Carmen Diez Simon

Aroma is considered as the superior component of flavour in food. The molecular compounds responsible for aroma are the so-called volatiles. Volatiles are small molecules that exhibit high vapour pressure under ambient conditions and have low boiling points, thus, they are present in the air phase as well as in the liquid and solid phases. Many volatiles are naturally present in plants, animals, microorganisms and the environment, however they can also be formed by the action of (enzymatic or non-enzymatic) processing reactions. When food is processed or cooked, different volatiles can be formed and these may influence the final flavour of food. Various analytical techniques (including metabolomics approaches) are being used to analyse the composition of food samples in order to generate new knowledge of the volatile composition of food samples, which is of great relevance when designing and formulating new superior flavourful products for the market.

In this PhD thesis I explored several methodologies to extract volatiles from complex food matrices, such as in processed food products (e.g. yeast derivative ingredients and instant soups). The major aim of this thesis was to design novel approaches to be able to zoom in on the volatile content of savoury food ingredients to better understand which ingredients and processing strategies can lead to superior and desired flavour qualities.

Therefore, first I focused on performing two literature studies, which covered the background knowledge of the savoury ingredients used in this thesis: yeast derivative ingredients in Chapter 2 and soy sauces in Chapter 3. Chapter 2 provides a comprehensive and up-to-date review of the major flavour compound classes described in processed savoury ingredients, including the formation of these compounds from their precursors. Special attention was given to the interconnections between Maillard reactions and the different amino acid, lipid, and carbohydrate degradation pathways. Furthermore, the chapter provides insights into advancing metabolomics applications that have not yet being exploited in depth for processed food ingredients.

As a result of the learning outcome from the first review, I proceeded with the reported literature on soy sauce composition. Chapter 3 summarises our knowledge on the chemical compounds known to be present in the most common soy sauces found on the market, and their potential sensory relevance. This review also presents a sensory wheel of taste and aroma attributes that characterize soy sauce flavour. Soy sauce is a condiment used worldwide and is perhaps one of the most complex fermented condiments containing many small molecules that boost the flavour of many dishes. Here also little has yet been done in implementing comprehensive metabolomics analyses to study the metabolite composition of soy sauces.

Intrigued by the knowledge of the volatiles reported in the literature and the lack of untargeted metabolomics techniques used for these specific food ingredients, I focused Chapter 4 on developing and comparing new methodologies that broaden the volatile spectrum as well as taking into consideration the ‘unknown’ compounds, which are also of great relevance in flavour studies. Four volatile trapping techniques were optimized and compared (SBSE, SPME, HSSE and DHS), and the results showed that SBSE performed better in terms of coverage of volatiles, while SPME was the most repeatable technique of all.

The lack of implementation of new methodologies, such as SBSE techniques, in soy sauce analysis led to Chapter 5, which focuses on analysing the volatile content of different soy sauce products made using contrasting production procedures and flavour characteristics. Fermented and non-fermented soy sauces showed the greatest differences in terms of the volatile content, and specific classes of volatiles (pyrazines in non-fermented and esters in fermented soy sauces) could be linked to the different types of soy sauces analysed. These results lead to the proposition that volatile content, and thus aroma, are greatly influenced both by the process used to produce the final product as well as the ratio of raw ingredients added. This resulted in a list of potential biomarkers for specific soy sauce types found on the global market. The implementation of new methodologies sets the basis for future research in the application of metabolomics for soy sauce products.

As a result of observing contrasting volatile profiles when using different volatile trapping techniques, I focused Chapter 6 on finding applicable tools/pipelines that help in deciding, in the early stages of large-scale experiments, which technique is best to use in a specific study. The workflow proposed showed that SPME is more repeatable than SBSE, however, SBSE was able to trap more volatiles than SPME and also that the additional (unique) volatiles that SBSE trapped increased the power of prediction for certain sensory attributes, such as for garlic flavour, as compared to SPME. Therefore, depending on what is the precise goal of a study, an informed decision can be taken on which technique is best to apply.

In order to confirm that volatiles measured by SBSE-GC-MS can be used to predict sensory characteristics in savoury products, results in Chapter 7 show that sensory attributes typical of bouillon-type soups, such as roast odour or chicken flavour, could be associated with certain volatile classes (i.e. short-chain aldehydes with chicken flavour), and thus the ingredients and processing steps of the bouillon samples that contained those volatile compounds could be linked.

All these chapters combined have shown that this untargeted metabolomics, data-driven approach is valuable for 1) scientists to understand aroma formation and chemical changes which occur during processing and; 2) industry to formulate improved food products in a more targeted way.