Profile of Polyphenols, Fatty Acids, and Terpenes in Henola Hemp Seeds Depending on the Method of Fertilization

2.3.2. Triterpenes and Tetraterpenes

During the conducted research, within a large group of terpenes, the presence of triterpenes, i.e., phytosterols, and tetraterpenes, i.e., carotenoids, was found (Table 3). In the literature on the subject, it is the commonly known sterol β-sitosterol [37]. Based on the results presented in Table 3, four phytosterols were detected among the triterpenes: campesterol, stigmasterol, β-sitosterol, and del-ta-5-avenasterol. It was found that the dominant phytosterol in Henola 2 seeds is β-sitosterol (NPK + microconditions = 179 mg/100 g). However, in previous studies, Frankowski et al. (2023) [10] found an approximately five lower content of these sterols. Based on the conducted research, the presence of tetraterpenes and three carotenoids, such as lutein, zeaxanthin, and beta-carotene, was also found (Table 3). It was found that the dominant carotenoid in Henola seeds (1 NPK + microconditions = 63.48 mg/kg) is beta-carotene. Similar observations were made by Frankowski et al. (2023) during his research (Table 5) [10].

Figure 2 visualizes the correlation matrices of phenolic acids in seeds under different soil conditions. There are both high correlations and relatively low correlations. Statistically insignificant correlations are plotted with a cross. On the other hand, among the flavonoids considered in the work, there is only a significant negative correlation between the content of naringenin and rutin under NPK conditions. Apart from this relationship, there are no significant correlations.

Under control conditions, only a significant positive correlation can be observed for p-coumaric and ferulic acids in Henola 1 plants. In contrast, a significant negative correlation is seen in Henola 2 only for catechins and syringing acids (Figure 2a,b).

In Henola 1 seeds in the conditions of enriched NPK soil, regardless of the content of microelements, the increase in the content of p-hydroxybenzoic acid has a positive effect on the content of p-coumaric. A similar tendency can be observed in the case of catechins and syringing acids (Figure 2c,e).

With NPK fertilization of the Henola 1 plant, the p-hydroxybenzoic acid content is significantly positively correlated with the content of catechins and syringing acids (Figure 2c). A similar trend can be observed in the case of vanillic and coffee, as well as sinapic and ferulic acids.

In Henola 2 seeds grown under NPK conditions, only a significant positive correlation was observed for benzoic and ferulic acids, Figure 2d).

In the conditions of fertilization with NPK+micro for Henola 1 seeds, the relationship between the content of vanillic acid and p-coumaric, ferulic, and p-hydroxybenzoic acid deserves attention (Figure 2e). With the increase in vanillic content, the content of ferulic, p-coumaric, and p-hydroxybenzoic acids increases, which is also correlated with the content of sinapic acid.

Only in plants growing in the best environmental conditions (i.e., NPK+micro), it was observed in Henola 1 seeds that with the increase in the content of p-coumaric acid, the content of ferulic acid increases, while in Henola 2 seeds, with the increase in the content of one of these acids, the content of the other decreases (Figure 2e,f).

Figure 3 presents a heat map with a dendrogram for the content of nine phenolic acids and two groups of seeds (Henola 1 and Henola 2) in three different soil conditions: Control, NPK, and NPK+micro. Analyzing all environmental conditions, due to the content of all phenolic acids, Henola 1 and Henola 2 do not form separate clusters. Under the conditions of NPK fertilization with possible micro enrichment, the content of ferulic, p-coumaric, vanillic, and sinapic acids is similar and significantly different from the content of syringing and benzoic acids (Figure 3b,c). Differences can be noted for catechins acid, whose content is similar to syringing in NPK conditions and to benzoic acid in NPK+micro conditions. In the conditions of fertilization with NPK+micro, the content of coffee and p-hydroxybenzoic acids is similar to the content of syringing, while only in the conditions of NPK, the content of coffee is comparable to the content of benzoic acid (Figure 3b,c). In the control conditions (Figure 3a), vanillic, syringing, and benzoic acids form the first cluster; the second cluster is coffee, p-hydroxybenzoic, p-coumaric, sinapic, ferulic, and catechins. Under the conditions of NPK fertilization (Figure 3b), catechins and syringing acids form the first cluster; the second includes coffee and benzoic acids; and the third is ferulic, p-coumaric, vanillic, sinapic, catechins, and benzoic. On the other hand, in plants growing in the best environmental conditions, NPK+micro (Figure 3c), p-hydroxybenzoic, coffee, and syringing acids include the first cluster, and the second one includes ferulic, p-coumaric, vanillic, and sinapic acids, and the third cluster contains catechins and benzoic acids.

Principal component analysis was carried out for the content of phenolic acids in hemp seeds grown under three different fertilization conditions in order to determine the similarities and differences in the quality of seeds from Henola 1 and Henola 2. Figure 4 presents the loads for the first two principal components for seeds grown in three different conditions, respectively. The loadings show us the contribution of individual base variables, i.e., phenolic acids, to the formation of principal components.

In turn, Figure 5a–c present individual observations and features (the content of phenolic acids) on the plane in the system of the first two main components in different fertilization conditions: Control, NPK, and NPK+micro. Under the control conditions (Figure 5a), the first two principal components explain over 84 of the total variability of the nine examined features. An analogous situation is visible when fertilizing with NPK+micro (Figure 5c). Here, too, the loss of information regarding the variability between the contents of phenolic acids at the transition from the nine-dimensional space of original variables to the space of the first two principal components is about 16. In Figure 5b, under the conditions of NPK fertilization, the two main components, which are a linear combination of the input hemp characteristics, explain over 82 of the total variability, i.e., the loss of information due to dimension reduction is about 18. Under the control conditions (Figure 4), the features most strongly correlated with the first principal component are the following acids: catechins, ferulic, syringing, and p-hydroxybenzoic. The most significant influences on the second main component are the acids: benzoic, sinapic, and coffee. When fertilizing with NPK, the features most strongly correlated with the first main component are the acids: p-hydroxybenzoic, ferulic, p-coumaric, and sinapic, while with the second main component are the acids: benzoic, syringing, and catechins.

Under NPK+micro conditions, catechins, ferulic, and vanillic acids significantly (positively) affect the value of the first component. The value of the second component is most positively influenced by benzoic acid, while the content of coffee, syringing, and p-hydroxybenzoic acids has a negative impact.

It can be observed that in all three conditions, the vast majority of acids have a positive effect on the first principal component. And on the second component, the greatest influences are negative (Figure 4 and Figure 5). It is also seen that catechins and ferulic acids have a large positive effect on the first principal component under all conditions.

In turn, syringing acid has a large negative effect on the first principal component under control conditions, but it also has a large negative effect on the second principal component already during fertilization. P-hydroxybenzoic acid has a large positive effect on the first component, but only under control and NPK conditions. On the other hand, benzoic acid clearly affects the form of the second main component in all three conditions of fertilization. Except that in control conditions and NPK, it has a negative effect, and in NPK+micro conditions it has a positive effect.

In Figure 5, no clear grouping of Henola 1 and Henola 2 can be observed.

Comparing Figure 5a–c, it can be concluded that the dispersion of individual observations is at a similar level in all fertilization conditions. The type of fertilization does not change the content of phenolic acids.

The similarities and differences in the content of phenolic acids in hemp seeds under different fertilization conditions were obtained by performing a principal component analysis for the content of the considered compounds in hemp seeds from Henola 1 and Henola 2.

The contribution to individual components, and thus the impact on the variability of the total acid content, is shown in Figure 6.

The relationships between the contents of phenolic acids expressed in terms of the first two principal components for Henola 1 and Henola 2 are shown in Figure 7.

In the experiment performed for Henola 1 seeds (Figure 7a), the first two principal components explain eighty-four of the total variability of the nine tested phenolic acids. In Figure 7b, in the experiment performed for Henola 2 seeds, two principal components, being a linear combination of the input hemp characteristics, explain slightly more than 67 of the total variability. So in this case, the information loss due to dimension reduction to two variables is higher than for Henola 1 seeds.

For Henola 1 (see Figure 7a), the features most strongly correlated with the first principal component are the following acids: sinapic, p-coumaric, ferulic, and p-hydroxybenzoic. Syringing and benzoic acids have the greatest influence on the second principal component. On the other hand, for Henola 2, catechins, benzoic, ferulic, coffee, and p-hydroxybenzoic acids significantly (positively) affect the value of the first component. The value of the second component is most influenced by syringing, sinapic, and p-coumaric.

It can be seen that in Henola 1 all acids have a negative effect on the first principal component, and in Henola 2 most acids have a positive effect on the first principal component (see Figure 6 and Figure 7).

In Figure 7, no clear grouping of fertilizations can be observed, respectively: Control, NPK, and NPK+micro. The type of fertilization does not change the content of phenolic acids.

Comparing Figure 7a,b, it can be concluded that the dispersion of individual observations is at a similar level for the content of acids in the seeds of Henola 1 and Henola 2. The method of cultivation does not change the content of phenolic acids.

Figure 7a shows that for Henola 1, there is a positive correlation between the variables, or it is at the level of zero. There are no negative correlations between phenolic acids. On the other hand, there are positive, negative, and zero relationships between the acid contents in Henola 2 seeds.

Principal component analysis was carried out for the content of flavonoids in hemp seeds grown under three different fertilization conditions in order to determine the similarities and differences in the quality of seeds from Henola 1 and Henola 2.

Figure 8 presents the loads for the first two principal components for seeds grown in three different conditions, respectively. The loads show us the contribution of individual base variables, i.e., flavonoids, to creating principal components.

In order to graphically present the variability of flavonoid content, principal component analysis was used. Figure 9a–c show individual observations and features (the content of flavonoids) on the plane in the system of the first two principal components under different fertilization conditions: Control, NPK, and NPK+micro.

Under the control conditions (Figure 9a), the first two principal components explain over 75 of the total variability of the three examined features. In Figure 9b, under the conditions of NPK fertilization, the two main components, which are a linear combination of the input hemp characteristics, explain over 95 of the total variability, i.e., the loss of information due to dimension reduction is about 5. An analogous situation is visible when fertilizing with NPK+micro (see Figure 9c). Here, too, the loss of information is very small at the transition from the three-dimensional space of original variables to the space of the first two principal components and amounts to about 3.

Under the control conditions (see Figure 8 and Figure 9), the features most strongly correlated with the first principal component are rutin and quercetin. Naringenin has the greatest influence on the second principal component. When fertilizing with NPK, the features most strongly correlated with the first principal component are naringenin and rutin. However, the second main component is quercetin.

Under NPK+micro conditions, the flavonoids naringenin and quercetin significantly affect the value of the first component. Rutin has the greatest influence on the value of the second component.

It can be observed that in the control conditions, the content of flavonoids has a positive effect on the first principal component, and in the NPK+micro conditions the opposite is true (see Figure 8).

In Figure 9, no clear grouping of Henola 1 and Henola 2 can be observed.

Comparing Figure 9a–c, it can be concluded that the dispersion of individual observations is at a similar level in all fertilization conditions. The type of fertilization does not change the content of flavonoids.

Similarities and differences in flavonoid content in hemp seeds grown in different fertilization conditions were obtained by performing a principal component analysis for the content of the considered compounds in hemp seeds from Henola 1 and Henola 2.

The contribution to individual components, and thus the impact on the variability of the overall flavonoid content, is shown in Figure 10. The relationships between the contents of flavonoids expressed in terms of the first two main components for Henola 1 and Henola 2 seeds are shown in Figure 11.

In the experiment performed for Henola 1 seeds (Figure 11a), the first two principal components explain over 89 of the total variability of the three flavonoids tested. In Figure 11b, in the experiment performed for Henola 2 seeds, two principal components, which are a linear combination of the input cannabis characteristics, explain slightly more than 88 of the total variability. The loss of information due to the reduction of the dimension to two variables is at a similar level for Henola 1 and Henola 2 seeds.

For Henola 1 seeds (see Figure 11a), the features most strongly correlated with the first principal component are naringenin and rutin. Quercetin and rutin have the greatest influence on the second principal component. On the other hand, for Henola 2, all three tested flavonoids significantly (negatively) affect the value of the first component. The value of the second component is most influenced by rutin and quercetin.

It can be seen that in Henola 1 and Henola 2 seeds, the content of naringenin has the greatest impact on the first principal component, but in Henola 1 it has a positive effect and in Henola 2 it has a negative effect. In turn, rutin and quercetin have the greatest influence on the second principal component, but these influences are opposite in Henola 1 and Henola 2 (see Figure 10 and Figure 11).

In Figure 11, no clear grouping of fertilizations can be observed, respectively: Control, NPK, and NPK+micro. The type of fertilization does not change the content of flavonoids.

Comparing Figure 11a,b, it can be concluded that the dispersion of individual observations is at a similar level for the content of flavonoids in the seeds of Henola 1 and Henola 2. The method of cultivation does not contribute to the change in the content of flavonoids.

The non-parametric Wilcoxon test was used to evaluate the differences between Henola 1 and Henola 2 seeds in the degree of phenolic acids and flavonoids content. The test was used with the assumption of significance at p < 0.05. Calculations were performed separately for each considered phenolic acid and flavonoids. Based on the obtained results (Table 6), it can be concluded that there are no differences between the distributions for Henola 1 and 2 seeds at the assumed significance level.

3.2.3. Statistical Methods

As part of the research, the seeds of the Cannabis sativa L. Henola variety were analyzed. The experiment concerned a plant grown only for seeds (Henola 1), cultivated to obtain essential oils, and then for seeds as seed (Henola 2). The research was carried out on two variants (depending on the intended use), each cultivated in three different series (depending on fertilization). Each of these series was fertilized accordingly: in the first case no fertilization was performed (Control), in the second case only NPK (NPK) was fertilized, and in the third case NPK and micro (NPK + micro). Each sample was analyzed in three replicates. Statistical analysis includes the determination of correlation coefficients between phenolic acids and between flavonoids in seeds. They define the relationships between individual variables. First of all, the dependencies that are statistically significant were indicated. The analysis also uses cluster analysis methods that allow for grouping the examined objects according to their similarity hierarchy. The results are presented in the form of a heat map with dendrograms. In turn, in order to reduce the number of variables and detect regularities between the examined features, a popular technique was used—principal component analysis (PCA). There are large differences in the scale of the data; therefore, before starting this analysis, data scaling [13] was used, consisting of calculating the quotient of the difference of a single measurement and the mean by the standard deviation in the sample. Thanks to it, it was possible to replace the original variables with new variables (components), mutually independent, which successively explain a specific part of the variance of the initial variables. On the basis of the scree chart, such a number of components was extracted so that the percentage of explained variance was sufficient. In the system of the first two components, the vectors of the examined hemp traits are presented. The longer the vector representing the given features, the greater the share of this variable in the values of the first two components. The direction of the arrows shows the influence of these features on the first and second components, respectively. The angle of intersection of the vectors is proportional to the relationship between the features. The points, on the other hand, represent the individual observations in the new layout. This analysis allowed us to determine the effect of fertilization on the chemical composition of seeds and to compare the properties of Henola 1 and Henola 2 seeds. Finally, a non-parametric Wilcoxon test was performed to check whether Henola seeds 1 and 2 came from the same distribution. All statistical analyses were performed using R 4.3.0 software [R Core Team, 2023]. The drawings in this document were created using the R ggplot2 (version 3.4.2), ggcorrplot (version 0.1.4), ComplexHeatmap (version 2.16.0), corrplot (version 0.92) packages.