Biological material

Three samples composed of bean seed from Santo Domingo del Estado, Oaxaca, were used and collected in 2016 at homes where bean mixtures with different colors and sizes are planted and used for self-consumption. Each sample was assigned a number (14 979, 14 981 and 14 983).

Morphological parameters of the seed

200 g of each composite sample were weighed and grouped by color and grain size. Each compound was divided into subsamples based on their size and color.

Color

It was determined visually and the seeds that constituted more than 10% in each sample, as they are the most representative.

Weight and size

The weight of 100 grains was measured, which were randomly selected and weighed on a digital semi-analytical balance. Those with a weight less than 25 g 100 seeds^-1 were classified as small grains, medium those whose weight was 25 to 40 g 100 seeds^-1 and large when their weight was greater than 40 g 100 seeds^-1 (Allende, 2006).

Analysis of the content of total sugars and crude protein

The analyzes were performed in duplicate. The total sugar content was determined using the Antrona method (^{Yemm y Willis, 1954}). The UV-VIS spectrophotometer was read at 630 nm. The protein content was determined according to the Kjeldahl method using the Kjeltec-1030 semi-automated equipment. The protein percentage was calculated from total nitrogen using factor 6.25.

The data were analyzed under a completely randomized design, three treatments (composite samples) and two to three subsamples, each with two repetitions. The General Linear Model (GLM) procedure of the Statistical Analysis System version 8 (^{SAS Institute, 1999}) was used. The comparison of means was made with the Tukey test (p< 0.05). In addition, a Pearson correlation matrix was obtained.

Table 1 shows the color of the subsamples with the highest to lowest predominance of the bean grains into which each composite sample was subdivided, as well as the weight of 100 seeds, the protein content and total sugars.

Four to five different colors and grain sizes were identified in each composite sample, including small (<25 g 100 seeds^-1), medium (between 25 and 40 g 100 seeds^-1) and large (> 40 g 100 seeds^-1). Small and medium grains were identified as common beans (Phaseolus vulgaris L.), while large ones were identified as ayocote. (Phaseolus coccineus L.) black and red P. coccineus beans, as well as yellow and black P. vulgaris beans, were the most frequent among the three composite samples; followed by pints, wine and beige, the latter represented a proportion less than 10% in the samples.

The grain weight varied from 20 g 100 seeds^-1 to 73.3 g 100 seeds^-1. Highly significant statistical differences were detected between the composite samples (p≤ 0.01) for the variable weight of 100 seeds. Sample 14 979 was made up of medium-sized P. vulgaris, although of different colors, weighing between 30 and 33 g 100 seeds^-1. The predominant grain size in two of the samples was the large one (>40 g 100 seeds^-1), since it is P. coccineus, which represented 74% of the 14 981 and 100% of the sample 14 983. The grains of size Small were rare and accounted for less than 2% of the sample 14 981.

Sample 14979 contained yellow, black, red and pinto grains. Three of the four colors that made it up are similar to those reported by ^{Espinosa-Pérez et al. (2015)}, for the 040 collection of Oaxaca, which is a Ensaladilla type with black, red and pinto black beans and its size between small and medium, does not exceed 36 g 100 seeds^-1. In sample 14 981 the red grain P. coccineus (73.3 g 100 seeds^-1) was the predominant and 24% of the sample consisted of yellow and black P. vulgaris of medium size (25.3 and 34.7 g 100 seeds^-1 respectively, and only 2% small-colored P. vulgaris beige (20 g 100 seeds^-1). Finally, in sample 14 983 the four beans were P. coccineus, 77% black and 23% colors pinto, wine and beige.

Total sugar content

The percentage of total sugars in grain types was 6.2 to 6.5%, in P. vulgaris grains, classified in medium size, while in P. coccineus, classified in large size, sugars accounted for 9.5 to 11.7%. Highly significant statistical differences were detected between the composite samples (p< 0.01) in the total sugar content. The concentration of total sugars was positively correlated with the grain size (r= 0.88^**).

Differences in the sugar content of the different grain colors were observed; however, the most notable differences were observed between the two species: P. vulgaris and P. coccineus. Also in green beans, differences between genotypes in the amount of total sugars have been observed, while in Puebla 152 common bean it was observed that as the development of the plant progressed the concentration of some sugars was modified, the sucrose content remained stable, while fructose and glucose increased (Vanden Langenberg et al., 2012).

The higher sugar content in P. coccineus compared to P. vulgaris, is associated with a sweeter taste, which according to the work of (^{Mkanda et al., 2007}; ^{Wszelaki et al., 2005}). It can motivate greater acceptance by consumers. The above is reaffirmed with testimonies of the women of Santo Domingo in the state, who affirmed that the black and red ayocote beans are the best tasting, so they are part of the traditional dishes; they are used to make bean atole (nakinj runee), one of the typical dishes of the High Triqui region, can be consumed on the day of the dead and in the wake (^{Martínez-López et al., 2018}). Figure 1 shows the total sugar content and the grain size of each bean subsample in the three composite samples.

Each bar texture groups the subsamples that make up a compound.

Figure 1 Total sugar content and grain size of three bean compounds (P. vulgaris and P. coccineus).  

The differences in the content of total sugars between the P. vulgaris and the P. coccineus that make up the sample, suggest that each of the samples, according to the predominance of one or the other of these bean species will have greater or lesser intensity Sweet taste and with it its organoleptic attributes will be different. Protein content showed significant statistical differences between the composite samples (p< 0.05). The variation between subsamples was from 21.7% in P. coccineus black, to 24.8% in P. vulgaris yellow (Table 1).

The protein content in the P. vulgaris genotypes was within the range reported for beans native to different areas of Mexico that was between 17.8 and 26.5% (^{Jacinto-Hernández et al., 2002}). The tendency to higher protein content was observed in P. vulgaris beans, whose weight varied between 25.3 and 34.7 g 100 seeds^-1, compared to P. coccineus that showed weight between 61.3 to 73.3 g 100 seeds^-1.

The above contributed to a negative correlation between grain size and protein content (r= -0.61^*). However, differences in protein content between genotypes of the same species were also observed. Figure 2 shows the protein content and grain size of each bean subsample in the three composite samples.

Each bar texture groups the subsamples that make up a compound.

Figure 2 Crude protein content and grain size of three bean compounds (P. vulgaris and P. coccineus).  

The composite sample 14 979 the protein range in the P. vulgaris samples ranged from 23.7 to 24.8%, the medium-sized yellow and red beans were the ones with the highest protein content with 24.8 and 24.5% (Table 1), while in sample 14981 both P. coccineus and P. vulgaris showed close values, 23.3 and 23.4% respectively.

Compound 14983 showed a difference between the two P. coccineus samples that comprise 21.7% black and 23.5% pinto- wine protein. Grain color was not associated with protein content and total sugars; however, the grain size does. A larger grain size, a higher percentage of sugars (r= 0.88^**) and a lower protein content (r= -0.61^*).

Conclusions

Higher sugar content was observed in P. coccineus beans compared to P. vulgaris; while the latter tended to show higher protein content. A larger grain size was associated with higher total sugar content (r= 0.88^**) and lower protein content (r= -0.61^*).

The composite sample 14 979, made up of different colors of P. vulgaris, showed higher average protein content (24.3%) while the sample 14 983 formed by P. coccineus showed the sweetest taste (10.7% total sugars), although lower average protein content (22.7%).

The values obtained indicated that the grain color was not associated with the protein or total sugar content.

The natural variation that tends to exist between the content of crude protein and total sugars between P. vulgaris and P. coccineus, can be used to design compounds with different sensory and nutritional characteristics that can be incorporated into both local markets in rural and urban populations.