Genetic variation underlying kernel size, shape, and color in two interspecific S. bicolor2 × S. halepense subpopulations

Abstract

Throughout the history of sorghum domestication, kernel traits have been subject to extensive selection. Breeding of grain sorghum is highly dependent on kernel morphology, which influences yield and quality. This study examined the genetic variation and architecture of kernel size, shape, and color among two BC 1 F 2 subpopulations, H4 and H6, derived from crossing Sorghum bicolor BTx623 and Sorghum halepense Gypsum 9E. We phenotyped 246 BC 1 F 2 families and their parents in two locations for two years using high-throughput digital imaging techniques, to determine fourteen kernel traits in five broad categories: size (area, length, width, and aspect ratio), shape (circularity and PC1), density (factor form density; FFD), weight (1000kernel weight) and color (RGB and CIE-L*a*b*). Based on single-trait, interval mapping, we identified 76 and 71 significant QTLs in the H4 and H6 subpopulations, respectively. Both parent genotypes contributed QTL alleles that conferred positive additive effects, indicating that Sorghum halepense contains alleles that may enhance some kernel-related traits of elite sorghums. Some genomic regions affect many traits-for example, a linkage group 6 homolog in the H4 subpopulation was associated with several kernel traits in the region between 0 and 57.3 cM, including FFD, area, length, and width; and in H6 between 56.8 and 185.9 cM contained QTLs associated with kernel color parameters (R, G, L* and B*), FFD, kernel shape (PC1), kernel area, aspect ratio, kernel width, and 1000-kernel weight. These results contribute to a better understanding of genetic factors governing sorghum kernel traits, providing a foundation for improving grain and quality traits in both annual and perennial sorghums using genomic tools.

Off-the-shelf image analysis models outperform human visual assessment in identifying genes controlling seed color variation in sorghum

“…Measurements of seed color made by extracting the RGB values of pixels corresponding to five seeds per genotype in photos of seeds from the sorghum association panel also identified signals from Y and tan1 (Zhang et al 2015). A more automated computer vision-based approach using the GRABSEEDS software package implemented within JCVI (Tang et al 2024) was also able to identify several QTL peaks, including peaks corresponding to tan1 and Y, when employed to quantify the seed colors of a set of several hundred BC 1 F 2 families (Nabukalu et al 2021).…”

Genetic Architecture of Grain Yield-Related Traits in Sorghum and Maize

“…In sorghum, nine QTLs have been detected for grain weight and grain width derived from wild Sorghum Virgatum and domesticated Sorghum bicolor [ 45 ]. In a recent study, fifteen significant QTLs for grain yield-related traits were identified from a population derived from wild Sorghum halepense and Sorghum bicolor [ 46 ].…”

“…To date, over 168 primarily mapped QTLs associated with grain size have been reported in sorghum. Among them, 155 QTLs located on ten chromosomes have been mapped for thousand grain weight (TGW) in various independent studies published between 1995 and 2021 [ 13 , 34 , 43 , 45 , 46 , 47 , 49 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 , 66 ] ( Figure 1 a and Table S1 ). Few QTLs related to grain size (21 QTLs for grain length, 23 QTLs for grain width, and 26 QTLs for grain number per panicle) have been identified.…”