The plant material of sweet potatoes—18 cultivars—was specifically selected and supplied by Bodles Research Station, Old Harbour, St. Catherine, Jamaica. The cultivars have been selected because of their predominance in the local market, and their agronomic and marketability are appreciated by the local farmers. The samples were transported to the laboratory, washed, left to drain at room temperature for two hours, and then frozen at –20°C until use.

The chemicals used in this study, methanol and HCl, were obtained from Sigma-Aldrich (St. Louis, Missouri, USA). The reagents and mixtures were freshly prepared prior to the analyses.

Total carotenoids and anthocyanins were extracted by the method described by Lindoo and Caldwell [12] using acidified methanol for optimal extraction with slight modifications. The modifications consist of extracting carotenoids and anthocyanins at room temperature for four hours, after samples have been homogenised in the extraction solvent. Briefly, 50 g of frozen sweet potato samples were homogenized with 75 mL of methanol:water:concentrated HCl (MeOH:H₂O:HCl/80:20:1) and left at room temperature to stand for 4 hours to ensure better stability and recovery rate (> 90%) of carotenoids. After, the homogenate was filtered through a cheesecloth. The filtrate was made up to 200 mL using acidified methanol. For the spectrophotometric analysis, 10 mL of the extracts were centrifuged at 2,000 × g for 3 minutes, and the collected supernatants were used for carotenoids and anthocyanins analyses.

Total carotenoid content was analysed using the method described by Sumanta et al. [13]. A sample of 0.5 mL extract was mixed with 4.5 mL of the extraction solvent (MeOH:H₂O:HCl) and vortexed for 10 s, and the mixture was used to assess carotenoid content. The absorbance of the mixture was measured at 470 nm, 652 nm, and 665 nm using a Genesys spectrophotometer (Model Genesys 10S, Thermo-Fisher Inc., Waltham, Massachusetts, USA), and the extraction solvent (MeOH:H₂O:HCl) was used as a blank. After calculating the concentrations of chlorophyll-a (C-a) and chlorophyll-b (C-b), total carotenoid content was calculated using the following equation:

C – a = 16.72 × A 665.2 - 9.16 × A 652.4

C – b = 34.09 × A 652.4 - 5.28 × A 665.2

C a r o t e n o i d s c + x = 1000 × A 470 - ( 1.63 C – a + 104.96 C – b ) 221

Where A is absorbance at the indicated wavelength (470 nm, 652.4 nm, and 665.2 nm), C-a = chlorophyll-a, and C-b = chlorophyll-b, c+x stands for carotenoids and xanthophylls.

Total anthocyanins were extracted by the same method used for the extraction of carotenoids and analysed using the method described by Lee et al. [14]. To perform anthocyanin analysis, centrifuged extracts were diluted using two pH buffers, one of pH 1.0 buffer (0.025 M potassium chloride) and the other of pH 4.5 buffer (0.4 M sodium acetate), and left to stand for 20–50 min at room temperature. Afterwards, the absorbance of the two test portions diluted with pH 1.0 buffer [C(a)] and pH 4.5 buffer [C(b)] was measured at 520 and 700 nm versus a blank cell filled with distilled water using Genesys spectrophotometer (Model Genesys 10S, Thermo-Fisher Inc., Waltham, Massachusetts, USA). Total monomeric anthocyanins are determined by the following equations:

(1) A = A 5 2 0 - A 700 p H   1.0 - A 5 2 0 - A 700 p H   4.5

(2) T M A μ g   p e r   g   F W = A × M W × D F × F V × 1000 ε × L

Where A is absorbance from Eq. 1, MW is the molecular weight of monomeric anthocyanins (MW = 449.2), ε is the molar absorptivity of cyanidin-3-glucoside (ε = 26,900), DF is the dilution factor, FV is the final volume of the extract, and L is the path length.

DM of the samples was determined by the method described by Yildirim et al. [15]. Samples of 5 g were dried at 65°C until constant weight. The final and initial weight differences were used to calculate the DM percentage. Ash content of the samples was determined using the method described by the AOAC [16]. Samples of 5 g were placed in a crucible and dried in an oven at 105°C for 3 h. Then, the samples were allowed to incinerate in a pre-heated muffle furnace at gradually increasing temperatures from 350°C until a temperature of 500°C for 2 h. The final and initial weight differences were used to calculate ash content.

Data were treated by analysis of variance (ANOVA) to compare the data of the different cultivars, and differences among means were determined by the Least Significant Difference (LSD) test with significance defined at P < 0.05. The data were standardised, and PCA (principal component analysis) and HCA (hierarchical cluster analysis) were performed using SPSS software package (version 29.2. IBM Corp., New York, USA). In order to ensure that groups share good correlation, for the HCA analysis, the Pearson correlation coefficients were selected as the measurement between the cultivars, the furthest neighbour as the clustering method, and the Squared Euclidean distance (SE) for interval measurement. The clustered heatmap was generated using SRPLOT free software (http://www.bioinformatics.com.cn/srplot).

Total carotenoids in the different cultivars of sweet potato varied significantly from as low as less than 0.5 μg/g FW in Cock Up cultivar to 79.35 μg/g FW in the Beauregard variety (Table 1). The classification to low, medium and high carotenoid contents shows six cultivars [Eustace, Fire on Land, Clarendon, Cock Up, TIS Clarendon, and Watson (Unknown2)] contained low carotenoids level (less than 5 μg/g FW), eight cultivars [Keri Reid, TIS3030, Blue Bud, Ms. Mac, Minda and Yellow Belly, Bonita (Unknown1) and Variety X] contained medium level of carotenoids (5 to 10 μg/g FW), and four cultivars (Okinawa Purple, Uplifter, Peter Clarke and Beauregard) contained more than 10 μg/g FW.

Among the cultivars analysed in this study, the Okinawa Purple cultivar had the highest anthocyanin content, yielding 3,382.88 μg/g FW, while in Ms. Mac, Cock Up, and Beauregard cultivars, anthocyanins were not detected (Table 2). Okinawa Purple cultivar was the only purple-fleshed cultivar analysed in this study, and as such, it was expected that this cultivar would contain a higher level of anthocyanins. In addition to the flesh colour of the cultivar, the colour intensity of the flesh also aided in the assumption that anthocyanin content would be very high.

The cultivar Clarendon followed Okinawa Purple with an anthocyanin content averaging 2,685.92 μg/g FW. This cultivar had a cream flesh colour, and the observed anthocyanin content may be due to the composition of the cultivar, which is not exhibited by its appearance. Beauregard cultivar was one of the three cultivars in which anthocyanins were not detected. Beauregard cultivar is characterized by an orange flesh colour and shows the highest carotenoid content among the analysed cultivars. This indicates that sweet potatoes with deep orange flesh colour do not contain or contain much less anthocyanins, which are present in other sweet potato cultivars. Additionally, in Cock Up cultivar, no anthocyanins were detected, while carotenoid content was less than 1 μg/g FW. This indicates that this cultivar is not a significant source of anthocyanins or carotenoids and, therefore, it is not suitable as an ingredient to nutritionally enhance other food products.

DM of the different cultivars varied significantly (Table 3). All the investigated cultivars yielded more than 30% DM except the Okinawa cultivar, which yielded 21.96%. On the other hand, thirteen cultivars yielded more than 40% of DM. Comparatively, cultivar Watson showed the highest DM content (46.46%) while the Okinawa Purple cultivar showed the lowest DM content (21.96%).

The Eustace cultivar had the highest ash content, yielding 1.20%, and the Peter Clarke cultivar the lowest, yielding 0.09% (Table 4). The ash content of the cultivar with the highest carotenoid content (Beauregard) and the highest anthocyanin content (Okinawa Purple) was much lower than cream coloured flesh cultivars such as Variety X, Eustace, and Fire on Land. On the other hand, the high ash content observed in the cream-coloured flesh cultivars might be due to the physiochemical composition of the cultivars and indicative of the presence of numerous minerals in sweet potato cultivars. Conversely, the low ash content observed in the purple and orange fleshed cultivars might be related to poor mineral content in comparison to their cream-coloured flesh counterparts.

Here, we are describing the use of the statistical hypothesis testing of factor loading in PCA using the analysed factors of the eighteen cultivars of sweet potato (Figure 1). PCA of the studies’ cultivars revealed two principal components and were cumulatively accounted for 79.6% of the total variance, with PC1 explaining 45.9% and PC2 explaining 33.7% of the total variance, respectively (Figure 1A). On the other hand, Figure 1B illustrates the corresponding loadings plot showing carotenoids, anthocyanins, DM, and ash for the classification of the different cultivars by PCA analysis. PCA revealed two principal components and were cumulatively accounted for 74% of the total variance, with PC1 explaining 38.9% and PC2 explaining 35.1% of the total variance, respectively.

To further clarify the difference between the studied cultivars, HCA was performed, hierarchically clustering each cultivar based on their anthocyanins, carotenoids, DM, and ash contents (Figure 2). Pearson’s coefficients were calculated to determine the significant correlations between the eighteen cultivars. In order to also have a clear visualization of anthocyanins, carotenoids, DM, and ash contents of the eighteen cultivars, these parameters were visualized according to a heatmap colour representation (Figure 3). The hierarchical clustering (dendrogram) and heat mapping evidenced the presence of three main groups (clusters). The first is represented by DM, indistinctively clustered in this group. The second group is instead represented by carotenoids, and the third group by anthocyanins.

Conclusively, in this study, an attempt was made to segregate eighteen varieties of sweet potato grown in Jamaica based on carotenoids, anthocyanins, DM, and ash contents. Four varieties showed carotenoid content higher than 10 μg/g FW, nine varieties showed anthocyanin content higher than 500 μg/g FW, and the highest carotenoid and anthocyanin content were found in Beauregard and Okinawa purple, respectively. Statistically, the PCA revealed two principal components, which cumulatively ranged from 74% and 79.6%, while the HCA evidenced the presence of three main groups, namely anthocyanins, carotenoids, and ash. The results of the current study provided a better understanding of the nutritional quality of the different cultivars of sweet potato grown in Jamaica, and the data can be recognized and used integrally to improve the diet and enhance some agro-processed foods. Therefore, from these findings, which showed a marked difference in carotenoid and anthocyanin contents of the numerous cultivars of sweet potato grown in Jamaica, we can recommend some cultivars of sweet potato that could contribute significantly to improving local human nutrition by developing new products such as snacks or baked products with enhanced nutritional value.