Characterization of Cantal and Salers Protected Designation of Origin Cheeses Based on Sensory Analysis, Physicochemical Characteristics and Volatile Compounds

Abstract

In this work, the aim was to characterize and differentiate two Protected Designation of Origin (PDO) semi-hard French cheese categories (Salers and Cantal cheeses) by focusing on their sensory, biochemical and volatile characteristics. A total of twelve cheeses, including six Cantal and six Salers cheeses, were analyzed. The provenance of milk from two dairy cow breeds (Salers and non-Salers) was discussed sensorially and chemically for each cheese sample and for each cheese category. Despite very few significant differences in biochemical parameters, differences were observed concerning the volatile composition and sensory profiles between each cheese category. Salers cheeses were clearly differentiated by their appearance and their more intense aromatic characteristics compared to Cantal cheeses. A large number of volatile compounds (VOCs) belonging to acids, alcohols, aldehydes, ketones and esters were detected in each cheese category (n = 78). The relative quantity of each compound varied depending on the cheese category but was lowly impacted by the origin of the breed’s milk. The results suggest that the provenance of milk (Salers vs. non-Salers) have a low impact on the chemical and sensory differentiation of cheeses regardless of the PDO cheese category. However, the PDO cheese categories (Salers vs. Cantal) were clearly differentiated by their volatile and sensory characteristics. The PDO Salers cheeses presented the highest flavor variability compared to the PDO Cantal cheeses due to compounds belonging to alcohols, acids, aldehydes and ester conferring ammonia, vegetal and animal flavors compared to the PDO Cantal cheeses that were perceived as more pungent and bitter.

1. Introduction

Cheese flavor is one of the most important quality factors and consumer acceptance criteria [1]. Monitoring and ensuring that cheeses have good and specific quality is a challenge for dairy companies and producers. However, during cheese elaboration, several parameters could influence its flavor, such as cheese-making practices, know-how and the raw milk material used.

Protected Designation of Origin (PDO) cheeses stand out for their strict specifications, particularly concerning the conditions of milk production and its origin, thus establishing a link with regional specificities. Salers and Cantal are two iconic PDO French semi-hard cheeses, with an approximate weight of 40 kg, produced in the Auvergne Rhône-Alpes region at annual amounts of 1076 T and 11,448 T, respectively. The cheese-making processes of the two PDO cheeses are very close, but the big difference concerns the use of milk which can be pasteurized and that is not subjected to a specific harvest period of the year for Cantal compared to Salers cheeses. Salers is an artisanal cheese made exclusively with raw cow’s milk curdled in a specific wooden vessel called “Gerle”. It has the particularity of being produced during a specific period of the year: from 15th April to 15th November. Different cow breeds can be chosen to obtain the milk for this kind of cheese-making, but when only milk from the Massif Central breed of cows, also called “Salers”, is used in its fabrication, the cheeses obtained are then called “Salers Tradition” [2,3]. The ripening time for PDO Salers cheeses varies from a minimum of three months to two years. In contrast, PDO Cantal cheeses can be made with Salers milk or with milk originating from other cow breeds such as Prim’Holstein, Aubrac or Montbéliarde breeds. For Cantal cheese, the minimum ripening time is one month to over twelve months.

Sensory differences can be perceived in the flavor and texture between each cheese’s category. In general, the flavor profile is determined by different volatile compounds included in cheese matrices. During the ripening time, different biochemical processes lead to the synthesis of volatile compounds belonging to different chemical families, such as ketones, esters, alcohols, acids, aldehydes and terpenes [4,5]. Depending on the type of cheese, the quality and the quantity (concentration) of these volatile compounds vary. The ripening time is not the only factor involved in flavor development. Indeed, numerous studies have shown that other factors, such as the extent of milk production (feeding diets and breed), milk characteristics (raw milk, heating treatment, milk microflora and milk composition) and the diversity in cheese-making practices, could have an impact on the development of flavor in cheese [6,7,8,9,10]. In reality, professionals of the supply chain think that there is an evolution of the flavors between Cantal and Salers cheeses due to the origin of the milk (Salers or other cow breed) after a long ripening time. However, this is just an observation. To date, no study has shown the differences between these two types of cheese depending on the origin of the milk. Several studies were conducted specifically on these types of cheese (Cantal or Salers, but separately), covering various topics such as microbiological aspects, the characterization of cheese texture and the aromatic profiles [6,11,12,13]; however, to date, the sensory and biochemical differentiations between these two categories of cheese have not been studied.

The aims of this exploratory study are to sensorially and chemically characterize the two categories of PDO cheeses at an advanced ripening period (>8 months) and to evaluate the differences that could come from those two categories of cheese (Cantal vs. Salers) and evaluate the differences in cheeses coming from different cow breeds (Salers vs. non-Salers).

2. Material and Methods

2.1. Sample Selection

Twelve uncooked PDO cheese samples belonging to two categories of cheese, Cantal and Salers, were collected from various producers and dairy companies in the Massif Central region. All cheeses were made from either Salers raw milk (Salers_M) or from milk of other cow breeds (OB_M) (this factor represents milk from different dairy cattle breeds) according to the specifications for PDO Cantal or Salers cheeses. In order to represent the existing sensory diversity for each of the categories of cheese studied, many samples were selected and manufactured from different producers or dairies, except for the Cantal category produced from Salers’s milk, where two samples that were studied came from the same dairy. All samples were stored in the same maturing cellar and were analyzed with a ripening time ranging from 8 to 12 months (

Table 1. Descriptions of the cheese samples studied.

Cheese	Cow Breed	Producer/Dairy	Cheese Sample (Code)
Cantal	Salers 1	Dairy 1	A
Cantal		Dairy 1	C
Cantal	Other cow breeds 2	Dairy 1	D
Cantal		Dairy 1	B
Cantal		Producer 1	E
Cantal		Producer 2	F
Salers	Salers 1	Producer 3	G
Salers		Producer 4	H
Salers		Producer 5	I
Salers	Other cow breeds 2	Producer 6	J
Salers		Producer 7	K
Salers		Producer 8	L

¹ Milk used from Salers cows only; ² milk used from other cow breeds (Montbéliarde or Aubrac breeds).

). Whole cheeses (approximate weight of 40 kg) were selected directly in the maturing cellar by supply chain professionals. One part of the cheese was stored in cold storage at 1 °C prior to sensory analysis. For physicochemical analysis, all samples were homogenized and stored at −20 °C until analysis.

2.2. Physicochemical Analysis

The physicochemical parameters of Salers and Cantal cheeses were analyzed in compliance with ISO standards. Cheese pH was measured using a CG 840-model penetration pH electrode (Schott, Mainz, Germany) by inserting the probe directly in a grated cheese sample. Dry matter (DM) was analyzed by desiccation in compliance with the ISO 5534 standard [14]. A weighted test portion mixed with sand was dried by heating it for 24 h at 102 °C. Fat was measured using the HEISS method described in NF V04-287 standard [15]. The fat-in-dry matter (FDM) ratio was calculated. Water activity (Aw) was measured in about 5 g of grated cheese samples at 20 °C as per the manufacturer’s instructions (Hygrolab C1, Rotronic, Bassersdorf, Schweiz). Total Nitrogen (TN) and Water-Soluble Nitrogen (WSN) were determined using the Kjeldahl method, in compliance with the ISO 8968-1 standard [16]. The WSN/TN ratio was calculated and can be used as a primary proteolysis indicator [17].

Chloride content was evaluated by potentiometry in compliance with the NF ISO 5943:2007 standard [18]. The chloride content was then converted into salt percent (sodium chloride). The salt-in-moisture ratio was calculated (salt/M). All measurements were carried out in triplicate.

2.3. Color

The color of the cheeses was determined using a CR-400 colorimeter (Konika-Minolta, Osaka, Japan). The CIELAB color model and D65 illuminant were used to perform the measurements. A piece of the cheese (40 × 3 cm) was cut along the height of the cheese, and 3 cm thick slices were obtained. Three random measurements were performed directly on the slice’s surface at room temperature (21 ± 1 °C).

2.4. Volatile Compound Analysis

Volatile compounds were determined using purge and trap extraction coupled with Gas Chromatography–Mass Spectrometry (GC-MS) as described by Caron et al. (2021) [19]. Three grams of the ripened cheeses (without rind) was frozen at −80 °C. Before the analyses, the cheese samples were stored at 4 °C for 16 h in order to stabilize the interactions between the matrix and volatile compounds.

Volatile compounds were extracted by using a purge and trap extraction system. The sample was shaken and heated (incubation). The station performs the purge of the headspace (GERSTEL MPS equipped dynamic Head Space (DHS)) by controlling the temperature, agitation, and inert gas flow using needles through the septum of the flask. The extracted compounds were trapped and concentrated on the temperature-controlled trap (tenax, 30 °C, volume of 300 mL He, flow rate of He 30 mL/min). The water in the trap was removed using a dry inert gas (dry purge: 30 °C, volume: 300 mL He, flow rate: He 50 mL/min).

The GC-MS analyses were performed on an Agilent 7890B (Santa Clara, CA, USA) equipped with a mass selective detector (Agilent 5977B, Santa Clara, CA, USA). Samples were analyzed on a PEG capillary column (HP-Innowax, 60 m × 0.32 mm, 0.25 µm film). The column carrier gas was helium at a constant flow rate of 1.6 mL·min⁻¹. Injection was performed in a splitless mode (1 min). The oven temperature was held at 40 °C for 5 min, increased to 155 °C at a rate of 4 °C·min⁻¹, increased from 155 to 250 °C at a rate of 20 °C·min⁻¹ and finally maintained at 250 °C for 5 min.

Compounds were identified by their retention time and by comparing their mass spectrum with those of the NIST 2017 Mass Spectral Library [20]. Only compounds with a score above 80% were selected.

The signal abundance (TIC-Total Ion Current) was used to relatively quantify the volatile compounds of cheeses by means of their chromatographic areas.

2.5. Sensory Analysis

2.5.1. Training Session

A panel of 10 panelists (females aged 40 to 65 years), from the external panel of VetAgro Sup’s Engineering School, were selected and trained according to the guidelines in ISO 8586-1 [21]. This panel had previous experience in the evaluation of cheeses. To describe cheeses, a sensory profile method was applied according to the recommendations of ISO standard 13299 [22]. A training session was conducted, including six 1 h sessions. During these training sessions, the panel generated numerous attributes to describe uncooked cheeses. A list of 31 sensory attributes was constructed to characterize cheeses as described by Bord et al. (2021) [23], and a cheese lexicon was generated, including the definition and the references for each sensory attribute. During this training session, the performance of the panel was validated in order to control-check for repeatability, discriminative capacity and consensus.

2.5.2. Evaluation Session

First, the sensory panel evaluated the appearance of each cheese. A slice, approximately 2 cm thick, was cut from each cheese, and the same slice was presented to the panel. For other attributes (odor, flavor and texture in mouth), only the cheese core (a triangle shape slice of 30 g (the rind was removed) was provided to panelists. Cheeses were taken out of the cold storage room 20 min before tasting, and then they were placed on a plastic plate, identified by a 3-digit code, and served at room temperature (20 ± 1 °C). The samples were presented in a monadic sequence and distributed according to a Williams Latin square design to consider the first effect of order and carry-over. Five samples were analyzed per one-hour session. All samples were duplicated.

A 10-point linear scale was used to evaluate each attribute, anchored from 0 (no intensity) to 10 (high intensity). To cleanse the mouth, mineral water (Evian, France) and unsalted crackers were provided between samples.

Data were collected with the Tastel software (version 2019; ABT Informatique, Rouvroy-sur-Marne, France).

2.6. Statistical Analysis

For the dataset of biochemical and volatile compounds, a two-way model ANOVA was used, taking into account the interactions (Category × Breed).

Data from sensory profiles were analyzed using a three-way mixed-model ANOVA. The model included cheese category (2 levels) and cow breeds (2 levels) as fixed effects and panelists as the random effect. The interactions of Category × Breed were included on the model. Tuckey’s test was applied for multiple comparison when significant differences were observed (p < 0.05). Statistical analyses were performed using the XLSTAT software, version 2020 (Addinsoft, Paris, France).

In order to show the impact of the cheese category or the cow breeds for each type of data (sensory and volatile compounds), a principal component analysis (PCA) was performed on the sensory and volatile compound parameters. PCA was performed using “FactoMineR” and “factoextra” to obtain confidence ellipses package in R.

The relationships between sensory characteristics and volatile compounds were explored by Partial Least Squares regression (PLS), which was carried out with significant volatile compounds as the X-matrix and significant sensory attributes as the Y-matrix. The variables were selected when a significant effect (p < 0.05) was observed according to the cheese category and/or the milk cow breed’s provenance.

3. Results and Discussion

3.1. Physicochemical Characteristics of Cheeses

The biochemical compositions of the cheeses are summarized in

Table 2. Biochemical characteristics for each cheese category and cow breeds used to produce cheeses.

Parameters	Cheese Category			Cow Breeds’ Milk			Category × Breed
	Cantal	Salers	Significance	OB_M	Salers_M	Significance	Significance
Dry matter-DM (%)	62.11 ± 0.5	62.38 ± 0.6	NS	62.49 ± 0.4	61.91 ± 0.4	NS	NS
Fat (%)	29.33 ± 1.29	29.25 ± 1.04	NS	30.40 a ± 1.12	27.73 b ± 1.15	***	NS
Fat in DM (%)	47.27 ± 2.47	46.83 ± 3.44	NS	48.67 a ± 1.83	44.78 b ± 2.56	***	NS
pH	5.50 b ± 0.11	5.59 a ± 0.12	*	5.54 ± 0.12	5.56 ± 0.09	NS	NS
Aw	0.91b ± 0.02	0.94 a ± 0.01	***	0.92 ± 0.03	0.93 ± 0.02	NS	NS
Salt (%)	2.34 ± 0.15	2.29 ± 0.24	NS	2.27 ± 0.14	2.37 ± 0.26	NS	NS
Salt in DM (%)	3.77 ± 0.28	3.67 ± 0.50	NS	3.63 ± 0.32	3.84 ± 0.49	NS	NS
TN	3.97 ± 0.21	4.12 ± 0.27	NS	3.97 b ± 0.24	4.15 a ± 0.19	*	NS
WSN	1.21 ± 0.11	1.23 ± 0.10	NS	1.23 ± 0.10	1.22 ± 0.12	NS	NS
WSN/TN (%)	30.62 ± 2.86	30.00 ± 3.02	NS	31.03 ± 2.92	29.29 ± 2.68	NS	NS
L* (Lightness)	78.60 a ± 4.13	73.20 b ± 3.00	***	76.27 ± 5.14	75.37 ± 4.20	NS	NS
a* (redness)	−2.75 b ± 0.53	−2.37 a ± 0.50	*	−2.56 ± 0.57	−2.56 ± 0.51	NS	NS
b* (yellowness)	25.81 b ± 1.50	27.83 a ± 2.35	***	27.35 a ± 2.28	26.07 b ± 1.94	**	***

Data are shown as means and standard deviation (n = 3). Means within row with differing superscripts differ (p < 0.05). NS = non significant; * p < 0.05; ** p < 0.01; *** p < 0.001. p-values are shown in Table S1. (Red color was added to significance when presenting no significant interaction between cheese category and cow breed). (Aw = water activity; TN = Total Nitrogen; WSN = Water-Soluble Nitrogen; WSN/TN = Ratio of Water-Soluble Nitrogen to Total Nitrogen; OB_M = Other Cow Breeds’ Milk; Salers_M = Salers Milk).

.

Concerning the effect of the cheese category, the pH and Aw values were significantly different, with the highest pH value being found for the Salers cheeses. The Salers cheeses had a slightly higher pH and Aw (5.59; 0.94) compared to the Cantal cheeses (5.50; 0.91), respectively. Similar results were found for Salers and Cantal cheeses in previous studies [11,13]. The Salers cheeses were less light (L) with a value of 73.20 and had a higher yellowness index (27.83) compared to the Cantal cheeses. In contrast, Salers cheeses had a higher redness index (−2.37) compared to Cantal cheeses. As described by Nozière et al. (2006) [24], the yellow color of cheeses can be related to the carotenoids concentration in fat. No significant differences were detected between cheese categories for the other physicochemical parameters (dry matter, fat, fat in DM, salt, salt in DM, TN and WSN/TN).

Regarding the cow’s breed provenance of the milk, the fat content and fat in dry matter (FDM) presented the lowest values (27.73 and 44.78, respectively) for the cheeses made with milk from Salers cows (Salers_M) compared to cheeses made with milk from other cow breeds (OB_Milk) (30.40 and 48.67, respectively). Indeed, the calf presence during milk harvest decreased the fat content of the milk and the fat in the dry matter of the cheese [25]. The TN content differed according to the milk category. The TN content in the cheeses reached the highest value for Salers_M (4.15) when compared to OB_M (3.97). In contrast, the ratio of WSN/TN decreased with Salers_M (29.29) compared with OB_M (31.03). This ratio shows that the cheeses were ripened (over 50 weeks) with a ratio of 30. In comparison, Manzocchi et al. (2021) [26] observed a ratio of 17, approximately, for 9-week-old Cantal-type cheeses.

When analyzing the interactions between cheese category and cow breeds’ milk, only one physicochemical parameter, the yellowness, appears to be significant (p < 0.05), meaning the observed differentiation between cheese categories and breeds for this parameter could not be generalized.

3.2. Volatile Compound Profiles in Cheeses

The volatile composition of all cheeses was relatively quantified, and the results are presented in

Table 3. Volatile compounds detected in cheese samples according to cheese category and cow breeds’ milk (arbitrary units (peak area divided by factor of 10⁵)).

Chemical Family	Volatile Compounds	Cheese Category			Cow Breeds’ Milk			Category × Breed
		Cantal	Salers	Significance	OB_M	Salers_M	Significance	Significance
Acid	4-Methyl-2-oxovaleric acid	0.0 ± 0.0	3.4 ± 3.3	**	1.9 ± 3.0	1.2 ± 2.6	NS	NS
	Acetic acid	773.5 ± 474.5	894.9 ± 390.1	NS	889.9 ± 403.3	827.7 ± 478.8	NS	*
	Butanoic acid	321.3 ± 222.5	261.7 ± 132.6	NS	308.7 ± 180.0	298.4 ± 204.2	*	NS
	Butanoic acid, 3-methyl-	3.0 ± 1.8	13.0 ± 8.0	***	8.6 ± 7.7	7.2 ± 8.6	NS	NS
	Hexanoic acid	59.6 ± 48.8	86.3 ± 47.9	NS	78.6 ± 47.5	67.1 ± 53.5	NS	NS
	n-Decanoic acid	2.7 ± 3.4	4.4 ± 2.7	*	3.9 ± 3.0	2.9 ± 3.1	**	NS
	Octanoic acid	9.0 ± 7.4	14.4 ± 7.9	*	12.8 ± 7.4	10.3 ± 8.3	NS	NS
	Pentanoic acid	1.6 ± 0.7	1.5 ± 0.8	NS	1.6 ± 0.8	1.6 ± 0.8	NS	NS
	Propanoic acid	30.3 ± 32.1	59.0 ± 66.1	NS	46.6 ± 55.0	27.6 ± 27.6	*	NS
	Propanoic acid, 2-methyl-	2.8 ± 1.8	9.8 ± 4.5	***	6.7 ± 4.9	5.4 ± 5.1	NS	NS
Alcohol	1-Butanol	118.2 ± 81.0	209.5 ± 293.7	NS	176.1 ± 221.6	110.1 ± 69.8	NS	NS
	1-Butanol, 3-methyl-	78.4 ± 77.5	2499.7 ± 904.4	***	1403.1 ± 1393.9	1094.6 ± 1551.7	**	**
	1-Butanol, 3-methyl-, acetate	12.2 ± 11.4	362.7 ± 165.9	***	204.2 ± 214.5	156.8 ± 222.3	NS	NS
	1-Hexanol	8.3 ± 3.9	23.2 ± 30.2	NS	16.6 ± 23.3	9.6 ± 4.1	NS	NS
	1-Octanol	1.3 ± 0.6	1.7 ± 1.6	NS	1.6 ± 1.3	1.3 ± 0.5	NS	NS
	1-Penten-3-ol	2.4 ± 2.8	0.8 ± 1.1	*	1.0 ± 0.9	2.0 ± 2.4	*	NS
	1-Propanol	205.3 ± 122.3	195.8 ± 53.7	NS	207.2 ± 93.7	196.1 ± 100.4	NS	NS
	1-Propanol, 2-methyl-	7.4 ± 1.5	533.2 ± 301.3	***	294.3 ± 345.4	239.7 ± 378.9	*	*
	1-Propanol, 3-(methylthio)-	0.0 ± 0.0	2.4 ± 1.3	**	1.3 ± 1.5	0.9 ± 1.5	NS	NS
	2-Butanol	1613.2 ± 450.3	1297.5 ± 572.1	NS	1503.1 ± 526.6	1513.7 ± 428.6	NS	*
	2-Heptanol	8.1 ± 4.4	27.3 ± 16.0	**	19.2 ± 14.9	13.2 ± 9.7	NS	NS
	2-Nonanol	0.8 ± 0.9	1.6 ± 1.6	NS	1.2 ± 1.4	0.8 ± 1.1	NS	NS
	2-Pentanol	65.9 ± 31.6	190.2 ± 83.2	**	138.0 ± 85.6	93.2 ± 52.2	NS	NS
	Ethanol	7.3 ± 3.5	34.0 ± 13.9	**	22.0 ± 17.1	18.1 ± 16.2	NS	NS
	Ethanol, 2-(dodecyloxy)-	1.3 ± 1.9	0.3 ± 0.6	NS	0.5 ± 1.3	0.9 ± 1.7	NS	NS
	Isopropyl Alcohol	0.6 ± 0.3	1.3 ± 0.7	**	1.0 ± 0.6	0.8 ± 0.4	NS	NS
	Phenol	3.3 ± 1.3	0.0 ± 0.0	***	1.5 ± 1.9	2.2 ± 1.9	NS	NS
	Phenylethyl Alcohol	0.3 ± 0.7	26.9 ± 25.1	***	14.8 ± 22.7	13.0 ± 25.6	NS	NS
Aldehyde	Benzaldehyde	4.4 ± 2.7	1.9 ± 0.6	***	2.5 ± 1.0	3.6 ± 2.5	*	NS
	Butanal, 3-methyl-	9.2 ± 2.6	10.2 ± 3.7	NS	9.6 ± 3.3	10.4 ± 3.1	*	NS
	Dodecanal	4.4 ± 2.1	0.0 ± 0.0	***	2.0 ± 2.7	2.9 ± 2.7	NS	NS
	Hexanal	2.2 ± 4.1	0.0 ± 0.0	**	0.2 ± 0.7	1.4 ± 3.4	*	*
	Nonanal	2.0 ± 0.7	1.3 ± 0.3	***	1.5 ± 0.4	1.8 ± 0.6	*	NS
Ester	Ethyl 9-decenoate	0.0 ± 0.0	2.5 ± 1.6	***	1.3 ± 1.7	1.1 ± 1.7	NS	NS
	Ethyl Acetate	14.9 ± 15.4	280.8 ± 172.5	***	161.0 ± 183.6	117.4 ± 170.9	NS	NS
	Acetic acid ethenyl ester	39.0 ± 19.6	29.9 ± 22.4	*	31.6 ± 19.7	33.1 ± 18.9	NS	***
	Acetic acid, 2-phenylethyl ester	0.0 ± 0.0	8.3 ± 7.7	**	4.6 ± 7.0	2.7 ± 4.7	NS	NS
	Butanoic acid, 1-methylpropyl ester	17.0 ± 13.1	25.5 ± 16.6	NS	23.1 ± 14.6	20.8 ± 16.5	NS	NS
	Butanoic acid, 2-methylpropyl ester	0.0 ± 0.0	3.0 ± 2.3	***	1.7 ± 2.2	1.2 ± 2.4	NS	NS
	Butanoic acid, 3-methyl-, ethyl ester	0.0 ± 0.0	2.2 ± 1.7	***	1.2 ± 1.7	1.1 ± 1.7	*	*
	Butanoic acid, 3-methylbutyl ester	0.0 ± 0.0	25.1 ± 14.8	***	13.7 ± 16.7	10.5 ± 18.1	NS	NS
	Butanoic acid, ethyl ester	20.6 ± 10.3	139.6 ± 72.4	***	86.2 ± 79.9	61.3 ± 61.3	NS	NS
	Decanoic acid, ethyl ester	1.4 ± 0.6	15.8 ± 9.8	***	9.3 ± 10.2	7.2 ± 10.1	NS	NS
	Dodecanoic acid, ethyl ester	0.0 ± 0.0	2.3 ± 1.4	***	1.3 ± 1.5	0.9 ± 1.5	NS	NS
	Hexanoic acid, 2-methylpropyl ester	5.8 ± 4.2	11.9 ± 11.1	NS	9.5 ± 8.8	8.7 ± 9.3	NS	NS
	Hexanoic acid, ethyl ester	32.8 ± 22.2	306.5 ± 156.0	***	183.7 ± 178.8	141.2 ± 172.0	NS	NS
	Hexanoic acid, propyl ester	2.3 ± 2.7	3.1 ± 1.2	NS	2.9 ± 2.1	2.5 ± 2.4	NS	NS
	Isoamyl lactate	0.0 ± 0.0	0.7 ± 0.4	***	0.4 ± 0.5	0.3 ± 0.5	NS	NS
	Isobutyl acetate	0.0 ± 0.0	14.3 ± 7.4	***	7.8 ± 9.0	6.5 ± 9.6	*	**
	Isopentyl hexanoate	0.0 ± 0.0	4.7 ± 3.1	***	2.5 ± 3.3	2.0 ± 3.6	NS	NS
	Octanoic acid, 2-butyl ester	1.4 ± 0.9	2.1 ± 1.7	NS	1.9 ± 1.4	1.7 ± 1.5	NS	NS
	Octanoic acid, ethyl ester	4.8 ± 3.5	69.0 ± 43.4	***	40.1 ± 45.2	31.6 ± 45.8	NS	NS
	Propanoic acid, 2-hydroxy-, ethyl ester	0.0 ± 0.0	13.2 ± 8.2	***	7.2 ± 9.0	5.5 ± 9.0	NS	NS
	Propanoic acid, ethyl ester	2.0 ± 1.9	35.9 ± 20.6	***	20.5 ± 22.9	13.0 ± 17.6	NS	NS
	sec-Butyl acetate	6.9 ± 9.2	60.0 ± 51.9	**	36.4 ± 46.3	24.4 ± 28.9	NS	NS
Ketone	2,3-Pentanedione	2.0 ± 1.8	0.0 ± 0.0	***	0.6 ± 1.1	1.3 ± 1.7	**	**
	2-Butanone	1930.9 ± 161.7	1604.0 ± 847.9	NS	1766.4 ± 648.7	1792.1 ± 535.8	NS	NS
	2-Heptanone	17.9 ± 13.3	18.8 ± 12.4	NS	19.0 ± 13.0	17.8 ± 13.4	NS	NS
	2-Hexanone	0.8 ± 1.3	2.1 ± 1.7	NS	1.6 ± 1.6	1.0 ± 1.5	NS	NS
	2-Hydroxy-3-pentanone	3.8 ± 1.8	1.1 ± 1.2	***	2.2 ± 2.0	2.9 ± 2.2	**	***
	2-Nonanone	9.9 ± 8.5	6.8 ± 2.2	NS	8.5 ± 6.6	9.1 ± 7.1	NS	NS
	2-Pentanone	22.4 ± 16.7	31.3 ± 20.7	NS	27.6 ± 19.6	23.2 ± 19.6	NS	NS
	2-Propanone, 1-hydroxy-	4.0 ± 1.5	2.9 ± 2.1	NS	3.3 ± 1.9	3.4 ± 1.6	NS	*
	2-Undecanone	0.9 ± 0.5	1.0 ± 0.5	NS	1.0 ± 0.5	1.0 ± 0.6	NS	NS
	4-Octanone, 5-hydroxy-2,7-dimethyl-	0.0 ± 0.0	4.5 ± 3.6	***	2.5 ± 3.5	1.7 ± 3.3	NS	NS
	8-Nonen-2-one	1.0 ± 1.0	1.0 ± 0.4	NS	1.1 ± 0.8	1.1 ± 0.9	NS	NS
	Acetoin	643.7 ± 470.1	254.4 ± 128.1	***	389.5 ± 350.8	488.5 ± 441.8	***	***
	Acetone	14.3 ± 6.9	15.4 ± 8.1	NS	14.4 ± 7.5	13.8 ± 5.9	NS	*
Terpene	D-Limonene	6.0 ± 11.2	0.4 ± 0.2	NS	3.4 ± 8.6	4.1 ± 9.4	NS	NS
Other	2(3H)-Furanone, dihydro-5-pentyl-	0.0 ± 0.0	2.7 ± 0.7	***	1.5 ± 1.5	0.9 ± 1.4	NS	NS
	3-Methyl-2-(2-methyl-2-butenyl)-furan	1.6 ± 1.9	2.4 ± 2.0	NS	2.2 ± 2.0	1.9 ± 2.3	NS	NS
	1,3-Dioxolane, 2-ethyl-2,4,5-trimethyl-	0.0 ± 0.0	4.1 ± 6.0	*	2.3 ± 4.8	2.2 ± 5.2	NS	NS
	2,2,4-Trimethyl-1,3-pentanediol diisobutyrate	1.5 ± 0.3	1.3 ± 0.2	**	1.3 ± 0.2	1.4 ± 0.3	NS	NS
	2-Butene	24.1 ± 18.4	26.5 ± 12.1	NS	26.0 ± 15.7	25.3 ± 15.2	NS	NS
	Dimethyl sulfone	3.9 ± 1.4	4.0 ± 2.4	NS	4.0 ± 2.0	4.3 ± 2.0	NS	*
	Propane, 1,2-dimethoxy-	7.8 ± 2.2	0.9 ± 2.1	***	4.1 ± 4.2	5.2 ± 4.2	NS	*
	Pyrazine, trimethyl-	0.5 ± 0.4	0.0 ± 0.0	***	0.1 ± 0.2	0.3 ± 0.4	***	***

Numbers represent mean ± standard deviation. Values represent peak area (arbitrary units). Values reported are means from duplicate cheese samples (n = 2) ± SD; NS = non significant; * p < 0.05; ** p < 0.01; *** p < 0.001. p-values are shown in Table S2. (red color was added to Significance column when presenting no significant interaction between cheese category and cow breed (OB_M = other cow breeds’ milk; Salers_M = Salers milk) or between cheese categories (Salers vs. Cantal)).

. In total, the analysis of VOCs revealed the presence of seventy-eight VOCs in the cheese samples, belonging to different chemical families: twenty-two esters, eighteen alcohols, thirteen ketones, ten acids, five aldehydes, one terpene and nine other compounds belonging to different chemical families (furan, pyrazine and heterocycle). To illustrate the differences among chemical families among cheeses, Figure 1 was drawn to represent the average percentage of the classes of volatile compounds found in the cheeses, classified in function of the cow’s breed provenance of milk and of cheese categories. First, it was observed that the volatile composition was very similar between OB_M and Salers_M cheeses, with alcohol contents between 45.4% and 48.2%, followed by ketone contents between 26.9% and 30.9%, acid contents 16.4% for the both and ester contents between 6.5% and 8.5%. However, sixteen volatile compounds presented statistical differences (p < 0.05) according to the provenance of cow’s breed milk (Table 3). Among them, only six did not present significant interactions with cheese category: butanoic acid, n-decanoic acid, 1-penten-3-ol, 3-methyl butanal, benzaldehyde and nonanal (Table 3).

Butanoic acid presented a weak statistical difference when taking into consideration the “milk breed”, with the amount in cheeses made with Salers’s milk being slightly higher than those made with milk from other dairy cow breeds. This compound was also identified in Salers cheeses by Callon et al. (2005) [6]. Butanoic acid is a major aroma compound with a cheesy, sharp aroma, and it plays an important role in the flavor of many cheeses [27]. The source of the butanoic acid in the cheese essentially results from lipolysis by endogenous milk-based or microbial-origin lipases, but it can also result from the metabolism of deamination of amino acids and lipid oxidation [5,28]. Butanal, 3-methyl- is more abundant in cheeses made with Salers milk than those made from other breeds’ milk. This compound is associated as a key flavor compound with many hard and semi-hard cheese varieties and is derived from the degradation of leucine. It is known that those compounds can be produced by different lactic acid bacteria, such as Lactococcus lactis, Streptococcus salivarius and Lactobacillus delbrueckii. Those species are part of the raw milk flora (in raw milk cheese) and starter cultures [29], which could explain why this compound is more abundant in the cheeses made with Salers milk.

The cheese categories showed bigger differences in chemical families compared to the cow’s breed origin of milk. PDO Cantal cheeses were characterized by ketones (42.8%), followed by alcohols (34.4%) and acids (19.34%). In contrast, Salers cheeses were characterized by a high content of alcohols (55.1%), followed by ketones (19.8%), acids (13.7%) and esters (10.8%). The main VOCs identified (Table 3) from the experimental cheeses were consistent with those previously reported in Cantal-type or Salers-type cheeses [6,30]. In our study, forty-seven VOCs enabled us to differentiate PDO Cantal cheeses and PDO Salers cheeses (Table 3). Thirty-seven presented no interactions with milk from different cow breeds. Among them, esters were the family that presented the highest number of compounds compared to the other analyzed volatile families, enabling us to discriminate PDO Cantal cheeses and PDO Salers cheeses. Among the list of twenty-two esters analyzed, seventeen esters were systematically higher in PDO Salers compared to PDO Cantal cheeses. In that list, three compounds (acetic acid ethenyl ester, butanoic acid 3-methylbutyl ester and isobutyl acetate) could not be used to differentiate the cheese categories due to their interaction with cow breeds’ milk. And ten were specific to PDO Salers cheeses, such as ethyl-9-decenoate; acetic acid; 2-phenylethyl ester; butanoic acid; 2-methyl propyl ester; butanoic acid, 3-methyl-ethyl ester; butanoic acid, 3-methylbutyl ester; dodecanoic acid, ethyl ester; isoamyl lactate; isobutyl acetate; isopentyl hexanoate and propanoic acid, 2-hydroxy-ethyl ester. Butanoic acid, 3-methyl-ethyl ester and isobutyl acetate could not be generalized due to the significant interaction with cow breeds.

Esters represent a class of VOCs produced by esterification reactions between alcohols, produced during lactose fermentation or amino acid catabolism or indirectly derived from FFA metabolism [31].

Generally, the type and concentration of esters found in cheeses vary according to cheese varieties and process conditions. Ethyl esters are known for their important role in the formation of a fruity character in cheese [32].

As suggested by Cornu et al. (2009) [30], the ripening time has an impact on the amount of some volatile compounds found between three and six months for Cantal cheeses. In the same way, relative amounts of volatile compounds in Salers and Cantal cheeses evolved upon cheese ripening (up to 6 months) with no specific observable trends, as previously observed in Cheddar cheeses [32]. Callon et al. (2005) [6] showed that the aromatic profile of Salers cheeses (150 days of ripening time) was modified according to the indigenous flora used.

Alcohols were the second most abundant family affected by the cheese category, presenting ten VOCs, enabling us to differentiate PDO Salers and PDO Cantal cheeses. Alcohols can generate a wide variety of aromas in cheese, for example, alcoholic, fruity, green, floral, rose, malty, honey, herbaceous, sweet, fragrant, caramel and others. Among the eleven alcohols, nine were more present in PDO Salers cheeses compared to PDO Cantal cheeses (1-butanol; 3 methyl; 1-butanol, 3-methyl-acetate; 1-propanol, 2-methyl; 1-propanol, 3-methylthio; 2-butanol; 2-heptanol; 2-pentanol; ethanol, propan-2-ol and phenylethanol). Phenol was only present in PDO Cantal cheeses and was not detected in PDO Salers cheeses. 1-propanol, 3-methylthio was found in our study to only be present in PDO Salers cheeses. Phenylethanol was relatively more abundant in Salers cheeses compared to Cantal cheeses. Its production originates from phenylalanine degradation and is among one of the most odorous aromatic alcohols [4,33]. Ethanol is also detected in higher amounts in Salers cheeses in comparison to Cantal cheeses, but it has often been detected in different categories of cheeses such as Cheddar [32] and Hispánico [34]. This compound is produced from lactose fermentation by starter bacteria [35] and yeasts [36]. Even if it has a limited role in the cheese aroma despite its high levels, it contributes to the formation of esters [37]. It should be noted that the higher presence of 1-butanol,3-methyl in Salers cheese vs. Cantal cheese could not be generalized due to its significant interaction with cow breeds’ milk.

Ketones were the third family to permit differentiation between the two cheese categories. Three VOCs were more present in PDO Cantal cheeses compared to PDO Salers cheeses, namely 2,3-pentanedione, 2-hydroxy-3-pentanone and acetoin, with 2,3-pentanedione only being found in PDO Cantal cheeses. These findings should be modulated due to the significant interaction of these three compounds with cow breeds’ milk. 4-octanone, 5-hydroxy-2-7-dimethyl was only present in PDO Salers cheeses. Generally, ketone or methyl ketones are recognized as key components in the flavor of mold-ripened cheeses such as Blue cheeses [19,38], but they can be identified in other kinds of cheeses like Cheddar [39], Ricotta [40], Parmigiano Reggiano [41] and Swiss Gruyere [4]. It has been suggested that methyl ketones in non-mold-ripened cheeses may be formed by the enzymatic oxidative decarboxylation of fatty acids by the presence of lactic acid bacteria [42]. Ketones have a typical odor with a low perception threshold [4]. Among these ketone compounds, acetoin is derived from citrate metabolism. Diacetyl can be converted to acetoin and 2.3 butanediol, which is often related to buttery notes. As previously identified by Callon et al. (2015), the content of acetoin varied in Salers cheeses manufactured with the same pasteurized milk, reinoculated with three different microbial communities, hence showing that the flavor production in cheese results from a complex ecosystem and can be varied by inoculating different microbial communities. In another work, it was reported that wild lactococci strains were positive for citrate formation. More particularly, L. Lactis strains used as starter cultures in these cheeses produced acetoin [43].

Concerning volatile acids, five (4-methyl-2-oxovaleric acid, 3-methylbutanoic acid, 2-methylpropanoic acid, n-decanoic acid and octanoic acid) were more concentrated in PDO Salers cheeses compared to PDO Cantal cheeses, and 4-methyl-2-oxovaleric acid was only present in PDO Salers cheeses. The presence of volatile acids in Salers and Cantal cheeses may play an important role in aroma production because they have been reported as possible precursors of other aromatic compounds, such as ketones, lactones, aldehydes, esters, methyl ketones and secondary alcohols [44]. Consequently, the PDO samples studied have a high ripening stage (>8 months of ripening) that could explain the high content of secondary aromatic metabolites identified.

Regarding aldehydes, four (benzaldehyde, dodecanal, hexanal and nonanal) were more present (p < 0.05) in PDO Cantal cheeses compared to PDO Salers cheeses. Such result should be modulated for hexanal, which presents significant interactions with cow breed’s milk. Aldehydes may have an important influence on cheese flavor because most of them are detected in other cheeses [33]. These are transitional compounds in cheese because they are quickly reduced to primary alcohols or oxidized to corresponding acids [45]. Benzaldehyde characterized Cantal cheeses and was the most abundant cyclic compound found in the present study. Benzaldehyde has previously been detected in other cheeses [46]. This compound originates from the oxidation reaction of cinnamic acid or phenylacetaldehyde and develops an aromatic note of bitter almond [47].

Only one terpene was detected (limonene), and no statistical difference was observed when comparing Cantal and Salers cheeses in our study. Terpene profile differences originated from multiple factors, such as physical, chemical or microbial mechanisms involved during the cheese-making process [12]. In addition, microbiological flora may have induced different chemical reactions on terpene molecules. A link has also been identified with the farm environment and, in particular, pasture feeding [48].

Regarding the other five VOCs differentiating the cheese categories, two (2(3H)-furanone, dihydro-5-pentyl; 1,3-dioxane, 2-ethyl-2-4-5-trimethyl) were only present in PDO Salers cheeses, and the other three (2,2,4-trimethyl-1,3-pentanediol diisobutyrate, propane 1,2-dimethoxy and pyrazine trimethyl) were more present in PDO Cantal cheeses compared to PDO Salers cheeses. Propane, 1,2-dimethoxy and pyrazine, trimethyl could not be generalized to differentiate cheese categories due to their significant interaction with cow breeds’ milk.

To better visualize the volatile composition among the twelve cheeses, analyzed in duplicate, a PCA was performed (Figure 2) on the totality of the VOCs. The first two components explained about 58% of the total variance. Figure 2B,C show the confidence ellipsoids representing the variability of the results over 12 cheeses (with replication). Concerning the effect of different cow breeds’ milk, the ellipses overlapping indicated a bad discrimination between cheeses made from different types of milk (Figure 2B). In contrast, the effect of the cheese category (Figure 2C) showed a good and clear distinction between both cheeses. Cantal cheeses and Salers cheeses were clearly separated based on the first component, dominated by previously described compounds such as acetoin, benzaldehyde, dodecanal and phenol. Taking into consideration the size of the ellipse, a broader chemical diversity could be observed in Salers cheeses vs. Cantal cheeses. More precisely, H_Salers cheese was described by elevated relative contents of phenylethanol and 3-methyl butanoic acid. In contrast, in the Cantal category, this chemical diversity was less pronounced.

The differences observed between each cheese category could be explained by different external factors, such as raw milk or the ruminant diet. Indeed, raw milk is an uncontrolled source of microflora; the most obvious source of variability appears to be microbial. It is well known that the potential for microflora to generate flavor variability is greater than the potential variability from different breeds or on-farm practices, especially when other aspects of cheese-making, such as curd composition and ripening conditions, are similar [49,50].

Another explanation for the differences between the two categories is the use of sheaves. In fact, the sheaf contains a microbial biofilm on the walls of the sheaf, which favors the transfer of microorganisms from the sheaf to the milk, which, in turn, favors the development of certain volatile compounds [51].

Furthermore, the relative percentage of each compound is mainly related to endogenous enzymes and microorganisms whose function can be strongly influenced by the bioactive compounds taken by animals with the diet and released in milk through the mammary gland. According to Manzocchi et al. (2021) [26], the nature and concentration of volatile compounds depend on the type of forage fed to cows for Cantal-type cheeses.

3.3. Sensory Properties of Cheeses

The sensory profiles, mainly the appearance and texture attributes, of the cheeses were affected by cheese category and milk origin (

Table 4. Intensity means from sensory profiles for each cheese category and cow breeds’ milk.

Sensory Attributes		Cheese Category			Cow Breeds’ Milk			Category × Breed
		Cantal	Salers	Significance	OB_M	Salers_M	Significance	Significance
APPEARANCE	Rind color	5.3 b ± 2.2	7.2 a ± 1.6	***	6.6 a ± 2.1	5.7 b ± 2.1	***	***
	Rind thickness	4.9 b ± 2.2	6.9 a ± 1.8	***	6.1 a ± 2.3	5.6 b ± 2.2	***	**
	Color_core	5.6 b ± 1.5	6.7 a ± 1.5	***	6.1 ± 1.7	6.2 ± 1.5	NS	***
	Color homogeneity	6.3 a ± 1.7	6.1 b ± 2.0	**	6.1b ± 1.9	6.4 a ± 1.8	**	***
	Marbled_core	4.6 b ± 2.1	5.1 a ± 2.3	***	5.1a ± 2.2	4.5 b ± 2.2	***	NS
	Cracked_core	5.0 a ± 2.1	3.4 b ± 2.5	***	4.5a ± 2.5	3.8 b ± 2.2	**	NS
	Firm_ in Touch	6.8 a ± 1.6	6.4 b ± 1.9	*	7.0a ± 1.7	6.1 b ± 1.7	***	NS
FLAVOR	Overall odor	5.8 b ± 1.4	6.3 a ± 1.3	***	6.0 ± 1.4	6.0 ± 1.3	NS	NS
	Animal odor	3.0 b ± 2.1	3.8 a ± 2.1	***	3.4 ± 2.2	3.4 ± 2.1	NS	NS
	Mushroom odor	1.6 b ± 1.8	1.8 a ± 1.9	*	1.6 ± 1.8	1.8 ± 1.8	NS	NS
	Lactic_odor	3.9 ± 2.3	3.7 ± 2.2	NS	3.8 ± 2.3	3.8 ± 2.2	NS	NS
	Global aroma	6.1 ± 1.4	6.3 ± 1.4	NS	6.2 ± 1.4	6.3 ± 1.4	NS	NS
	Mushroom aroma	1.8 ± 1.8	2.0 ± 2.0	NS	1.8 ± 1.8	2.0 ± 1.9	NS	NS
	Vegetable aroma	2.9 b ± 1.9	3.4 a ± 2.1	***	3.1 ± 2.0	3.4 ± 2.1	NS	NS
	Nutty aroma	2.3 a ± 2.0	1.9 b ± 2.0	***	2.1 ± 2.0	2.0 ± 2.2	NS	*
	Animal aroma	2.2 b ± 1.9	2.8 a ± 1.9	***	2.3 b ± 1.9	2.8 a ± 1.9	*	NS
	Ammoniac aroma	1.3 ± 1.7	1.5 ± 2.1	NS	1.3 b ± 1.8	1.6 a ± 2.2	*	NS
	Atypical aroma	1.7 b ± 2.2	2.4 a ± 2.5	***	1.9 ± 2.3	2.2 ± 2.4	NS	NS
	Lactic aroma	3.3 ± 2.2	3.2 ± 2.3	NS	3.3 ± 2.2	3.2 ± 2.3	NS	NS
	Persistance	5.6 ± 1.8	5.8 ± 2.0	NS	5.7 ± 1.9	5.8 ± 2.0	NS	NS
	Salty	6.5 ± 1.3	6.6 ± 1.3	NS	6.5 ± 1.3	6.7 ± 1.3	NS	NS
	Sour	3.3 b ± 2.1	3.8 a ± 2.4	***	3.4 ± 2.2	3.7 ± 2.3	NS	NS
	Bitter	3.3 a ± 2.1	2.7 b ± 2.2	***	3.0 ± 2.1	3.1 ± 2.2	NS	**
	Pungent	3.5 a ± 2.6	3.1 b ± 2.6	***	3.3 ± 2.6	3.4 ± 2.7	NS	**
TEXTURE IN MOUTH	Firm_Mouth	5.8 a ± 1.8	5.5 b ± 2.1	*	5.8 a ± 1.9	5.4 b ± 2.0	**	NS
	Melty	3.7 ± 2.0	3.9 ± 2.2	NS	3.8 ± 2.1	3.9 ± 2.1	NS	NS
	Crumbly	3.6 ± 2.1	3.6 ± 2.3	NS	3.7 ± 2.2	3.4 ± 2.1	NS	NS
	Sticky	3.8 ± 2.0	3.6 ± 1.8	NS	3.8 a ± 1.9	3.5 b ± 1.9	*	NS
	Grainy	3.8 ± 2.0	3.5 ± 2.1	NS	3.7 ± 2.1	3.6 ± 2.1	NS	NS
	Fatty	4.7 ± 1.8	4.9 ± 1.9	NS	4.8 ± 1.8	4.7 ± 1.9	NS	NS
	Residue	4.2 ± 2.3	4.1 ± 2.4	NS	4.2 ± 2.3	4.1 ± 2.4	NS	NS

Values corresponding to the mean of intensity scores calculated by nine panelists for three replications. The intensity of an attribute was scored on a 10 cm linear scale (0 = no perception; 10 = very intense perception) (OB_M = Other cow breeds milk; Salers_M = Salers milk). NS = non significant; * p < 0.05; ** p < 0.01; *** p < 0.001. p-values are shown in Table S3. (a red color was added to the Significance column when presenting no significant interaction between cheese category and cow breed). Different lowercase letters (a,b) represent significant differences (p < 0.05) between cheeses according to Tukey’s test.

). More precisely, eighteen sensory attributes were significantly different between both cheese categories, and ten were different among cow breeds’ milk.

Concerning the appearance, PDO Salers cheeses had the darkest (7.2) and the thickest rind (6.9) in comparison with PDO Cantal cheeses (5.3 and 4.9, respectively). Salers cheeses had a yellower rind (6.7) than Cantal cheeses (5.6, in compliance with instrumental data). In contrast, Cantal cheeses had the most cracked core (5.0) when compared to Salers cheeses (3.4). Concerning the aromatic profile, Salers cheeses had a more intense global odor (6.3) than Cantal cheeses (5.8) and presented the highest flavors, such as odors of mushroom (1.8 vs. 1.6) and animal (3.8 vs. 3.0); vegetal (3.4 vs. 2.9), animal (2.8 vs. 2.2) and atypical (2.4 vs. 1.7) aromas; and a sour taste (3.8 vs. 3.3) compared to Cantal cheeses. In contrast, Cantal cheeses presented more intense pungent (3.5 vs. 3.1), bitter (3.3 vs. 2.7) and nutty (2.3 vs. 1.9) flavor notes compared to Salers cheeses. Poveda et al. (2008) [52] showed that an increased bitter taste may be related to high contents of octanoic and decanoic acids, which lead to a rancid and pungent taste in cheese. In contrast, the bitterness could be modulated by the presence of specific volatile compounds. In Salers cheese, the relative quantity of esters is more abundant than in Cantal cheese. Esters are widely associated with a pleasant aroma characterized by fruity and floral notes. Esters could reduce the bitterness and sharpness of cheeses, very often due to the high contents of amines and FFAs [31,53].

Seven texture attributes described the cheeses, but only the firmness (by touch and in the mouth) was significantly different between Cantal and Salers cheeses. Cantal cheeses had the highest firmness by touch and in the mouth (5.3 and 5.8, respectively). The softer texture and stronger flavor of Salers cheeses may also be explained by their higher levels of certain bacteria. Some studies show the relationships between the level of lactic acid bacteria in cheeses, which may contribute to their stronger flavor [54,55]. For example, Lactobacilli are known to have proteolytic activity in cheeses [56,57], increasing the production of small peptides and resulting in a higher secondary proteolysis [58], hence leading to the production of specific volatile compounds.

In addition, Lebecque et al. (2001) [13] observed a large diversity in texture in 25 Salers cheeses (3.5-month ripening time).

Regarding the milk origin, cheeses from OB_M had the highest score for color and rind thickness (6.6 and 6.1, respectively) compared to cheeses made with Salers_M (5.7 and 5.6). The core is described by a marbled (5.1) and cracked appearance (4.5). Concerning cheese flavor, cheeses with Salers milk (Salers_M) were characterized with the highest animal (2.8) and ammonia (1.6) aromas compared to OB_M. Concerning texture, cheeses made with OB_M were firmer by touch (7.0) and texture in the mouth (5.8) than those made with Salers_M (6.1 and 5.4, respectively). In the study conducted by Guiadeur et al. (2011) [59], cheeses made from Salers cow milk were firmer and more granular (less melty) than cheeses made from Holstein cow milk. On the other hand, cheeses with Salers milk were described by an intense aromatic profile. Indeed, they obtained the highest scores for animal and ammonia aromas.

Concerning the interactions of Category × Breed, four appearance attributes (rind color, rind thickness, color core and color homogeneity) and three flavor attributes (nutty, bitter and pungent) were significant. So, these interactions show that those sensory attributes depend both on the cheese category and the origin of the milk.

Figure 3 shows a PCA plot made based on the significant sensory attributes used to visualize the twelve cheeses, analyzed in triplicate. The first two components explained 49.38% of the total variance. The first component was characterized by flavor and appearance attributes (ammonia aroma, animal aroma, atypical aroma, core color, rind color and rind thickness). These attributes particularly describe Salers cheeses in contrast to Cantal cheeses. The second dimension enabled less discrimination and was characterized by firmness by touch and in the mouth for positive values on PC2 and by sticky quality on the negative values on PC2.

On the observation plot, there is a distinction between Cantal and Salers cheeses regarding the sensory properties (Figure 3C). However, we can observe that the ellipses are quite large, particularly for Salers cheeses, showing the sensory diversity for this cluster. In contrast, the ellipses overlap for cow breeds’ milk underlining, that this effect should have a smaller influence on the sensory qualities of cheeses (Figure 3B). However, it is recognized that Salers cheeses have great sensory diversity, which has been associated with a wide diversity of microbial dynamics [6,60].

3.4. Correlations Between Volatile Compounds and Flavor Sensory Attributes

The partial least squares regression (Figure 4) was constructed with the cheese samples (n = 12) using only the significant flavor sensory variables (Y, n = 11) and significant volatile compounds (X, n = 45) in order to analyze the correlations between flavor sensory attributes and volatile compounds.

The first three PLS factors explained 74% (Xcum) and 54% (Ycum) of the variance in the data. According to the results of Figure 4, Salers cheeses can be clearly distinguished from Cantal cheeses along the first axis of the PLS biplot.

PDO Salers cheeses were distinguished from PDO Cantal cheeses by their more intense flavor and sensory attributes (mushroom/animal odors, ammonia/vegetal/atypical/animal aromas and sour taste) and by their highest relative quantity of thirty-three VOCs. Among them, phenylethyl alcohol (44), butanoic acid, 3-methyl (36) and isopentyl hexanoate (30) were positively correlated with the aroma of ammonia describing H_Salers cheese. A vegetable aroma was positively correlated with ethyl acetate (1), ethanol (3), isobutyl acetate (5), butanoic acid, ethyl ester (6) and propanoic acid, 2-hydroxy-, ethyl ester (23). The mushroom odor specifically described the H_Salers cheese, which was correlated with phenylethyl alcohol (44), isopentyl hexanoate (30) and 3-methyl- butanoic acid (36). Generally, this odor could be associated with the specific molecule 1-octen-3-one, but in this case, this volatile compound was not detected [4]. Alcohols can generate a wide variety of aromas in cheese, for example, alcoholic, fruity, green, floral, rose, malty, honey, herbaceous, sweet, caramel and others [61,62].

In contrast (for negative values along the first axis of the PLS biplot), PDO Cantal cheeses were clustered in the PLS biplot due to their flavor attributes (bitter and pungent tastes and nutty aroma) and twelve VOCs. Only the bitter taste attribute was correlated with volatile compounds. Indeed, a bitter taste was clearly correlated with acetoin (20), nonanal (26), trimethyl-pyrazine (27), 2-hydroxy-3-pentanone (25) and 2,3-pentanedione (9), which describe A_Cantal_cheese in particular. Such correlations were unexpected but revealed interesting synergetic sensorial interactions in Cantal cheeses’ taste. On the one hand, bitterness can be generally linked to the peptide and amino acid compositions, and on the other hand, nonanal compounds were characterized by green grass and herbaceous aromas. Indeed, it has been shown that straight-chain aldehydes were very unpleasant when their concentrations exceeded certain thresholds. It could be supposed that high levels of the grass flavor of nonanal compounds and their association with other volatile compounds could increase the intensity of the bitter taste in Cantal cheese. We can suppose a similar phenomenon with nonanal compounds compared to ethyl hexanoate compounds (another aldehyde). Indeed, Castada et al. (2019) [63] identified a correlation between ethyl hexanoate and bitterness in Swiss cheese. Normally, at a low concentration, ethyl hexanoate carries a fruity flavor note in Swiss cheese. However, when the concentration increases, ethyl hexanoate gives a bitter perception in Swiss cheese.

It could be very interesting to conduct a further analysis to quantify the volatile compounds and determine the odor activities’ values [62] and identify those contributing to the aromas of Cantal or Salers cheeses. Indeed, among these significant compounds detected in this study, only five have previously been identified as odor-active compounds in Cantal-type cheeses by Cornu et al. (2009) [30]: nonanal compounds, ethanol, hexanal, acetic acid and butanoic acid.

4. Conclusions

This work provides additional information concerning the sensory and chemical characterization of PDO Salers and Cantal cheeses with a long ripening stage (>8 months). It was demonstrated that the cow breeds’ milk provenance poorly impacts the chemical and organoleptic quality of cheeses. On the other hand, the cheese category drives the sensory attributes and volatile compounds of PDO cheeses. PDO Salers cheeses are characterized by typical flavors with strong intensity (e.g., animal/ammonia/atypical/animal flavors) compared to PDO Cantal cheeses, which are characterized by a nutty aroma and bitter and pungent tastes. PDO Salers cheeses present a higher diversity of volatile compounds belonging to esters, alcohols and ketones. Our study also highlights the uniqueness of each cheese category (PDO Salers vs. PDO Cantal) with specific known VOCs resulting from the complex biotransformations occurring in cheeses due to cheese-making practices, microbial interactions and ripening times.

Our study also reveals the relationships existing between VOCs and sensory perceptions of cheese in the two cheese categories (PDO Salers vs. PDO Cantal), with both contributing to their organoleptic specificities. These results could be used in future research to increase our knowledge of the origin of cheese flavors.