**2. Materials and Methods**

#### *2.1. Working Group Organization and Training*

A working team composed of two coordinators and a committee (including researchers, computer scientists and compilers, all of whom were dietitians and nutritionists) was established for the preparation of the S4H FCDB. Both the coordinators and the compilers completed the e-learning course offered free of charge by FAO/INFOODS [57]. The e-FoodComp course on food composition was designed by experts to be used by different professional users. The course consisted of 14 lessons structured in five units, for a total of approximately 10 h. The course offers a large number of examples and exercises suitable for on-the-job training. In addition, different guides and research were chosen to be used as a reference for the standardization and harmonization processes [3,25,32,33,41–51,57–68]. The coordinators established the general guidelines, and also helped choosing and obtaining the FCDBs used. In addition, they were subsequently responsible for checking and assessing the quality of the harmonized procedures and data. The remaining committee members performed the rest of the tasks.

## *2.2. Data Collection, Harmonization and Standardization Methods*

A personalized nutrition intervention for different populations in Spain, Germany and Greece will be carried out within the S4H Project [56]. For this reason, the three national FCT/FCDB of the intervention countries were used as references [69–71]. These FCT/FCDB were completed with values of nutrients, bioactive compounds, such as polyphenols, and foods from different databases [14,26,27,30,31,72–75] (Table 1). All FCT/FCDB were either free of charge or permissions were granted when needed. The original FCDB data, such as original name or food identifier, were kept for the purpose of future checks or updates. In addition, quality and traceability of the documented data was guaranteed. However, the data needed to undergo some conversions before being added to our FCDB. All data were harmonized in order to obtain standardized foods and nutrients. Subsequently, all the information was entered into dynamic spread-sheets that related the data and characteristics to each other. As all foods were not in one single language, names and recipes were translated into English. All foods were uniquely identified using the standardized food classification and description system proposed by EFSA FoodEx2 [40,42]. The coding was carried out by qualified compilers and the last version of FoodEx2 system was used [40]. FoodEx2 allowed coding of all foods and beverages present in the FCDB into 20 main food categories, divided into subgroups up to a maximum of four levels [68]. Fortified foods, dietary supplements, food commercial brands, recipes or prepared dishes were discarded from the FCDB. Cooked foods were included, and the cooking method was extracted as an additional data element. Generic unbranded processed foods (such as canned foods, pickles, processed meats or pastries, among others) were also included.


**Table 1.** Compilation of FCDBs used in the construction of the S4H FCDB.

<sup>1</sup> Number of foods used out of total. <sup>2</sup> Number of foods included. <sup>3</sup> Number of items collected, including information on food nutrients and other compounds and data.

> The complete dataset was examined and converted into standard units [3,43]. The tagnames for food components developed by INFOODS were used for this purpose [33,60]. In order to ensure harmonization, standard tagnames were designed for each compound. The original FCDB compounds that were in different units or did not correspond to those described in the INFOODS tagnames, were transformed and recalculated to match the one expressed in the standard tagname (i.e., change of units from grams to milligrams) [33,43].

Only in specific cases were tagnames not modified (as in the case of some polyphenols) where the coordinators decided that it was more functional to leave all compounds with the same units. Those compounds that did not have labels were assigned one that was proposed by compilers. The labels and units can be found in Supplementary Material S1 (Excel sheet). All compounds were expressed in amount per 100 g or 100 mL of food and edible portion values were extracted for further calculations as recommended [3,33]. All changes were made manually or semi-automatically in spreadsheets. All changes were monitored and subsequently validated as described in Section 2.5.
