Cheeses were studied to evaluate the consistency of chemical composition, acceptability of cheeses by a sensory panel. An exploration of cheeses, cheese spreads, and traditional butter by GC-MS was performed to confirm the differences in the aroma profile.

Cheeses and cheese spreads and butter were obtained from cheese producer Dubník, the Slovak Republic in three different batches. Samples of five types of hard cheese differ in flavouring spices and length of ripening. Hard cheeses were analyzed for basic physical and chemical composition by NIR and sensory panel. 11 samples of dairy products (including above mentioned 5 types of hard cheese, 5 types of cheese spreads, and one type of traditional butter) differed in the technological process of their production ). All 11 samples were analyzed by GC-MS for aroma profile determination. The hard cheese samples were analyzed by NIR and sensory panel.

All analyzed samples are characterized as follows: no. 1 – hard cheese with chili flakes „Ľahké pierko s čili“; no. 2 – 6 months rippening hard cheese „Kryštálová Kamenica“; no. 3 – 3 months rippening hard cheese „Sýta Rubaň“; no. 4 – hard cheese with chives „Jemný Dubník pažítka“; no. 5 – hard cheese with 4 spices „Ľahké pierko 4 korenia“; no. 6 – cheese spread with smoked cheese „Syrová nátierka s údeným syrom“; no. 7 – cheese spread with onion „Syrová nátierka sladká cibuľka“; no. 8 – cheese spread with hot paprika „Syrová nátierka pálivá paprika“; no. 9 – cheese spread with sweet paprika „Syrová nátierka dolniacka paprika“; no. 10 – cheese spread with emmental cheese „Syrová nátierka Ementálček“; no. 11 – traditional butter „Tradičné maslo“.

The analyses did not require special chemicals. We used deionized water and ethanol (96%, p.a., nondenatured, Centralchem Ltd., Slovakia) for washing the equipment.

No animal and biological material was used.

NIR spectrometer, type MPA (Bruker Optik GmbH, USA) was applied for the determination of the basic chemical composition of cheeses. Selected chemical parameters (dry matter, fat, salt, pH, proteins) were measured at the AgroBioTech SPU research center. The cheese samples were homogenized into a fine powder, then transferred to a Petri´s dish (type Duraplan) and placed in the apparatus. The NIR instrument was calibrated by O.K. SERVIS BioPro, Ltd. for hard cheeses in the range as follows: fat (5 – 43.6%), NaCl (0.04 – 2.85%), proteins (13.6 – 37.10%), dry matter (24.2 – 72.6%) and pH (4.93 – 6.35).

The aroma profile was analyzed by gas chromatography coupled to mass spectrometry (GC-MS). Agilent 7890B gas chromatograph coupled with an Agilent 5977A mass spectrometer (Agilent Technologies Inc., Palo Alto, CA, USA) was used. Separation was performed by a DBWAXms column (30 m × 0.32 mm × 0.25 μm; Agilent Technologies) working with the temperature program and MS conditions according to Sádecká et al. (2014). Identification of the compounds was performed by matching the measured peaks with Kovats´ retention indices with NIST library (The National Institute of Standards and Technology Library) by software Alpha Soft V14 (Alpha M.O.S.) (Štefániková et al., 2019).

The chemical composition results are summarized in Table 1 and Figure 1. The dry matter in the analyzed cheeses ranged from 54 to 68.4%. In general, it can be stated that cheeses containing herbs had a lower salt content compared to ripening cheeses such as “Kryštálová Kamenica” or “Sýta Rubaň”.

The fat content ranged from 17.42 to 30.88% and the salt from 1.2 to 2.1%. Proteins accounted for 27.4 to 32.43%. The acidity of the cheeses, characterized by a pH value, was in the range of 5.04 to 6.03. The highest values in the most of parameters were determined in Sýta Rúbaň cheese, except for the protein content.

The lowest values of these parameters were found in cheese with 4 types of spices, again except for protein content. A visual expression of the consistency of the individual cheese chemical composition data is shown in Figure 1. In our case, the results indicate the consistency of the chemical composition of the samples over time. Since the data were collected from 3 series of measurements during the 3 months (the cheese producer provided samples at monthly intervals), the results reflect the stability of the chemical composition of the products over time. As can be seen in the case of dry matter content, the closest data were measured for the “Kryštálová Kamenica” and “Ľahké pierko s čili” cheeses. “Kryštálová Kamenica” is a cheese that has been ripening for 6 months, therefore the composition of this cheese becomes stable. Cheeses such as “Sýta Rubaň” and “Ľahké pierko 4 korenia” had a higher dry matter content, which may be related to seasonal changes in milk parameters. In the case of fat content, the closest data were observed in the “Sýta Rúbaň” cheese and the largest variance was recorded in the “Ľahké pierko 4 korenia”. The salt as a parameter showed a relatively constant proportion in cheeses such as “Jemný Dubník pažítka“, “Ľahké pierko s čili” and “Ľahké pierko 4 korenia”. We found the highest variability of the salt content values in the “Sýta Rubaň” cheese. The pH values were most balanced in the “Kryštálová Kamenica” and “Jemný Dubník pažítka”. On the other hand, “Ľahké pierko 4 korenia” had the highest variance in pH values. In the protein content, we recorded the most consistent data for all analyzed cheeses.

Different types of cheese were analyzed in literature to evaluate basic chemical composition. Ricotta cheeses were analyzed by Madalozzo et al. (2015) NIR method for protein and fat content. The average fat content was 10.63%, protein 12.37%, and moisture 70.23%. Margolies and Barbano (2018) analyzed Cheddar cheeses using the MIR technique. The average fat content was 34.0%, the protein content 24.0%, the moisture content 36.82% and the salt 1.8%. Manuelian et al. (2017) analyzed commercial cheeses using the NIT technique. The moisture, fat, and protein contents averaged 43.24 ±0.97%, 27.24 ±0.47%, and 24.87 ±0.54%. Jo et al. (2018) analyzed the sensory and chemical properties of Gouda cheeses. All cheeses had a water content of less than 45%, a fat content of more than 46%, and a pH value in the range of 4.9 to 5.6. The salt content of the cheese differs significantly from the type of cheese, ranging from about 0.5 – 0.7% in acidic cottage cheese and Emmental cheeses to about 4-6% in pickled cheeses such as Domiati and feta (Guinee, 2004).

Kuflu cheese, a Turkish mould-ripened variety was studied by Hayaloglu et al (2008). This type of cheese was characterized by moisture content 49.97%, 12.18% fat content, 37.84% total protein contents, and pH 6.29. Our results corresponded most to the chemical parameters in ripening Cheddar cheeses.

Sensory analysis was performed in the sensory laboratory. Trained sensory panels are important tools for assessing the quality of food and non-food products (Tomic et al., 2010). Sensorial features of cheese are more or less a consequence of the combined actions of all steps involved in its processing (Kongo and Malcata, 2015). The organoleptic properties of the individual samples were evaluated using a 1 – 9 point test (1 – insufficient to 9 – very intense). The sensory panel evaluated the attributes: appearance, overall aroma, aroma odour, consistency, overall taste, flavour, overall acceptability. The data obtained from the sensory evaluation were processed and statistically tested by the Shapiro-Wilk test for normality at a significance level of P=0.05. A principal component analysis (PCA) was used to determine element variability and visualization of sensory analysis data. The result is PCA biplot (Figure 2). For a better interpretation of the results obtained using the statistical method of analysis of the main components, we used visualization using the so-called biplot. This PCA map represents a biplot that shows simultaneously our observations as well as the variables examined in a single graph. The coordinate axes (x and y) of the biplot explain 79.52% of the total variation of observation.

This type of graph allows a better interpretation of the share of individual characteristics in the main components. A PCA biplot shows both PC scores of samples (dots) and loadings of variables (vectors). The further away these vectors are from a PC origin, the more influence they have on that PC.

In our biplot, the most important variables in the sensory evaluation were overall acceptability and appearance, almost no influence had consistency and aroma odour. The results of the sensory analysis (Figure 2) showed that sample no. 5 “Ľahké pierko 4 korenia” was rated within all attributes as the best. Sample 3, 2, and 1 followed. Sample 4 “Ľahké pierko s pažítkou” received the lowest score. Toro, Valencia, and Molina (2016) evaluated nutritional parameters and sensory features of processed cheese. For the development of this type of food product, it is necessary to consider nutrition value for different groups of people and its versatility and ease of consumption.

The GC-MS method is often used to characterize the aromatic profile of dairy products. We have found the presence of the following aromatic compounds. The volatile compounds that were identified belong to several chemical groups: alcohols, aldehydes, ketones, carboxylic acids, esters, amides, amines, imines, and terpenes (Table 2a and 2b). Data obtained by GC-MS analysis and area percentages were visualized by heat map (Figure 3). The vertical dendrogram represents volatile compounds in the samples and the horizontal dendrogram the reordered samples. The heat map reflects the range of area percentages as well as the proportion of specific volatile aromatic compounds in the individual samples. The clustering showed that the compounds are divided into four groups (left dendrogram) according to the observed similarities. The pattern inside the map shows the groups of compounds described above as follows: the first cluster is represented by the compounds 5- methyl-2-heptene to 2-methyl-butanal, the second cluster is represented by compounds from acetone to propanoic acid. The third group can be clearly distinguished by the presence of 2-pentanone to butanoic acid. This means agglomeration appears to have moderate amounts of most of the compounds. Compounds from D-limonene to 2,3- butanedione represent group four. Horizontal clustering revealed the fact that almost no similarities were observed between the samples, based on the composition of the aroma profile and the amount of volatiles.

Aldehydes (specifically 3-methylbutanal and benzaldehyde) were an important group of volatile compounds in cheeses with a presence in several samples. Due to their low odour detection thresholds, these compounds contribute significantly to the aroma features of the cheeses. Among the carbonyl compounds, ketones were also represented, namely acetoin, 2,3,3-trimethylcyclobutanone, 2-heptanone, 2,3-butanedione, acetone, 2- butanone, 2-nonanone, 2-pentanone. Alcohols also have a low odour detection threshold, which gives dairy products herbal, woody and fatty tones. We identified alcohols represented by 3-methyl-1-butanol in our cheese samples. Terpenes (specifically β-phellandrene, α-phellandrene, β- myrcene, o-cymene, D-limonene, and caryophyllene) were also identified in sample 5 (cheese “Ľahké pierko 4 korenia”) to confirm the spicy aroma of the sample. Carboxylic acids such as acetic acid, butanoic acid, caproic acid (hexanoic acid), heptanoic acid, malonic acid, and 3- methyl-butanoic acid were present in the cheese samples.

Sample 1 with chive flavour had the largest proportion of acetic and butyric acid. In sample 2 with chili flavour, acetoin was present in the highest concentrations, followed by 3-methyl-1-butanol, 2,3-butanedione. Sample 3 was characterized by the highest proportion of acetoin and acetic acid, followed by (2-aziridinylethyl)amine. In sample 4 (Kamenica matured Parmesan-style cheese) the most abundant was butanoic acid, followed by acetone.

In the last sample were dominant 2,3-butanedione, (2- aziridinylethyl) amine, acetoin, D-limonene, followed by terpenes such as α-phellandrene, caryophyllene, o-cymene, β-phellandrene, which originated from the addition of spices.

Cheddar cheeses were analyzed on aroma profile by GCMS by the authors Chen et al. (2020). They found fortynine flavour compounds in the samples, including 11 acids, 5 aldehydes, 8 alcohols, 8 ketones, 9 esters, 3 aromatic hydrocarbons and heterocyclic compounds, and 4 other compounds. The active ingredients were mainly aldehydes (2-methylbutanal, 3-methylbutanal, benzaldehyde), ketones (2-heptanone, 2-octanone), carboxylic acids (butyric, isovaleric acid, octanoic acid, n-decanoic acid, hexanoic acid), and esters (ethyl hexanoate, ethyl acetate, ethyl butyrate, and butyl acetate). The authors identified, according to the value of the aroma intensity (AI values), the compounds with the highest aromatic activity in the cheese. Several compounds have relatively high aroma intensity values such as butyric acid (cheesy/rancid), ethyl acetate (fruity/sweet), 2-heptanone (cheesy/spicy/fruity), ethyl butyrate (fruity/sweet), benzaldehyde (almond), 3- methylbutanal (nutty/malty), 2-methylbutanal (nutty/chocolate), and 3-methylbutanoic acid (cheesy/sweaty/rancid). Of these, the odours of 3- methylbutanal and 2-methylbutanal were rated by the sensory evaluation panel as pleasantly nutty, and the almond odour imparted by benzaldehyde was also considered to be a good nutty flavour. Štefániková et al. (2019) analyzed by GC-MS traditional slovak „Parenica“ cheese. A Parmesan and parmesan-style cheese aroma were studied by Barbieri et al. (1994), Moio and Addeo (1998). The backbone of Grana Padano cheese aroma seemed to consist of carboxylic acids and 14 potent neutral odourants imparting fruity, green, nutty, and coconut notes. The authors reported that the concentration of volatile components responsible for the fruity and green notes was inversely proportional to the length of ripening, whereas the concentration of volatile agents with spicy, nutty, and earthy notes tended to increase during maturation. Bryndza cheese aroma profile was studied by Sádecká et al. (2014). The overall aroma of Bryndza cheese was found to be formed by the following volatile components: esters, alcohols, aldehydes, ketones, fatty acids, and hydrocarbons. Nineteen of twenty-seven aroma-active compounds were identified. Mascarpone cheese aroma was analyzed by Capozzi et al. (2020) and the volatile compounds contributing to aroma were classified into 9 chemical classes (9 ketones, 5 alcohols, 4 organic acids, 3 hydrocarbons, 2 furans, 1 ester, 1 lactone, 1 aldehyde, and 1 oxime).

The most abundant compounds were ketones (especially 2-heptanone and 2-pentanone), then 2-propanone, 2- nonanone, 2-butanone, 1-pentanol, 2-ethyl-1-hexanol, furfural, and 2-furan methanol. Thomsen et al. (2014) studied different French commercial cheeses belonging to the same type: non-cooked semi-hard cheeses, including famous types like French Port Salut or Gouda. The volatiles found in the samples belonged to sulfur compounds (e.g. dimethyl disulfide, described as cauliflower, garlic flavour), alcohols (3-methyl butane-1-ol was found in the greatest amount and this alcohol has been described in cheeses with a fruity aroma (Pérès, Viallon and Berdagué, 2001), aldehydes (e.g. 3-methylbutanal, responsible for cheesy aroma), ketones (the most abundant ketone was 3- hydroxybutan-2-one or acetoine, perceived alone with a dairy aroma), carboxylic acids (the most abundant carboxylic acid was 3-methylbutanoic acid, described as a potent odourant in goat cheese flavour (Poveda et al., 2008), esthers (e.g. ethyl acetate and ethyl propanoate), lactones and furans.

Our results corresponded more closely with analyzes of the Cheddar cheeses, which belong to ripening cheeses, and therefore we also found these compounds in our samples of ripening cheeses. The similarities in dominant compounds were also found with non-cooked semi-hard cheeses.

Processed cheese is an attractive product that enjoys great popularity. The basic raw material for processed cheese and spreads are natural cheeses and other dairy materials and non-dairy raw materials, as well as emulsifying salts (Černíková et al., 2017).

These components contribute to the aroma of the products. The aroma profile of the cheese spreads in our study was characterized by the following aromatic compounds.

Oxygen derivatives of hydrocarbons such as alcohols (represented by 2-methyl-1-butanol, 3-methyl-1-butanol, and cyclopropane-1,2,3-d3-methanol), aldehydes (2- methylbutanal, 3-methylbutanal, benzaldehyde) occurred ), ketones (acetoin, 2-heptanone, 2-pentanone, 2-butanone, and 2,3-butanedione), carboxylic acids (propanoic acid, butanoic acid, hexanoic acid, octanoic acid, acetic acid, 2- acetonyl-3-cyano-2,3-dimethylcyclobutane-1-carboxylic acid) and esters (ethyl butanoate and ethyl hexanoate), the nitrogen derivatives were amines (dimethylamine) and oximes (methoxyphenyloxime) and amides (butyl(formyl) formamide).

Acetoin was dominant in sample 6 with the addition of smoked cheese, followed by 2-heptanone, butyl (formyl) formamide, 2-pentanone. Sample 7 with onion flavour had the largest proportion of hexanoic acid, butanoic acid, and acetoin.

Sample 8 was spicy with the addition of hot peppers, and hexanoic acid, butanoic acid, and acetoin were the most abundant. Acetoin and 2-pentanone had the largest proportion in sample 9. And in the last Emmental-flavoured sample 10, propanoic acid was the most abundant, followed by acetoin, acetic acid, cyclopropane-1,2,3-d3-methanol.

Ningtyas et al. (2019) determined flavour profile in reduced-fat cream cheese. The volatile profiles of the cream cheese contained β-glucan, esterified phytosterols, and emulsified phytosterols. Twenty-nine volatile compounds at varying levels were identified: 9 fatty acids, 9 ketones, 9 aldehydes, 2 alcohols. Additionally, high levels of acetic acid (vinegar), hexanoic acid (pungent/sour), octanoic acid, and decanoic acid (rancid/fatty) were also identified in the samples. El-Wahed and Hassanien (2019) analyzed sensory characteristics of cheese spreads and cheese spreads analogues. In control, processed cheese spread the highest concentration of major short-chain volatile fatty acids namely butyric, caproic, caprylic, and capric acids were found. Hickey et al. (2018) reported volatile compounds from the comparative study of processed Cheddar cheese. The treatment of the experimental variant lied also of buttermilk. The differences were found in the abundance of the compounds as follows: hexanol, 2-methyl-propanal, 2- methyl-butanal, benzeneacetaldehyde, nonanal, butanoic acid, hexanoic acid, octanoic acid, methional, σ- decalactone.

Butter has appreciated sensory attributes such as flavour, aroma, and melting near the body temperature, which results in a pleasant sensation in the mouth (Galindo- Cuspinera, Valenҁa de Sousa and Knoop, 2017). The volatile compounds methoxyphenyloxime, benzaldehyde, and benzyl alcohol were found in our butter samples. Tamura et al. (2021) analyzed butter for volatile compounds and their contribution to fresh sweet creamy butter flavour. Out of the sixty-four identified compounds, 23 were selected, which, according to the authors, had the most significant effect on the aroma. They belonged here e.g. compounds such as δ-dodecalactone, δ-decalactone, γ- decalactone, acetaldehyde, and decanoic acid (capric), which were the most prominent aromatic compounds responsible for the butter flavour. Yoshinaga et al. (2019) also confirmed the presence of lactones in butter and fermented butter. The main lactones were δ-decalactone, δ- dodecalactone, δ-tetradecalactone, δ-hexadecalactone, and γ-dodecalactone. In margarine, dominant lactones were δ- decalactone and δ-dodecalactone. Our results, represented by benzaldehyde with a fruity aroma, benzyl alcohol with floral tones, and methoxyphenyloxime, the aroma of which is not described, do not correspond to the results of the above study.

We assume that an insufficient number of standards have influenced the identification of GC-MS compounds, so a more detailed study of the aroma profile in butter would need to take this into account, and investment in standards for compounds that have the highest probability to occur in butter should be considered.