Color
The CIELAB system, established by the International Commission on Illumination, was employed to analyze the color of different rose accessions as indicated in Table 1. The system utilizes L* to quantify the lightness of the color, ranging from white to black. Additionally, a* and b* indicate distinct color directions, with a* ranging from green to red and b* ranging from blue to yellow. Lastly, c* is used to measure the chroma of the color. Table 1 shows that the R4 accession had the lightest L* value (79.16), which was similar to the R10 accession (76.06) for white and yellow flowers, respectively. The R3 accession had a L* value of 63.55, followed by R8 with a value of 61.08. The darkest L* value of 18.99 was observed in the R7 accession, corresponding to blackish red roses. In terms of other color coordinates, the a* value varied from 48.49 to -3.66. The highest value of 48.49 was observed in R6, which represents red color flowers. This value was similar to R2 (46.90, hot pink color) and R1 (44.85, orange color), but significantly different from the other rose accessions. The lowest value of -3.66, indicating white color flowers, was recorded in the R4 accession. In addition, the values of b* for the other color directions varied from positive to negative, ranging from 60.13 to -3.66. This corresponds to the color spectrum from yellow to purple in a flower. The highest b* value, 60.13, was obtained from R10, which showed a statistically significant difference compared to all other rose accessions. Following R10, R9 had a b* value of 26.07, while R5 had the lowest value of -3.67. The rose accession R10 exhibited the maximum color saturation, with a C* value of 60.19, which was statistically distinct from the other rose accessions. Nevertheless, R6 exhibited the second highest value (50.68), which was statistically comparable to R1 (47.77) and R2 (46.91), while R4 had the lowest vividness of color (13.35). However, the decrease in values of a* and the increase in values of b* are associated with the perception of darkness and lightness. Similarly, the highest luminosity (h°) was observed at R10 (87.56), indicating a brightening of rose petals close to a yellow color. This was followed by R9 (47.13), R7 (19.89), R1 (19.19), and R6 (16.86). The last three values were statistically similar to each other but different from the rest. Therefore, the values of the parameters accurately depicted the color patterns of the flower, aligning with the visual observations of the rose accessions' blossom color.
pH and TSS (°Brix)
A notable fluctuation in pH and total soluble solids (TSS) (°Brix) among the 10 rose accessions was demonstrated in Fig. 1. The acidity of rose petals varied among different accessions, with the lowest pH recorded in R1 (4.50), which was comparable to R7 (4.50), R6 (4.57), and R2 (4.70). On the other hand, the highest pH was observed in R4 (5.60), which was similar to R8 (5.50) and R9 (5.40), and statistically similar to R3 and R5 (5.30). Conversely, the highest total soluble solids (TSS) level was found in R1 (9.40 °Brix), while the lowest was seen in R3 (7.10 °Brix).
Total Carotenoids and β-carotene (mg/100g)
There was a significant statistical variation (P < 0.05) in the total carotenoids and β-carotene concentration noticed across ten different rose accessions (Fig. 2). The rose accessions R10 had the highest total carotenoids content at 108 mg/100 g fresh weight (FW). The second highest was R1 at 72 mg/100 g FW, followed by R9 at 67 mg/100 g FW. The lowest carotenoids content was found in R4 at 6 mg/100 g FW. All rose accessions were statistically distinct from each other. From a β-carotene perspective, the highest concentration of β-carotene was found in R10 (47 mg/100 g FW), followed by R9 (42 mg/100 g FW), and R1 (39 mg/100 g FW). These values were statistically distinct from each other. Furthermore, all the remaining accessions exhibited statistically equivalent levels of accumulation, which were below 5 mg. However, the lowest accumulation was recorded in the rose accessions R3, with a value of 1 mg per 100 g. Notably, the rose accessions R5 and R7 did not exhibit any detectable β-carotene content.
Total Anthocyanin and Betacyanin (mg/100 g)
The column graph (Fig. 3) clearly demonstrates that there were statistically significant variations in the overall anthocyanin and betacyanin content among the 10 rose accessions. In terms of total anthocyanin content, the accession R7 accumulated the maximum quantity of anthocyanin (196 mg/ 100g FW), which was comparable to R6 (191 mg/ 100g FW), R2 (189 mg/ 100g FW), and R1 (183 mg/ 100g FW). On the other hand, R4 accumulated the lowest amount (3 mg/ 100g FW). The total anthocyanin content varied between 196 and 3 mg per 100 g FW.
Conversely, the betacyanin content was determined based on the dry weight of the samples. The highest amount of betacyanin was found in the R7, with a concentration of 22.63 mg per 100g of dry weight (DW) which is equivalent to the amount of anthocyanin. However, it also indicates more than double the concentrations than that of the R1 (10.43 mg/ 100g) and R6 (9.91 mg/ 100g) accession while the lowest betacyanin content was observed in R10.
Tocopherol Content (Vit.E) (mg/ 100 g DW)
The tocopherol content in the rose accessions examined ranged from 400.08 to 300.95 mg/100 g DW. The highest concentration was found in the R8 accession (400.08 mg/100 g DW), which was significantly different from the other accessions. However, the second greatest value was achieved from R6 (400.05 mg/ 100 g DW), which was equal to the value acquired from R1 (400.05 mg/ 100 g DW). On the other hand, the lowest value was seen in R10 (300.95 mg/ 100 g DW), which was statistically close to the value in R9 (300.97 mg/ 100 g DW) (Table 2).
The total phenolic content (TPC) (mg GAE/ 100 g, FW)
Table 2 demonstrates that the various rose petal extracts exhibited significant variance in their total phenolic content (TPC). The R6 accession exhibited the maximum concentration of TPC 533.18 mg GAE/ 100 g, FW, whereas the lowest concentration was observed in the R4 accession, with a value of 241.87 mg GAE/ 100 g, FW.
The total flavonoid content (TFC) (mg QE/ 100 g, FW)
Upon examining Table 2, it becomes evident that all the rose accessions exhibited significant variations in the accumulation of flavonoid contents (TFC). The highest level of flavonoid content was observed in R3 (27.77 mg QE/ 100 g, FW), whereas the lowest concentration was reported in R1 (0.76 mg QE/ 100 g, FW).
Total Antioxidant Activity (IC50) (µg/ mL FW)
The antioxidant activity of rose petals was assessed by determining the IC50 value, which represents the concentration of the sample needed to block 50% of DPPH free radicals. Thus, in the DPPH experiment, greater IC50 values indicate lesser antioxidant activity, and vice versa. Table 2 clearly shows that the IC50 values of 10 rose extracts had a substantial impact on their ability to scavenge free radicals (p < 0.05). The R1 rose accession had the most potent antioxidant activity, as evidenced by its lowest IC50 value of 82.60 µg/mL FW. Among the other rose accessions, R5, R6, R7, R8, R9, and R10 exhibited an IC50 value greater than 250 µg/mL FW, indicating their inactivity in free radical scavenging action. Regrettably, the antioxidant activity of two rose accessions, R3 and R4, was not observed in this experiment.
Minerals (g/100g) and Moisture content (%)
The results from Table 3 clearly demonstrate that the accumulation of mineral elements such as sodium (Na), potassium (K), calcium (Ca), iron (Fe), and moisture content varied significantly among the ten rose accessions, with the exception of magnesium (Mg). Moreover, within the composition of these minerals, the concentration of potassium was particularly notable in the rose accessions. The sodium (Na) concentration in the petals of ten different rose accessions varied significantly. The highest accumulation of Na was seen in accession R5 (0.092 g/100g DW), followed by accession R9 (0.085 g/100g). Simultaneously, the lowest recorded value was obtained from R10 (0.062 g/100g), which was exactly the same as R7 (0.065 g/100g). The analysis found that the potassium level varied between 1.408 and 0.984 g/100g where the maximum concentration was seen in R2, while the lowest value was found in R4, which was statistically similar to R6. The average value of calcium content differed across the roses where the R1 and R3 had the highest accumulation of Ca at a concentration of 0.19 g/100g, whereas R6 and R7 had the lowest concentration at 0.14 g/100g. The accumulation of Fe, varied greatly, ranging from 0.090 to 0.017 g/100g. The top accumulator was R4, while R9 was the lowest accumulator.
Meanwhile, the rose petals exhibited a significant variation in moisture content, with the greatest reported in the rose accession R6 (87.31%). This value was statistically equivalent to that of R5 (87.07%) and the water content in R3 was significantly low, measuring at 68.51%.
Antinutrient properties (g/ 100 g DW)
Upon examining Table 4, it is evident that there were notable variations in the levels of alkaloids, phytate, saponin, and tannins among the rose accessions. The alkaloid content of rose petals in this investigation varied from 1.24 to 14.64 g/100 g DW (Table 4). Of the ten rose accessions, R2 had the greatest alkaloid content accumulator (14.64 g/ 100 g DW), followed by R1 (9.52 g/ 100 g DW), and R10 (1.24 g/ 100 g DW) had the lowest. Of these three metrics, the phytate content indicated a very small quantity of present (Table 4). However, R6 differs greatly from the others due to its higher amount (0.63 g/100 g DW). The majority of the saponin (14 g/100 g DW) was found in the rose petals of R6 and R9, with R3 and R7 following closely behind (12 g/100 g DW), and R1 containing the least (4.03 g/100 g DW). The data displayed in Table 4 indicates that there was a notable variation in the tannin content among the ten rose accessions, ranging from 143.55 to 198.05 mg TAE/ 100g DW. The highest tannin conserver in this testing was R7 (198.05 mg TAE/ 100g DW), followed by R6 (180.57 mg TAE/ 100g DW). Conversely, the R2 accession had the least amount of tannin (143.55 mg TAE/ 100g DW), and it was statistically comparable to the R10 and R4 containers (143.97 and 145.47 mg TAE/ 100g DW, respectively).
Antinutrients to molar ratios
The molar ratios of [PHT]: [Ca], [Ca]: [PHT], [PHT]: [Fe], [PHT]: [K], [Mg]: [PHT], [TNN]: [Fe], and [PHT + TNN]: [Fe] were determined based on the analyzed data of antinutrients and minerals of rose accessions and presented in Table 5. Out of the ten rose accessions, all except for R6 exhibited [PHT]: [Ca] ratios below the crucial value of 0.24 and [Ca]: [PHT] ratios above the critical value of 0.6. The rose accession R6, however, had [PHT]: [Ca] and [Ca]: [PHT] ratios of 0.273 and 3.667, respectively. Based on the table 5, it is evident that the rose accessions R6 and R7 exhibited a [PHT]: [Fe] ratio that exceeded the threshold value of 1, indicating a lower bioavailability of iron. On the other hand, R4 had the lowest ratio of 0.019, suggesting the highest absorption of iron in nutrition. Regarding [PHT], the 10 rose accessions exhibited variances in [K] molar ratios, which varied from 0.001 to 0.033. Table 5 shows that the molar ratios of [Mg]: [PHT] in rose accessions varied from 4.396 to 139.819. Among them, R10, R4, R8, R3, and R9 had larger amounts of magnesium, with molar ratios of 139.819, 138.462, 137.104, 54.842, and 39.560, respectively. The molar ratios of [TNN]: [Fe] were adjusted within the range of 0.142 in R4 to 0.871 in R7 rose accessions. The molar ratio of [Fe] for [PHT + TNN] ranged from 0.161 in R4 to 2.884 in R6, with R6 having the greatest ratio.
Qualitative Assessment of bioactive compounds
The phytochemical screening tests largely determined the presence or absence of steroids, coumarins, quinones, anthraquinones, and phlobatannins in the ten rose genotypes. This was done using color reactions, as shown in Table 6. The presence of phytosteroids, coumarin anthraquinones, quinones, and phlobatanin was confirmed through various chemical reactions, including the formation of a brown ring, the development of a yellow-colored solution, the precipitation of red-colored solids, a color change from red to blue, and the precipitation of pink-colored solids, respectively (Fig. 4. A1, A2, A3, A4, A5). The analysis of the ten rose accessions revealed the presence of steroids in R1, R4, and R10 accessions, coumarines in R1, R3, R4, R9, and R10 accessions, quinones in R1, R2, R6, and R7 accessions, anthraquinone in R1, R3, R6, and R7 accessions, and phlobatanin exclusively in the R6 rose accession.
Correlation coefficient analysis
The Pearson correlation coefficient was employed to evaluate the intra and interrelationships among the 25 variables under investigation. The correlation matrix visually represents the degree of both positive and negative association between the colorimetric parameters, secondary metabolites, and mineral contents (Fig. 5A). In the event of a positive correlation, an increase in one variable will result in a corresponding increase in another variable, whereas a negative correlation indicates that an increase in one variable will lead to a decrease in another one. The circles in Fig. 5, colored in blue and red, indicate positive and negative correlations, respectively. The intensity of the color represents the strength of the correlation between the variables. Vacant cells indicate an inconsequential association at a 5% level of significance. The Pearson's correlation coefficient and correlation matrix revealed a significant link between colorimetric features and secondary metabolites, ranging from moderately strong to extremely strong. However, this correlation was not observed with mineral matters. Among these factors, a significant positive correlation (R2 = 0.95) was observed between the total carotenoid and β-carotene content. This suggests that an increase in total carotenoid content is associated with a corresponding increase in the concentration of β-carotene. The total carotenoid concentration exhibited a significant positive correlation with h° and b* (R2 = 0.71, 0.79). Moreover, there was a significant and positive relationship between the β-carotene content and both the b* value and hue angle, with R2 values of 0.84 and 0.79, respectively. Conversely, the overall amount of anthocyanin was highly and positively associated with a*, phytate, and betacyanin (R2 = 0.89, 0.86, and 70), whereas it exhibited a substantial negative association with L* and pH (R2 = 0.914, 0.819). Furthermore, there was a significant positive relationship between the betacyanin and tannin concentration (R2 = 0.75), as well as a substantial negative association with L* (R2 = 0.874). In addition, there was a significant positive association between the phytate content and a* (R2 = 0.89), as well as a negative link with L* (R2 = 0.77). Additional observations revealed a negative correlation between the minerals (Na, K, Ca, Mg, and Fe) and the levels of anthocyanin, betacyanin, and phytate. It was observed that the rose blossom became darker in color as the levels of anthocyanin, betacyanin, and phytate increased. Conversely, the flower's color faded with an increase in mineral content. From an antioxidant perspective, the IC50 value demonstrated a significant inverse relationship with Ca (R2 = 0.85), similar to the correlation between TSS and pH (R2 = 0.82) (Fig. 5).
A heatmap with a dendrogram was created using 25 factors to cluster ten rose accessions for the purpose of cluster analysis (Fig. 5B). The analysis showed that the 25 factors were divided into two primary clusters, each making a major contribution to grouping the ten rose accessions into three clusters. Regarding the variable cluster, Cluster I consist of the elements Na, K, Alk, TFC, L, pH, Ca, Fe, whereas Cluster II comprises the remaining 17 variables h, b, VitA, X. carn, MMC, Mg, VitE, PHT, an AOA, BTC, TNN, TPC, SPN, IC50, C, and TSS. Cluster I accounted for 32% of the variables, whereas Cluster II accounted for 68%. Once again, cluster II was subdivided into two subclusters, which were then further fragmented into smaller clusters (Fig. 5B). On the other hand, these 25 factors categorized the ten rose accessions into three clusters. Cluster I consisted of R4, R3, R5, and R8. Cluster II included R10 and R9, while cluster III comprised R6, R7, R1, and R2.
Principal component analysis (PCA)
The previous section findings revealed that the variables made a substantial contribution to the classification of the rose accessions. The PCA was conducted to ensure consistency of the data and assess the extent of variation among variables. Principal Component Analysis (PCA) is a form of multivariate analysis that transforms large, intricate datasets with associated variables into groupings in order to uncover the most influential characteristics. The PCA biplot diagram visually represented the relationships, both similarities and dissimilarities, among the various parameters in Fig. 7A and 7B. The diagram specifically focused on the first dimension (PC1) and the second dimension (PC2); the initial two principal components (PCA), account for 54.6% of the overall data variance. Specifically, PC 1 and PC 2 individually account for 32.4% and 22.2% of the variance, respectively. The factor loadings and scores for the first two principal components (Dim 1 and Dim 2) of color parameters, secondary metabolites, and mineral matters of ten rose accessions (Fig. 6) indicated that a positive score on Dim1 was associated with L, b, pH, Vit A, X.Carn, TCP, Na, Ca, Mg, and Fe (ranging from 0.05 to 0.3), while all the other variables had moderate to low negative scores (ranging from − 0.04 to -0.25). Conversely, eleven variables, namely pH, Vit E, TFC, AOA, BTC, Na, Fe, PHT, ALK, SPN, and TNN, exhibited a positive score in Dim 2. In contrast, the variables L, b, c, h, MC, TSS, VitA, X. carn, TCP, IC50 K, Ca, and Mg contributed negatively to the score in Dim 2.
Upon examination of Fig. 7A and 7B, it is clear that the variables X. carn, b, VitA, h, c, a, L, pH, TSS, PHT, BTC, and AOA had the greatest influence on the selection of rose accessions. Consequently, the 10 different rose genotypes were clearly separated into three groups as a result of a positive association in both directions. In this case, the two accessions R9 (salmon color) and R10 (yellow color) that have the greatest influence on the variables may be easily differentiated from the others due to their strong positive correlation with both dimensions of the biplot. However, the biplot displays both the observations and variables in a given orientation along the PC axis (Dim1 and Dim2) concurrently. The orientation of the variable arrows signifies the direction in which the contribution of the related variable experiences the greatest rise, while the length of the arrows represents the magnitude of the change in that direction.
Cluster analysis
Cluster analysis was conducted using the K-means algorithm to group ten rose accessions based on 25 quantitative attributes. To do this, a dendrogram was constructed and then divided at a rescaled distance of 7.5. This division resulted in the formation of three separate clusters of roses, each exhibiting significant similarities in terms of the studied attributes (Fig. 8). Table 7 provides a compilation of three clusters of rose accessions in relevance with the CIELAB system and their visual evidence. The cluster I consisted of four rose accessions (R3, R4, R5, R8), which accounted for 40% of the plant population, just like cluster III (R1, R2, R6, R7). Cluster II consisted of R9 and R10 rose genotypes, which accounted for 20% of the population. Considering the contribution of rose accession in the CIELAB it has been revealed that the cluster I with the highest contribution of L* and the lowest b* value and which visualized the light color flower, cluster II with the highest contribution of b*, h° and cluster III with the highest contribution of a*, C* and appeared as bright color and dark color, respectively.