JIA-2019-11

2453 CHEN Bing-ru et al. Journal of Integrative Agriculture 2019, 18(11): 2446–2456 of substrate and is one of the largest protein families. Sobic.006G237200 encodes an ATP synthase subunit D. KASP can be used for genotyping a broad range of species for various purposes, including quality control (QC) analysis, quantitative trait loci (QTL) mapping, marker- assisted recurrent selection (MARS), and allele mining applications (Semagn et al. 2013). In this study, 100 accessions were genotyped by KASP markers, A009649, one of seven KASPmarkers (Fig. 4-A), and the association of A009649 and starch content were significant using Tassel 5 (Table 3). The C allele was correlated with a high starch content, while the T allele was linked with a low level of starch content (Fig. 4-B). 4. Discussion GWAS has become a powerful tool for mapping quantitative traits and is now routinely applied in a range of plants and animals including Arabidopsis (Atwell et al. 2010), rice (Huang et al. 2010, 2012), foxtail millet (Jia et al. 2013), maize (Li et al. 2013), cattle (Olsen et al. 2011), and mouse (Flint et al. 2012). GWAS is used to map multiple traits and is also used for sets of related characters or a single character (Cui et al. 2016; Zheng et al. 2018). In the present analysis, sorghum grain association signals for starch, AM, and AM/ AP were located using a GWAS. 4.1. Phenotype variation of traits and correlation between starch and AP The starch content of sorghum grains in this population ranged in 60.28–74.03%. AM content ranged in 0.18– 28.21% andAP content ranged in 71.79–99.82%. Sorghum was shown to have extensive variation for starch (Shewayrga et al. 2012; Sukumaran et al. 2012; Rhodes et al. 2016). In non-waxy sorghum, starch and AP contents were positively correlated, but in waxy sorghum, starch content was negatively correlated withAP content. We deduced that when waxy gene was mutated, AM contents decreased, and starch contents correspondingly decreased. Meanwhile, the percentage of AP increased relative to starch. This mutation did not increase starch contents. So, starch contents of glutinous sorghum were lower than non-waxy sorghum. Another reason may be that there were few waxy sorghum accessions in this study and more waxy sorghum accessions need to be tested and verified. When AP content is about 90%, the color complexes of starch and iodine solution were reddish brown and the accession was called waxy sorghum; this result was consistent with waxy sorghum breeding. Based on the non-waxy sorghum breeding process, development and utilization of high grain starch sorghums breeding has occurred for more than 20 years. The positive correlation between grain yield and starch content is most likely the root cause of why high starch lines are widely used in the sorghum breeding (Boyles et al. 2017). We can cross or backcross between high starch non-waxy sorghum and waxy sorghum and select germplasm with high AP and starch contents to cultivate elite waxy sorghum parental lines with high grain yield. 4.2. Marker-trait associations and potential genes involved in starch biosynthesis by GWAS Starch synthesis in plant seeds is a very complicated process controlled by many genes with small effects. SNP rs12508174, which is significantly associated with starch content, explains 5.7% of the variation in the trait. The gene Sobic.006G036500.1 was 58.9 kb away from rs12508174; this gene encodes peptidyl-prolyl cis - trans -isomerase CYP38, which is responsible for hexose monophosphate shunt (HMS). The reaction initiator of HMS is glucose-6- SNP ID Allele P -value Associate gene Gene location The distance SNP away of candidate gene (kb) GO term 2) PFAM annotation Chr9.S8773750 G, A 3.22E-07 Sobic.009G072400 8881404– 8887144 107.654 ATP binding F ; plasma membrane C ; ATP catabolic process P ; ATPase activity, coupled to transmembrane movement of substances F ATPase domain Chr9.S8773750 G, A 3.22E-07 Sobic.009G072700 8916935– 8923148 143.185 carbohydrate metabolic process P ; response to glucose P ; chloroplast envelope C Membrane-bound O-acyltransferase family 1) SNP, single nucleotide polymorphism; PFAM, the PFAM database . 2) C , P , and F indicate cellular components, biological processes, and molecular functions in Gene Ontology (GO) terms, respectively. ATP, adenosine triphosphate; NADP, nicotinamide adenine dinucleotide phosphate. Table 2 (Continued from preceding page)