*2.5. Statistical Analysis*

The data processing was performed using statistical software Minitab 17.1.0. One-way ANOVA analysis of variance was performed in order to assess the significant differences of the variable between monofloral and polyfloral honey samples. The confidence level (*p* = 0.05) was used for every ANOVA test. Tukey comparison procedure for assuming equal variances was used for every variable obtained from IRMS and UHPLC-HRMS methods while Fisher comparison was used once for total N assessment in honey proteins. Principal component analysis was performed for data reduction in order to find potential chemical compound biomarkers for floral origins. The correlation matrix was used for analysis. As a pre-step, the software performed standardization of variables, meaning a variable was rescaled to have a mean of zero and a standard deviation of one. Principal component scores and their correlation coefficients are stored in a Supplementary Excel file. The formation of monofloral group clusters or positions in the score plot was used to assess the potential of variable capability as a marker.

### **3. Results and Discussion**

### *3.1. IRMS Analysis of Honey Proteins*

C and N isotope ratio and total weight fraction of monofloral and polyfloral honey samples is presented in Table 1. The ANOVA one-way results show that there was no significant variance, with a confidence level of 95% between monofloral and polyfloral

honey, by using δ13C values (*p* = 0.08). Tukey test simultaneous differences of means for δ13C are described in Supplementary Material (Figure S1).


**Table 1.** IRMS analyses results by floral origins of honey.

The carbon isotope ratio in honey proteins is directly influenced by carbon fixation in plants from which bees are gathering honey. Therefore, carbon isotope analysis is mainly used for C4 plant additive determination in honey. Nevertheless, the δ13C values are often used for floral origins determination [3,22].

After extractions of sugars, the honey proteins showed δ13C values in a range of −25.47‰ to −29.64‰, which is characteristic for C3 plants [23]. Moreover, in comparing polyfloral honey proteins (δ13C = −27.4 ± 0.9‰) with other results, we found that δ13C values are more depleted than of Mediterranean region honey proteins. δ13C values are dependent on the amount of sun exposure to plants and air humidity; therefore, an increase of sunny days and less precipitations increases the δ13C values [24].

The nitrogen isotope ratio for honey proteins reflects the nitrogen content of the soil where plants from which bees have gathered the nectar grow. δ15N > 0.0‰ values indicate that the nitrogen is biologically fixed in soil, and values near 0.0‰ show that the nitrogen is obtained from air. Results show clover honey proteins are enriched with heavy nitrogen isotope, although the plant is considered as gathering nitrogen via *Rhizobium* bacteria [25]. Exceptional honey proteins were extracted from heather honey, indicating depleted nitrogen ratio values and statistically different significance (*p* = 0.001) using ANOVA one-way Tukey tests (see Figure S2). In total, 11 out of 78 honey proteins showed negative δ15N values. These samples of honey were heather monoflorals and polyflorals that had reported the presence of heather (*Calluna vulgaris*) pollen (see the Supplementary Excel file).

Total carbon and nitrogen in honey proteins were found to have no particular statistical difference using the ANOVA test. *p*-values were found for total carbon (*p* = 0.5) and total nitrogen (*p* = 0.06), although total nitrogen *p*-values were close to 0.05, which suggests that results could be capable for further floral origin discrimination investigation. Using the Fisher test, we found that there are differences in heather and buckwheat (increased total nitrogen) honey proteins between willow and rapeseed (decreased total nitrogen) (see Figure S3). Total nitrogen in proteins generally is ≈16% [26]. Obtained nitrogen mass fraction results sugges<sup>t</sup> that after dialysis, pure protein is not obtained, but instead a mixture of protein and other molecularly large compounds that could not be separated via dialysis such as lipids and pollen [27,28].
