Measuring all nutrients present in foods cannot be performed by using a single method of chemical analysis. For instance, it is impossible to extract both highly polar water-soluble compounds like phytate and highly apolar lipid-soluble compounds like carotenoids and vitamin E in a single extraction solvent. Similarly, minerals and phytonutrients cannot be analysed in the same manner. Thus, dedicated methods are needed for both extraction and detection of various nutrients.

Measuring nutrients not only involves a good analytical practice, but also a well-defined aim of research, well-designed experimental set-up and correct handling of materials to be analysed. Only a good framework from the start of plant growth until the final nutrient analysis will enable getting reliable conclusions on, as examples, the effects of cultivar, plant treatment or growth conditions. Sufficient number of independent biological replica’s needs to be included in the analyses to enable answering the specific research question. For instance, when comparing different cultivars for their nutrient levels at a specific growth condition, or when different growth conditions are compared within a single cultivar, care should be taken to!) grow sufficient plants (biological repetitions) to allow good statistics, and II) each plant should be representatively sampled by taking a representative leaf or fruit sample. Representative sampling can be performed by pooling per plant a few leaves or fruits, at the same stage of development, rather than a single leaf or fruit, respectively. Moreover, the analytical approach applied may depend on the exact research question: for instance, if the aim is to know if a specific growth condition affects the level of a certain nutrient in your plants, generating relative levels may be sufficient to compare conditions. On the other hand, if quantitative levels are required, the method applied should be well checked and validated, as each plant sample or food material is different and this matrix effect can significantly influence the quantification. Nevertheless, even if nutrients are quantified, a comparison of nutrient levels detected in your plant or food material with those reported in literature is hard and frequently unreliable, also within the same crop, as the actual growth conditions and developmental stages are inescapably different. For plant mineral nutrition examples are provided in Chapter 5 (Figure 5.x). For instance, under hydroponic conditions the same wheat cultivar can have 130 mg/g of Zn and 220 mg/g of Fe in the grain versus a grain content of 31 and 33 mg/g of Zn and Fe, respectively, under field conditions (Welch et al. 2005).

In addition, nutrient levels obtained depend on the specific methods used for sample handling, extraction and the analysis itself. The nutrient of interest may be present in different forms: e.g. vitamin C can be present in both its reduced form (ascorbate) and in its oxidized form (dehydro-ascorbate, or DHA; see §3.2.4) and the ratio of these forms may differ between food materials; folic acid is not present in plants as such, but principally conjugated with a varying number of glutamate units attached to its pteroyl structure; flavonoids and other phenolic structures are mainly present as conjugated forms, e.g. decorated with different sugar groups and stored as so-called glycosides (§3.2.2). Moreover, quantified levels are highly dependent on the actual purity of the standards used for making the calibration curves: the purity of a standard compound provided by the chemical supplier is not always met and decreases after opening of the bottle. Such purity loss of standards may simply result in a marked overestimation (factor 2 or even more) of nutrient levels, especially of labile nutrients such as carotenoids, tocopherols and ascorbic acid. As a result, the levels in food tables provided for phytonutrients may markedly differ depending on the exact protocols used for their analysis. Therefore, the best is to compare plant cultivars or cultivation practices for effects on nutrient levels directly, by growing the plants within the same experiment and using the same protocols for harvesting and methods and standards for the analyses. Alternatively, appropriate quality control samples should be included in the analysis to enable comparison of nutrient data from different experiments, like different harvest times.


Welch RM, House WA, Ortiz-Monasterio I, Cheng Z. 2005. Potential for Improving Bioavailable Zinc in Wheat Grain ( Triticum Species) through Plant Breeding. Journal of Agricultural and Food Chemistry 53: 2176–2180. DOI: 10.1021/jf040238x.