E-nose and E-tongue for sensory quality evaluation of canestrato cheese produced with kiwifruit extract

The development of texture, aroma, and flavor of most cheese types is directly affected by casein proteolysis, which is one of the major biochemical changes occurring during cheesemaking and ripening [1]. The most widely used proteases in cheesemaking is the animal rennet, which is extracted from the abomasum of the newborn ruminants. Recently, plant derived enzymes are proposed as valid substitute for the milk clotting and among them the kiwifruit extract appears very promising [2]. Up to now, little data are available on the effects of kiwifruit extract on the organoleptic characteristics of the cheese. The aim of this work was to assess odor and taste properties of cheese made with kiwifruit extract as milk coagulant, using a sensor-based instruments platform. The study was conducted on two set of 6 samples each of fresh cheeses (15 days of seasoning), obtained from a blend of pasteurized cow and ewe’s milk, which was coagulated using kiwi extract (H cheese). Cheese samples made with commercial lamb rennet was used as control (C). The cheese samples were analyzed using an E-nose (FOX 4000, Alpha M.O.S., Toulouse, France) with 18 MOS sensors and a potentiometric E-tongue (αAstree, Alpha M.O.S., Toulouse, France) with 7 chemical sensors. For each sample, 10 replicates were analyzed using both instruments. Exploratory data analysis was performed with principal component analysis (PCA) and expressed based upon the discrimination index (DI). The E-tongue PCA plot shows that the C group was well discriminated from H cheese with a discrimination index of 92%. The first two components (PC1 and PC2), represent 99.70% of the total variance between the sample measurement. The sensors named AHS and SCS showed a greater intensity of response towards the H cheese. The two sensors respond to acid and bitter taste respectively [3] highlighting a strong difference between the two groups of cheeses, in accordance with other studies [4]. The E-nose PCA plot shows a discrimination index of 91%. Moreover, PC1 and PC2, explain a variance of 99.94%, showing also a clear separation between the two groups considered. Specifically, the LY-type sensors, sensitive to short-chain volatile fatty acids and aldehydes, moving towards the C group, while the H group is defined by the P-type sensors, sensitive to methane, propane and other aliphatic non-polar molecules and the T-type sensors, sensitive to polar alcoholic and chlorinated compounds [5]. The artificial senses showed the effect of bitter compounds in the final products underlying the limitations of plant coagulants in cheese production; these results, although preliminary, can be very useful for developing cheese making strategies aimed at reducing the bitterness in the final products. Certainly, at the end of the trial, a panel test combined with E-nose and E-tongue will be carried out to quantify the effect on the consumer.