P., Kebede B. Characterization of blue cheese volatiles using fingerprinting, self-organizing maps, and entropy-based feature selection. Food Chem. 2020;347(15) doi: 10.1016/j.foodchem.2020.128955. [DOI] [PubMed] [Google Scholar]
  14. Huang W., Gumuly Y., Wätjen A.P., Marcellin E., Prakash S., Bang-Berthelsen C.H., Turne M.S. Enhancing dairy-like aromas and reducing beany flavours in mould-ripened soy cheeses. Int. J. Food Microbiol. 2025;443 doi: 10.1016/j.ijfoodmicro.2025.111408. [DOI] [PubMed] [Google Scholar]
  15. Kumura H., Ohtsuyama T., Matsusaki Y.H., Taitoh M., Koyanagi H., Kobayashi K., Hayakawa T., Wakamatsu J.I., Ishizuka S. Application of red pigment producing edible fungi for development of a novel type of functional cheese. J. Food Process. Preserv. 2018;42(10) doi: 10.1111/jfpp.13707. e13707.13701-e13707.13708. [DOI] [Google Scholar]
  16. Li J., Lu T., Wang Y., Tang M., Ma T., Awasthi M.K., Sun X., Fang Y. Sensory-directed flavor decoding: key aroma compounds determination in new Chinese apple cultivars using GC × GC-QTOFMS approach. Food Chem. 2025;495(3) doi: 10.1016/j.foodchem.2025.146662. [DOI] [PubMed] [Google Scholar]
  17. Li X., Zhang C., Li S., Wang H., Yu P., Shao H., ShenhaoWang, Wang H., Jin F. Unveiling key aroma-active compounds of two distinct aroma-types oriental melon: an integrated sensomics approach utilizing GC×GC-O-TOF-MS, GC–MS, aroma recombination and omission experiments. Food Chem. X. 2025;28 doi: 10.1016/j.fochx.2025.102517. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Li X., Zhang W., Zeng X., Xi Y., Li Y., Hui B., Li J. Characterization of the major odor-active off-flavor compounds in normal and lipoxygenase-lacking soy protein isolates by sensory-directed flavor analysis. J. Agric. Food Chem. 2023;71(21):8129–8139. doi: 10.1021/acs.jafc.3c00793. [DOI] [PubMed] [Google Scholar]
  19. Li X., Zhang Z., Li B., Luo Q., Chen Q., Chitrakar B., Wu G., Huang L. Characterization of key odorants and their content variation in sichuan hotpot oil with different chili pepper cultivars by means of molecular sensory science. Food Sci. Nutr. 2025;13(11) doi: 10.1002/fsn3.71098. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Liao G.Q., Han H.L., Wang T.C., Li H.R., Qian Y.Z., Zhu M.X., Jia Q., Qiu J. Comparative analysis of the fatty acid profiles in goat milk during different lactation periods and their interactions with volatile compounds and metabolites. Food Chem. 2024;460(Part1) doi: 10.1016/j.foodchem.2024.140427. [DOI] [PubMed] [Google Scholar]
  21. Luo H., Akkermans S., Verheyen D., Wang J., Polanska M., Van Impe J.F.M. Tuning and modeling cheese flavor. Compr. Rev. Food Sci. Food Saf. 2024;23(5) doi: 10.1111/1541-4337.13420. [DOI] [PubMed] [Google Scholar]
  22. Mahmoud M.A.A., Zhang Y. Enhancing odor analysis with gas chromatography–olfactometry (GC-O): recent breakthroughs and challenges. J. Agric. Food Chem. 2024;72(17):9523–9554. doi: 10.1021/acs.jafc.3c08129. [DOI] [PubMed] [Google Scholar]
  23. Martin C., Harel-Oger M., Garric G., Marette S. Impact of sensory properties and their appreciation on willingness to pay for innovative cheeses with health benefits. Food Qual. Prefer. 2024;118 doi: 10.1016/j.foodqual.2024.105207. [DOI] [Google Scholar]
  24. Moio L., Piombino P., Addeo F. odor-impact compounds of Gorgonzola cheese. J. Dairy Res. 2020;67(2):273–285. doi: 10.1017/S0022029900004106. [DOI] [PubMed] [Google Scholar]
  25. Nedele A.-K., Mayer N., Feller N., Hinrichs J., Zhang Y. Off-flavor in soy drink: development, optimization, and validation of an easy and fast method to quantify the key odorants. Talanta: The Inter. J. Pure and Appl. Analytical Chem. 2021;229 doi: 10.1016/j.talanta.2021.122251. [DOI] [PubMed] [Google Scholar]
  26. Ning H., Yang A., Zhou W., Li S., Tang Y., Ma W., Liu P. Key volatile off-flavor compounds identification in chicken meat and deodorizing effects of polyphenol and spice extracts. Food Chem. X. 2025;29 doi: 10.1016/j.fochx.2025.102659. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Palermo C., Mentana A., Tomaiuolo M., Campaniello M., Iammarino M., Centonze D., Zianni R. Headspace solid-phase microextraction/gas chromatography–mass spectrometry and chemometric approach for the study of volatile profile in X-ray irradiated surface-ripened cheeses. Foods. 2024;13(3):416. doi: 10.3390/foods13030416. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Plutowska B., Wardencki W. Application of gas chromatography–olfactometry (GC–O) in analysis and quality assessment of alcoholic beverages – a review. Food Chem. 2008;107(1):449–463. doi: 10.1016/j.foodchem.2007.08.058. [DOI] [Google Scholar]
  29. San A.T., Joyce D.C., Hofman P.J., Macnish A.J., Webb R.I., Matovic N.J., Williams C.M., Voss J.J.D., Wong S.H., Smyth H.E. Stable isotope dilution assay (SIDA) and HS-SPME-GCMS quantification of key aroma volatiles for fruit and sap of Australian mango cultivars. Food Chem. 2016;221:613–619. doi: 10.1016/j.foodchem.2016.11.130. [DOI] [PubMed] [Google Scholar]
  30. Schulze L.J., Beier R., Hartmann B., Krammer G.E. Molecular-sensory decoding of the citrus latifolia aroma. J. Agric. Food Chem. 2024;72(26):14874–14886. doi: 10.1021/acs.jafc.4c02059. [DOI] [PubMed] [Google Scholar]
  31. Syed N., Singh S., Chaturvedi S., Nannaware A.D., Khare S.K., Rout P.K. Production oflactonesfor flavoring and pharmacological purposes from unsaturated lipids: an industrial perspective. Crit. Rev. Food Sci. Nutr. 2023;63(29):10047–10078. doi: 10.1080/10408398.2022.2068124. [DOI] [PubMed] [Google Scholar]
  32. Trabue S.L., Anhalt J.C., Zahn J.A. Bias of tedlar bags in the measurement of agricultural odorants. J. Environ. Qual. 2006;35(5):1668–1677. doi: 10.2134/jeq2005.0370. [DOI] [PubMed] [Google Scholar]
  33. Wang J., Yang Z., Xu L., Wang B., Zhang J., Li B., Cao Y., Tan L. Key aroma compounds identified in cheddar cheese with different ripening times by aroma extract dilution analysis, odor activity value, aroma recombination, and omission. J. Dairy Sci. 2021;104(2):1576–1590. doi: 10.3168/jds.2020-18757. [DOI] [PubMed] [Google Scholar]
  34. Wang Y., Zeng H., Qiu S., Han H., Wang B. Identification of key aroma compounds and core functional microorganisms associated with aroma formation for Monascus-fermented cheese. Food Chem. 2024;434 doi: 10.1016/j.foodchem.2023.137401. [DOI] [PubMed] [Google Scholar]
  35. Wang Y., Zhang Q., Yousaf J., Meng F., Wang Y., Wang B. Analysis and comparison of lipids in monascus-fermented cheese from different ripening periods based on UHPLC-QTRAP MS quantitative lipidomics. J. Agric. Food Chem. 2025;73(7):4385–4395. doi: 10.1021/acs.jafc.4c10355. [DOI] [PubMed] [Google Scholar]
  36. Weimer B.C. Woodhead Publishing Limited: Improving the flavor of cheese; 2007. Improving the Flavor of Cheese. [Google Scholar]
  37. Wu S., Yu H., Liu Z., You C. Influence of Monascus purpureus BD-M-4 on the physicochemical properties, proteolysis and volatile compounds of surface mould-ripened cheese. Food Sci. Biotechnol. 2019;(1):129–138. doi: 10.1007/s10068-018-0459-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Xu S., Zhang Q., Dong G., Liu Z., Fu J., Shi S. Advances in the microbial biosynthesis of methyl ketones. Renew. Sustain. Energy Rev. 2025;208 doi: 10.1016/j.rser.2024.115038. [DOI] [Google Scholar]
  39. Xu Y., Wang X., Liu X., Li X., Zhang C., Li W., Sun X., Wang W., Sun B. Discovery and development of a novel short-chain fatty acid ester synthetic biocatalyst under aqueous phase from Monascus purpureus isolated from Baijiu. Food Chem. 2020;338 doi: 10.1016/j.foodchem.2020.128025. [DOI] [PubMed] [Google Scholar]
  40. Yang C., Zhang Y., Jia M., Qu Q., Li Z., Zhou Q., Feng M., Tian Y., Ren W., Shi X. Fermentation interactions and lipid-lowering potential of Monascus purpureus fermented Ginseng. Food Biosci. 2025;66 doi: 10.1016/j.fbio.2025.106260. [DOI] [Google Scholar]
  41. Yu P., Zheng Z., Wang M., Luo Q., Lao F., Ren K., Dong W., Wu J. Characterization of key aroma-active compounds in onion seasoning oil via sensory-directed flavor analysis and gas chromatographysulfur chemiluminescence detection. Food Res. Int. 2025;221(Part2) doi: 10.1016/j.foodres.2025.117292. [DOI] [PubMed] [Google Scholar]
  42. Zeng H., Wang Y., Han H., Cao Y., Wang B. Changes in key aroma compounds and esterase activity of monascus-fermented cheese across a 30-Day ripening period. Foods. 2022;11(24) doi: 10.3390/foods11244026. Article 4026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Zhang C., Cheng Y., Qin Y., Wang C., Wang H., Ablimit A., Sun Q., Dong H., Wang B., Wang C. Occurrence, risk implications, prevention and control of CIT in monascus cheese: a review. J. Agric. Food Chem. 2024;72(17):9567–9580. doi: 10.1021/acs.jafc.4c00588. [DOI] [PubMed] [Google Scholar]
  44. Zhang W., Chen C., Li Y., Guo F., Liu W., Liu S., sun Y., Wang X., Shen Y., Wang P. Analysis of composition and source of the key aroma compounds in stir-fried pepper tallow. Food Chem. 2024;441 doi: 10.1016/j.foodchem.2023.138321. [DOI] [PubMed] [Google Scholar]
  45. Zhu L., Lin Z., Zheng Y., Liang J., Li Y., Kramp S., Zhang Y., Xiang C., Chen L., Rigling M., Hannemann L., Oellig C., Zhang Y. Decoding the aroma of Jägermeister liqueur through sensory-directed flavor analysis combined with solvent-assisted flavor evaporation and headspace-stir bar sorptive extraction. LWT. 2025;230 doi: 10.1016/j.lwt.2025.118265. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Multimedia component 1
mmc1.docx (36.2KB, docx)

Articles from Current Research in Food Science are provided here courtesy of Elsevier

RESOURCES