Consumer perception of smoked Cheddar cheese.

Consumer perception of smoked cheese was evaluated through focus groups, surveys, and central location testing. Three focus groups (n = 29) were conducted with consumers of smoked cheese. Subsequently, 2 online surveys were conducted. The purpose of the first survey (n = 1,195) was to understand types of smoked cheeses consumed and if consumers associated specific wood smokes with smoked cheese. Next, an adaptive choice-based conjoint (n = 367) was designed to evaluate consumer perception of different attributes of smoked cheese. Maximum difference scaling and familiarity questions were also included in the adaptive choice-based conjoint survey. Following the surveys, a central location test (n = 135) was conducted with cheeses smoked with 3 different woods at a low and high intensity (6 cheeses total). Hierarchical Bayesian estimation, 1-way ANOVA, agglomerative hierarchical clustering, and 2-way ANOVA (smoke type × intensity level) were used to interpret the collected data. Results from the focus groups indicated that smoked cheese was perceived as an artisan, high-end product and that appearance and price were strong purchase factors. In general, consumers were not aware of how smoked flavor was imparted to cheese, but when informed of the processes, they preferred cold-smoked cheese to the addition of liquid smoke flavor. Results from both surveys confirmed focus group observations. Consumers perceived flavor differences among different wood smokes and smoked products. Method of smoking, smoke intensity, type of wood, and type of cheese were the most important attributes for purchase of smoked cheese. When tasting, consumers differentiated smoke aroma and flavor among cheeses and preferred cherry wood smoked cheeses over apple wood or hickory smoked cheeses. Understanding consumer perceptions of smoked cheese will give insight into the desired experience that consumers expect when purchasing smoked cheese.

Uma análise histórico-crítica sobre o desenvolvimento das normas brasileiras relacionadas a queijos artesanais / [A critical-historical analysis of the continuous development of Brazilian legislation related to artisanal cheeses]

Diferentes tipos de queijos artesanais são produzidos, comercializados e consumidos no Brasil, o que impulsiona o constante desenvolvimento de normas por órgãos oficiais, como o Mapa. A criação do Suasa e do Sisbi-POA foi fundamental para esse setor, por permitir um sistema de equivalência na fiscalização e por ampliar a distribuição. Ainda, o Mapa passou a permitir que queijos artesanais produzidos com leite cru pudessem ser maturados em um período inferior a 60 dias, desde que comprovada sua inocuidade. A redução do tempo de maturação é um tema controverso e polêmico, já que não há critérios específicos que estudos científicos devem contemplar, o que permite múltiplas interpretações de dados. Com a criação e a regulamentação do selo Arte, a fiscalização dos produtos artesanais foi designada aos órgãos de agricultura, pecuária e de saúde pública, em complementação à atribuição já prevista pelo Mapa e pelo Sisbi-POA. Ainda, o selo Arte atribui aos órgãos de inspeção uma função orientadora, atividade que deveria ser prioritariamente executada por agências de extensão e associações. As normas que balizam a produção e comercialização de produtos artesanais devem ser frequentemente atualizadas, devido aos constantes avanços científicos na área e para assegurar a oferta de produtos com qualidade e inócuos aos consumidores.(AU)

Different artisanal cheeses are produced, commercialized and consumed in Brazil, leading to a constant development of related rules by the MAPA and other official agencies. The establishment of two national programs (SUASA and SISBI-POA) allowed an equivalence in inspection system and an expanded distribution. Also, MAPA allowed ripening time lower than 60 days for artisanal raw milk cheeses, based on scientific studies that assure their safety. However, lowering the ripening period is still controversial, once there are no proper established criteria for such scientific studies, leading to potential multiple interpretation of data. The newly established ARTE certification transferred the inspection responsibilities of artisanal products to secretaries of agriculture, livestock and health, in support of what was already predicated by MAPA and SISBI-POA. Based on ARTE certification, the inspection service must also provide orientation guidance to producers, which should be done specifically by extension organs and associations. The norms that guide the production and commercialization of these artisanal products often need to be updated, but based on well-established methodologies and procedures, to ensure the distribution of suitable products to consumers.(AU)

Non-thermal atmospheric gas plasma for decontamination of sliced cheese and changes in quality.

The atmospheric-pressure non-thermal dielectric barrier discharge (DBD) plasma has recently emerged as an efficient decontamination method for the food safety enhancement. Thus the objective of this study was to evaluate the effectiveness of a simple DBD plasma treatment, with a relatively low-frequency power supply operating at 60 Hz, for microbial inactivation. A parametric study of operating conditions for bacterial inactivation was conducted using nutrient agar inoculated with Escherichia coli (2.28--6.28 log CFU/ml). The microbial log reduction was enhanced with increasing input power (30, 50, 70 W) and plasma exposure time (0, 1, 3, 5, 7 min). The inactivation effect was increased by decreasing inter-electrode gap (2, 1.5, 1 cm) and by reducing the initial microorganism concentration. Accordingly, a DBD plasma treatment at 50 W for 10 min could lead to complete killing of E. coli and partial inactivation of Listeria innocua on cheese (mean log reduction: 4.75 ± 0.02 and 0.72 ± 0.01, respectively). The decontamination efficacy of DBD plasma was affected by the types of microorganisms. The changes in hardness and color of cheese were unnoticeable after 10 min treatment with a power of 50 W. Overall, the results suggested that the DBD plasma can be potentially exploited to improve the food safety.

Avaliação da composição de queijos parmesão ralados de diferentes marcas comerciais / Assessment of the composition of parmesan cheeses rolled from different trademarks

O queijo parmesão é obtido a partir de massa pré-cozida e prensada, caracterizado por apresentar baixa umidade e tempo de maturação igual ou superior a 6 meses. Objetivou-se analisar a composição de amostras comerciais de queijos parmesão ralados. Foram adquiridas 3 diferentes marcas (A, B, C) de queijo parmesão ralado, em embalagens hermeticamente fechadas, submetidas a análises de composição química e características físico-químicas. Os resultados foram analisados estatisticamente por meio de análise variância univariada (ANOVA) e as médias foram submetidas ao teste Tukey ao nível de 5% de probabilidade. Observou-se que uma das amostras diferenciou-se estatisticamente das demais, estando fora dos padrões legais exigidos pela legislação para queijo parmesão ralado, saindo dos padrões para queijo parmesão ralado.

Avaliação de métodos de contagem de BAL e da qualidade microbiológica de queijo Minas Frescal comercializado no município de Duque de Caxias/RJ / Microbiological analysis of “Minas Frescal” cheese commercialized in the municipality Duque de Caxias/RJ

O queijo Minas Frescal é um derivado lácteo de amplo consumo no Brasil, entretanto suas características favorecem o crescimento de microbiota contaminante. As amostras foram submetidas às seguintes análises microbiológicas: enumeração de coliformes a 35⁰C e a 45⁰C, contagem de Staphylococcus coagulase positiva e enumeração de Bactérias Ácido Láticas (BAL). Em 10 amostras foi observada contagem de coliformes a 45⁰C acima do padrão estabelecido na legislação, confirmadas como Escherichia coli. Em todas as amostras a contagem de Staphylococcus coagulase positiva foi superior ao permitido em legislação. Nas BALs, foi observada baixa contagem em todas as amostras. Diante dos resultados obtidos, se faz necessária a adoção de medidas a fim de melhorar a qualidade microbiológica da matriz alimentícia e garantir a inocuidade alimentar.

Características dos estabelecimentos produtores e do processo de fabricação do queijo colonial da microrregião de Francisco Beltrão-PR / Characteristics of the production establishments and the process of manufacturing the colonial cheese of the Francisco Beltrão-PR microregion

O queijo Colonial é tradicionalmente presente e apreciado na região Sudoeste do Paraná, e muito frequentemente fabricado de forma artesanal, o que ilustra sua importância socioeconômica. No entanto, são escassos os dados sobre este queijo, não há padrão de identidade e nem legislação a seu respeito. Considerando o exposto, este trabalho teve por objetivo a elaboração de fluxos de processo e verificação das condições de boas práticas de fabricação em estabelecimentos produtores de queijo Colonial. Foram avaliados 5 laticínios da microrregião de Francisco Beltrão – PR, onde foram verificadas diferenças sobretudo quanto ao tratamento do leite utilizado, adequação às boas práticas de fabricação, controle de qualidade e inconformidades quanto aos padrões microbiológicos.

Short communication: Physicochemical and antioxidant properties of Cheddar-type cheese fortified with Inula britannica extract.

Cheddar-type cheese was fortified with the antioxidant Inula britannica flower extract (IBE). Cheddar-type cheeses manufactured with varying concentrations of IBE (0, 0.25, 0.5, 0.75, and 1% wt/vol) were analyzed during storage at 4°C, 0, 1, 2, and 3 wk after production. Higher IBE concentrations resulted in higher protein and ash contents, with a concomitant decrease in pH, total solid, and fat content relative to the unfortified control cheese. The total phenolic content also increased with IBE concentration, but decreased over longer storage periods. The antioxidant activities of the cheeses, determined as 2,2-diphenyl-1-picrylhydrazyl (DPPH) free-radical scavenging activity and ferric thiocyanate assay results, increased proportionally to the total phenolic content. The highest antioxidant effect was observed in the 1% IBE-fortified cheese, showing 79 and 86% antioxidant effects in the DPPH and ferric thiocyanate assays, respectively. At the 1-wk time point, the 5 cheese preparations underwent sensory evaluation for odor, taste, texture, color, and overall quality, determined using a descriptive analysis by a trained panel (n=20). The addition of IBE resulted in some increases in extract odor and taste. Overall, IBE showed good potential as an antioxidant supplement for dairy products.

Monitoring the ripening process of Cheddar cheese based on hydrophilic component profiling using gas chromatography-mass spectrometry.

We proposed an application methodology that combines metabolic profiling with multiple appropriate multivariate analyses and verified it on the industrial scale of the ripening process of Cheddar cheese to make practical use of hydrophilic low-molecular-weight compound profiling using gas chromatography-mass spectrometry to design optimal conditions and quality monitoring of the cheese ripening process. Principal components analysis provided an overview of the effect of sodium chloride content and kind of lactic acid bacteria starter on the metabolic profile in the ripening process of Cheddar cheese and orthogonal partial least squares-discriminant analysis unveiled the difference in characteristic metabolites. When the sodium chloride contents were different (1.6 and 0.2%) but the same lactic acid bacteria starter was used, the 2 cheeses were classified by orthogonal partial least squares-discriminant analysis from their metabolic profiles, but were not given perfect discrimination. Not much difference existed in the metabolic profile between the 2 cheeses. Compounds including lactose, galactose, lactic acid, 4-aminobutyric acid, and phosphate were identified as contents that differed between the 2 cheeses. On the other hand, in the case of the same salt content of 1.6%, but different kinds of lactic acid bacteria starter, an excellent distinctive discrimination model was obtained, which showed that the difference of lactic acid bacteria starter caused an obvious difference in metabolic profiles. Compounds including lactic acid, lactose, urea, 4-aminobutyric acid, galactose, phosphate, proline, isoleucine, glycine, alanine, lysine, leucine, valine, and pyroglutamic acid were identified as contents that differed between the 2 cheeses. Then, a good sensory prediction model for "rich flavor," which was defined as "thick and rich, including umami taste and soy sauce-like flavor," was constructed based on the metabolic profile during ripening using partial least squares regression analysis. The amino acids proline, leucine, valine, isoleucine, pyroglutamic acid, alanine, glutamic acid, glycine, lysine, tyrosine, serine, phenylalanine, methionine, aspartic acid, and ornithine were extracted as ripening process markers. The present study is not limited to Cheddar cheese and can be applied to various maturation-type natural cheeses. This study provides the technical platform for designing optimal conditions and quality monitoring of the cheese ripening process.

Proteolysis and biogenic amine buildup in high-pressure treated ovine milk blue-veined cheese.

Penicillium roqueforti plays an important role in the ripening of blue-veined cheeses, mostly due to lactic acid consumption and to its extracellular enzymes. The strong activity of P. roqueforti proteinases may bring about cheese over-ripening. Also, free amino acids at high concentrations serve as substrates for biogenic amine formation. Both facts result in shorter product shelf-life. To prevent over-ripening and buildup of biogenic amines, blue-veined cheeses made from pasteurized ovine milk were high-pressure treated at 400 or 600 MPa after 3, 6, or 9 wk of ripening. Primary and secondary proteolysis, biogenic amines, and sensory characteristics of pressurized and control cheeses were monitored for a 90-d ripening period, followed by a 270-d refrigerated storage period. On d 90, treatments at 400 MPa had lowered counts of lactic acid bacteria and P. roqueforti by less than 2 log units, whereas treatments at 600 MPa had reduced lactic acid bacteria counts by more than 4 log units and P. roqueforti counts by more than 6 log units. No residual α-casein (CN) or κ-CN were detected in control cheese on d 90. Concentrations of ß-CN, para-κ-CN, and γ-CN were generally higher in 600 MPa cheeses than in the rest. From d 90 onwards, hydrophilic peptides were at similar levels in pressurized and control cheeses, but hydrophobic peptides and the hydrophobic-to-hydrophilic peptide ratio were at higher levels in pressurized cheeses than in control cheese. Aminopeptidase activity, overall proteolysis, and free amino acid contents were generally higher in control cheese than in pressurized cheeses, particularly if treated at 600 MPa. Tyramine concentration was lower in pressurized cheeses, but tryptamine, phenylethylamine, and putrescine contents were higher in some of the pressurized cheeses than in control cheese. Differences in sensory characteristics between pressurized and control cheeses were generally negligible, with the only exception of treatment at high pressure level (600 MPa) at an early ripening stage (3 wk), which affected biochemical changes and sensory characteristics.

Characterization by solid-phase microextraction-gas chromatography of the volatile profile of protected designation of origin Montasio cheese during ripening.

Montasio is a typical protected designation of origin (PDO) Italian semi-hard and semi-cooked cheese produced in northeast Italy from raw or thermized cow's milk. The PDO label implies that the product has distinctive characteristics that are connected to traditional production methods. The aim of this work was to precisely characterize the volatile fraction of this Italian cheese. The volatile profile can be considered a fingerprint because the flavor of a cheese variety is the result of a specific balance between the volatile compounds produced during the ripening process. Analysis of the volatile profile of Montasio cheese was performed by solid-phase microextraction-gas chromatography. Six cheesemaking trials were performed, each in a different dairy located within the Montasio cheese production area. Cheeses were analyzed at 5 stages of ripening (60, 90, 170, 300, and 365 d). Only 11 compounds were identified and measured: 5 fatty acids, 3 alcohols, 2 ketones, and 1 ester. The limited number of volatile compounds measured in the headspace of the Montasio cheese is probably due to the specific making process of this cheese, which affects evolution of the microflora and the biochemical processes of ripening. The total volatile fraction profile progressively increased from 60 to 170 d, after which time it remained almost steady. The most important contributors were found to be ethanol, short-chain fatty acids (C(2) to C(6)), diacetyl, and ethyl hexanoate. Ethanol and short-chain fatty acids increased up to 170 d, diacetyl increased up to 300 d and then declined, and ethyl hexanoate increased until the final stage.

Effect of emulsifying salts on the physicochemical properties of processed cheese made from Mozzarella.

The aim of this study was to investigate the effect of different types and concentrations of emulsifying salts (trisodium citrate, tetrasodium pyrophosphate, sodium tripolyphosphate, sodium hexametaphosphate, and disodium orthophosphate) on the physicochemical properties of processed cheese. The physicochemical composition, texture profile, degree of casein dissociation, fat particle size, color, and nuclear magnetic resonance profile (NMR) of processed cheese were determined. Hardness, degree of casein dissociation, and pH increased as the concentration of emulsifying salts increased. The fat particle size of processed cheese was significantly influenced by the type of emulsifying salts, with processed cheese made with sodium hexametaphosphate having larger particles (4.68 µm) than cheeses made with the other salts (from 2.71 to 3.30 µm). The processed cheese prepared with trisodium citrate was whiter than those prepared with the other emulsifying salts. The NMR analysis showed that the relaxation time of processed cheese of 10 to 100 ms accounted for a major proportion, indicating that the moisture in processed cheese was mainly bound water combined with the fat globule and hydrated casein.

Removal of cadmium and lead from aqueous solution by Enterococcus faecium strains.

Foods may be contaminated with heavy metals, which, even in small quantities, possess detrimental effects on human health. The aim of the present study was to investigate the uptake of cadmium or lead from an aqueous medium frequently found in foods, by 2 Enterococcus faecium strains (E. faecium EF031 and E. faecium M74). Also, the effects of the bacterial viability, incubation (contact) time, and pH on the binding capacities and binding stability were assessed. The results showed that both of the strains efficiently removed cadmium and lead. While EF031 removed 77.3% to 98.1% of cadmium and 66.9% to 98.9% of lead, M74 removed 53.5% to 91% of cadmium and 42.9% to 93.1% of lead throughout a 48 h incubation period at pH 5. It was found that, at 1 h, EF031 and M74 strains removed cadmium and lead, which was more than 60% of total removed cadmium and lead throughout the whole incubation period of 48 h. It suggests that the uptake of cadmium and lead by EF031 and M74 strains is a rapid process. The binding of both heavy metals increased with increasing pH of an aqueous medium and was the highest at pH 5. Also, the complexes formed between both heavy metals and bacterial cells were found to be stable. These findings indicate that E. faecium strains used in the study are able to bind the 2 heavy metals and may be used in the production of fermented functional foods, which will be healthy via its detoxification properties.

Multivariate analysis of proteolysis patterns differentiated the impact of six strains of probiotic bacteria on a semi-hard cheese.

The individual contribution of 6 strains of probiotic bacteria (3 of Lactobacillus acidophilus and 3 of the Lactobacillus casei group) to proteolysis patterns in a semi-hard cheese was assessed. Control cheeses (without probiotics) and 2 types of experimental cheeses (with the addition of probiotics either directly to milk or by a 2-step fermentation method) were manufactured. Cheeses containing Lb. acidophilus showed the most extensive peptidolysis, which was evidenced by changes in the peptide profiles and a noticeable increase of free amino acids compared with control cheeses. The strains of the Lb. casei group showed a lower contribution to cheese peptidolysis, which consisted mainly of free amino acid increase. Two-step fermentation improved peptidolytic activity for only one of the cultures of Lb. acidophilus tested. The addition of Lb. acidophilus strains into cheese may be suitable not only for their beneficial health effect but also for their influence on secondary proteolysis, consistent with acceleration of ripening and improved flavor formation.

Isolation of high quality DNA: a protocol combining “rennet” and glass milk

High quality DNA is essential for many molecular biology techniques. However, the reagents used for that purpose usually are expensive and/or cause a high environmental impact. Here, we describe two alternative protocols that use inexpensive reagents and are not hazardous to the environment. The first protocol utilizes the enzyme chymosin, normally used as “rennet” in cheese production and which is easily obtained on the commercial market. The second protocol uses “rennet DNA extraction protocol” combined with the DNA binding capacity of glass powder (glass milk), which can easily be “home made”. The first protocol is used when a high yield of DNA is needed, whereas the second protocol is used for production of a higher quality DNA, being able to work with sparse samples.

Effects of gamma-irradiation on Listeria monocytogenes population, colour, texture and sensory properties of Feta cheese during cold storage.

Feta, a white brine cheese, was produced and contaminated with Listeria monocytogenes. Contamination occurred either at the beginning (pre-process contamination) or at the end of Feta manufacturing (post-process contamination). In the first case the milk was contaminated with 10(3) cfu/ml, and 2 months later, in the final product, the L. monocytogenes population was approximately 10(5) cfu/g. In the second case, the brine (NaCl, 7% w/v), in which the Feta was packaged, was contaminated with 10(3) cfu/ml. Contaminated Feta samples were vacuum-packaged and exposed to irradiation doses of 1.0, 2.5 and 4.7 kGy and stored at 4 degrees C for a month. In the pre-process contaminated samples none of the irradiation doses eliminated L. monocytogenes; however the highest dose reduced the viable population to a level which is in compliance with EC regulations. In the post-process contamination, the 2.5 kGy and 4.7 kGy doses reduced L. monocytogenes counts below the detection limit. Irradiation had no effect on the texture of Feta. Irradiation at 4.7 kGy increased Feta's redness and decreased its yellowness and lightness. Sensorial analyses showed that at the 4.7 kGy dose, the aroma profile of Feta was temporarily affected, since it was restored after 30 days of cold storage.

Diversity of lactic acid bacteria population in ripened Parmigiano Reggiano cheese.

The diversity of dominant lactic acid bacteria population in 12 months ripened Parmigiano Reggiano cheeses was investigated by a polyphasic approach including culture-dependent and independent methods. Traditional plating, isolation of LAB and identification by 16S rDNA analysis showed that strains belonging to Lactobacillus casei group were the most frequently isolated. Lactobacillus helveticus, Lactobacillus delbrueckii subsp. lactis, Lactobacillus parabuchneri, and Lactobacillus buchneri species were detected with lower frequency. PCR-denaturing gradient gel electrophoresis (DGGE) applied to DNA extracted directly from cheese samples and sequencing of rDNA amplicons confirmed the complex microbiological pattern of LAB in ripened Parmigiano Reggiano cheeses, with the significant exception of the Lactobacillus fermentum species, which dominated in several samples, but was not detected by cultivation. The present combination of different approaches can effectively describe the lactic acid bacteria population of Parmigiano Reggiano cheese in advanced stages of ripening, giving useful information for elucidating the role of LAB in determining the final cheese quality.

Quantitative studies and taste re-engineering experiments toward the decoding of the nonvolatile sensometabolome of Gouda cheese.

The first comprehensive quantitative determination of 49 putative taste-active metabolites and mineral salts in 4- and 44-week-ripened Gouda cheese, respectively, has been performed; the ranking of these compounds in their sensory impact based on dose-over-threshold (DoT) factors, followed by the confirmation of their sensory relevance by taste reconstruction and omission experiments enabled the decoding of the nonvolatile sensometabolome of Gouda cheese. The bitterness of the cheese matured for 44 weeks was found to be induced by CaCl2 and MgCl2, as well as various bitter-tasting free amino acids, whereas bitter peptides were found to influence more the bitterness quality rather than the bitter intensity of the cheese. The DoT factors determined for the individual bitter peptides gave strong evidence that their sensory contribution is mainly due to the decapeptide YPFPGPIHNS and the nonapeptides YPFPGPIPN and YPFPGPIHN, assigned to the casein sequences beta-CN(60-69) and beta-CN(60-68), respectively, as well as the tetrapeptide LPQE released from alphas1-CN(11-14). Lactic acid and hydrogen phosphate were identified to play the key role for the sourness of Gouda cheese, whereas umami taste was found to be due to monosodium L-glutamate and sodium lactate. Moreover, saltiness was induced by sodium chloride and sodium phosphate and was demonstrated to be significantly enhanced by L-arginine.

Levels of benzo[a]pyrene and benzo[a]anthracene in smoked "Provola" cheese from Calabria (Italy).

Levels of benzo[a]pyrene (BaP) and benzo[a]anthracene (BaA) in "Provola" cheese samples from Calabria, smoked naturally and using commercial smoke-flavouring, have been determined. A comparative study of BaP and BaA concentrations was carried out on rind, exterior zone, core and slice of these samples. Quantitative determination of BaP and BaA was ascertained by HPLC using fluorescence detection. Levels of BaP and BaA were found in all smoked "Provola" cheese samples, but higher in cheese smoked with natural methods (0.46-1.13 µg kg⁻¹ for BaP and 1.38-9.29 µg kg⁻¹ for BaA) compared to samples smoked using commercial flavouring (0.085-0.32 µg kg⁻¹ for BaP and 1.20-2.98 µg kg⁻¹ for BaA). The smoking process contributes to polycyclic aromatic hydrocarbons contamination; however, for improved food safety, the use of commercial buffered smoke is preferable to traditional smoking procedures.

Sensory and mass spectrometric analysis of the peptidic fraction lower than one thousand daltons in Manchego cheese.

A total of 107 different peptides, all derived from alphaS1-, alphaS2-, and beta-casein, were identified in different fractions of artisan or industrial Manchego cheese at 4 and 8 mo of ripening, and their sequences were examined. Most of these peptides are described for the first time in Manchego cheese. Taste characteristics (umami and bitter) were assigned based on their AA sequence and the position of these AA within the sequence. The umami taste was predominant in all fractions analyzed by the panelists, and the peptides EQEEL, QEEL, and EINEL, containing a high number of glutamic residues, were found within the fractions. However, in several fractions described as having umami characteristics, no peptides responsible for this taste were detected. Therefore, compounds other than peptides seem to be involved in the umami properties of water-soluble extracts lower than 1,000 Da of Manchego cheese.

Levels of cystathionine gamma lyase production by Geotrichum candidum in synthetic media and correlation with the presence of sulphur flavours in cheese.

Geotrichum candidum is a cheese-ripening agent with the potential to produce sulphur flavour compounds in soft cheeses. We aimed to develop an alternative test for predicting the aromatic (sulphur flavours) potential of G. candidum strains in soft cheese. Twelve strains of G. candidum with different levels of demethiolase activity (determined by a chemical method) in YEL-met (yeast extract, lactate methionine) medium were studied. We investigated cgl (cystathionine gamma lyase) gene expression after culture in three media - YEL-met, casamino acid and curd media - and then carried out sensory analysis on a Camembert cheese matrix. We found no correlation between demethiolase activity in vitro and cgl gene expression. Sensory analysis (detection of sulphur flavours) identified different aromatic profiles linked to cgl expression, but not to demethiolase activity. The RT-PCR technique described here is potentially useful for predicting the tendency of a given strain of G. candidum to develop sulphur flavours in cheese matrix. This is the first demonstration that an in vitro molecular approach could be used as a predictive test for evaluating the potential of G. candidum strains to generate sulphur compounds in situ (Camembert cheese matrix).