Prof Dolores T. Reece
Food Chemistry Room 1A
Every day, numerous analyses are carried out on food products to answer a host of problems, in modern and natural food science.
These analyses are ultimately aimed at the improvement of food quality, reduction in price, increase in production yield, the elimination of undesirable effects during preservation, and determining the correlation between the food and the frequency of some diseases. In addition, food scientists are requested to assess some kind of guarantee of food quality, with regard to its content, geographic origin and age.
Many parameters determine the composition of food: class, order, family, species, variety of breed of the food producers (plant or animal), age, soil composition or nourishment, climate, preservation and treatments of the food, etc.; these parameters constitute the “cause space”.
On the other hand, food composition determines its quality, preservation characteristics, price, food value, appearance, taste, and odour; these parameters constitute the “space of effects”.
The chemical composition of samples we collect is the “chemical space” which represents the intermediate between cause and effect space, and which is necessary for the interpretation of cause-effect relationships.
The chemical space is described qualitatively and quantitatively and its description can be rough or very detailed. We can obtain the determination of all components, at the trace level as well. Real descriptions in practice, made by measuring a few components, in some cases belonging to a particular chemical class, or by”measuring physical quantities related to composition, such as spectral variables.
Cause space, as well as chemical space, can be described more or less in detail: e.g., climate description can range from a brief description (warm, cold) to daily temperature, moisture, lighting intensity at every incident frequency, etc., for all samples. Currently, the cause space is described roughly (mean temperature, etc.) or with a little more detail (seasonal or monthly temperature, precipitation, thermal swing, etc.).
Effect space is often the hardest to describe. Only recently were nutrients systematically connected with longevity or with the frequency of some diseases. Food quality is rarely broken down into its components (taste, smell, appearance, etc.), which are very difficult to measure. However, also in this space,
feature description is tending to become quantitative (panel scores, quality index).
Improved descriptions of these three spaces are necessary to determine cause-effect relationships. Today, we can obtain very efficient chemical descriptions and possibly also good descriptions of some features of the cause and effect spaces.
Because of the consequent high number of descriptors and of samples, powerful means are necessary to find useful relationships in this great quantity of data.
These tools are chemometric methods by which one may find relationships between cause space and chemical features, and between these and affect space, explaining them in the light of chemical composition.
However, one cannot think of obtaining more and more detailed descriptions for a greater and greater number of samples, because of economics and time considerations. So, chemometrician must study food problems to evaluate the performance of reduced sets of chemical components.
Moreover, in spite of the great improvement in analytical instrumentation, and the spreading of the principles of multivariate analysis in chemistry, a great many food analyses are carried out by univariate criteria, that is, seeking the parameters that are correlated to particular effects on the basis of experience. But this experience is old. For example, in the study of many foods such as fats and beverages, the ratio between absorption at two wavelengths is used; the whole information contained by remaining part of the spectrum is ignored because the old experience could only recognize the most evident aspect of the information.