Figure 2. Plant height and seeds weight scaling in wild and cultivated barley. A. pattern of scaling in wild barley, landrace and cultivars. B. Kolmogorov-Smirnov plot showing the phenotypic distribution of plant height in in wild barley, landrace and cultivars. C. Kolmogorov-Smirnov plot showing the phenotypic distribution of seeds weight in in wild barley, landrace and cultivars.
Both barley cultivars and landraces show a positive plant height and seeds weight scaling with r = 0.422, p = 0.0001 for cultivars (N=3390), and r = 0.230, p = 0.0001 for landraces (N5740), respectively (Figure 2A). However, plant height and seeds weight scaling is not evident in wild barley (N=292) withr = 0.0368 and p = 0.5233) (Figure 2A), which is probably due to relatively small number of wild barley samples in the analysis. Significant difference between phenotypic distribution in wild barley, landrace and cultivar were observed. Kolmogorov-Smirnov test for equal distribution for plant height between cultivar and landrace returned a D = 0.128 and p = 0.0001, and a D = 0.127, p = 0.0001 for seeds weight. Domestication led to a shift to shorter status, from a median of 107.9 cm in wild barley to 98.8 cm in landraces and 102.2 cm in cultivars (Figure 2B). Meanwhile, the shift of plant height to shorter status was accompanied by a shift to smaller seeds, as it would be expected from a positive plant height and seeds weight scaling. Median seeds weight decreased from 46.5 g in wild barley to 45.3 g in landraces and 43.5 g in cultivars (Figure 2C).
With the filtrations, a total of 133,588 SNPs were obtained for the 12,828 samples. Wild barley has a relatively higher genomic diversity (nucleotide diversity π = 0.0673, N = 292) than both landraces (π = 0.0535, N = 5740) and cultivars (π = 0.0485, N = 3390). Plant height is highly heritable with SNP-based heritability hSNP2 of 0.603 ± 0.095, while SNPs explained less variation in seeds weight with hSNP2 of 0.322 ± 0.100. Plant height and seeds weight are genetically correlated (r G = 0.272 ± 0.020,p < 0.0001). Barley accessions with closer genetic relatedness tend to have similar plant height and to have similar seeds weight, as indicated by the phylogenetic signal lambda of 0.822 andp < 0.001, while by the lambda of 0.853 and p< 0.001 for seeds weight. Generalised least-squares fit by REML indicate that the two traits are phylogenetically correlated (p < 0.001), implying that the two traits tend to vary correlated toward the similar direction of the phenotypic spectrum.
Genome-wide association studies revealed 314 SNPs, or 0.23% of the 133,588 SNPs in total, associated with plant height, and 190 (0.14%) with seeds weight. Among them, eight SNPs are associated with both traits (Figure 2). The eight SNPs, forming two clusters in chromosome 2H and 5H (Figure 2), can be traced to at least three functional genes (Table 1), with two genes in cluster 1 (an Expansin B3 gene and an Elongation Factor G gene), and one gene in cluster 2 (Vacuolar protein sorting-associated protein 18 gene).