Genetic Diversity, Population Structure, and Andean Introgression in Brazilian Common Bean Cultivars after Half a Century of Genetic Breeding

Abstract

Brazil is the largest consumer and third highest producer of common beans (Phaseolus vulgaris L.) worldwide. Since the 1980s, the commercial Carioca variety has been the most consumed in Brazil, followed by Black and Special beans. The present study evaluates genetic diversity and population structure of 185 Brazilian common bean cultivars using 2827 high-quality single-nucleotide polymorphisms (SNPs). The Andean allelic introgression in the Mesoamerican accessions was investigated, and a Carioca panel was tested using an association mapping approach. The results distinguish the Mesoamerican from the Andean accessions, with a prevalence of Mesoamerican accessions (94.6%). When considering the commercial classes, low levels of genetic differentiation were seen, and the Carioca group showed the lowest genetic diversity. However, gain in gene diversity and allelic richness was seen for the modern Carioca cultivars. A set of 1060 'diagnostic SNPs' that show alternative alleles between the pure Mesoamerican and Andean accessions were identified, which allowed the identification of Andean allelic introgression events and shows that there are putative introgression segments in regions enriched with resistance genes. Finally, genome-wide association studies revealed SNPs significantly associated with flowering time, pod maturation, and growth habit, showing that the Carioca Association Panel represents a powerful tool for crop improvements.

Keywords: Carioca variety; Phaseolus vulgaris L.; common bean; genome-wide association study; genomic introgression; variability gain.

Figure 1

(a) Mean values of L (k) for the 20 runs of each K. (b) Plot of ΔK for each K value. (c) Structure analysis for 185 Brazilian cultivars grouped by commercial class (i.e., Black, Carioca, Special) and year of launch for the Carioca group (i.e., Carioca lines, modern Carioca, old Carioca) for K = 2 (top clustering), K = 5 (middle clustering), and K = 9 (bottom clustering).

Figure 2

(a) Principal component analysis based on the 2827 SNPs for the total set (n = 185) (b) and on only the Mesoamerican accessions (n = 175).

Figure 3

Discriminant analysis of the principal components carried out for the 185 Brazilian common bean cultivars. (a) Scatter plot of the two main components grouped by commercial class (colors and ellipses). (b) Representation of the compoplot function, showing the membership coefficient of each cultivar to each group (i.e., commercial class) ordered according to commercial class (i.e., Black, Carioca, Special), following the scatter plot colors. (c) Loading plot generated by the 2827 SNPs, showing the SNPs that contribute most to the separation of the cultivars between the three commercial classes.

Figure 4

Linkage disequilibrium (LD) decay determined by the LD measurements (r²) against the distance between SNPs (Mb) for the 11 chromosomes (Pv) adjusted according to the model proposed by Hill and Weir (1988). (a) Brazilian diversity panel controlled for population structure. (b) Carioca panel controlled for relatedness. The horizontal and vertical lines represent the standard LD decay thresholds (r²) and distances (Mb), respectively.

Figure 5

Structure analysis for each chromosome individually plotted for K = 2 to identify events of chromosome introgression of the Andean gene pool (red cluster) into the Mesoamerican accessions (n = 175). Each plot represents one chromosome (as indicated), and the bars represent the Mesoamerican cultivars.

Figure 6

Distribution of the 1060 diagnostic SNPs in the common bean genome. The 82 SNPs with Andean allele frequency from 11% to 15%, 16% to 19%, and > 20% in the Mesoamerican pool are in green, blue, and red, respectively. SNPs with Andean allele frequency < 10% are in black. The centromeric region of each chromosome is in pink.

Figure 7

Genome-wide association study of the Carioca association panel. (a–c) Manhattan plots showing the association between the SNP markers and the growth habit (a), flowering time (b) and pod maturation (c). The green dotted line passes through the SNPs with the lowest p-values for all of the trait associations on Pv01. (d) Linkage disequilibrium (r²) heatmap in the 10 Mb region surrounding the most significant SNPs (ss715646076, ss715639272) associated with all of the traits, which were located 389 kb and 299 kb, respectively, upstream of the PvTFL1y gene (Phuvl.001G189200). (e) Boxplots illustrating the relationships between alleles and phenotypes (as indicated) for the significant SNPs on Pv01.