Genome-wide association studies of five free amino acid levels in rice

Abstract

Rice (Oryza sativa L.) is one of the important staple foods for human consumption and livestock use. As a complex quality trait, free amino acid (FAA) content in rice is of nutritional importance. To dissect the genetic mechanism of FAA level, five amino acids’ (Val, Leu, Ile, Arg, and Trp) content and 4,325,832 high-quality SNPs of 448 rice accessions were used to conduct genome-wide association studies (GWAS) with nine different methods. Of these methods, one single-locus method (GEMMA), seven multi-locus methods (mrMLM, pLARmEB, FASTmrEMMA, pKWmEB, FASTmrMLM, ISIS EM-BLASSO, and FarmCPU), and the recent released 3VmrMLM were adopted for methodological comparison of quantitative trait nucleotide (QTN) detection and identification of stable quantitative trait nucleotide loci (QTLs). As a result, 987 QTNs were identified by eight multi-locus GWAS methods; FASTmrEMMA detected the most QTNs (245), followed by 3VmrMLM (160), and GEMMA detected the least QTNs (0). Among 88 stable QTLs identified by the above methods, 3VmrMLM has some advantages, such as the most common QTNs, the highest LOD score, and the highest proportion of all detected stable QTLs. Around these stable QTLs, candidate genes were found in the GO classification to be involved in the primary metabolic process, biosynthetic process, and catalytic activity, and shown in KEGG analysis to have participated in metabolic pathways, biosynthesis of amino acids, and tryptophan metabolism. Natural variations of candidate genes resulting in the content alteration of five FAAs were identified in this association panel. In addition, 95 QTN-by-environment interactions (QEIs) of five FAA levels were detected by 3VmrMLM only. GO classification showed that the candidate genes got involved in the primary metabolic process, transport, and catalytic activity. Candidate genes of QEIs played important roles in valine, leucine, and isoleucine degradation (QEI_09_03978551 and candidate gene LOC_Os09g07830 in the Leu dataset), tryptophan metabolism (QEI_01_00617184 and candidate gene LOC_Os01g02020 in the Trp dataset), and glutathione metabolism (QEI_12_09153839 and candidate gene LOC_Os12g16200 in the Arg dataset) pathways through KEGG analysis. As an alternative of the multi-locus GWAS method, these findings suggested that the application of 3VmrMLM may provide new insights into better understanding FAA accumulation and facilitate the molecular breeding of rice with high FAA level.

Genome‐wide association studies for rice seed germination under drought stress using 3VmrMLM

“…As an advanced tool, the 3VmrMLM first selected all putative QTNs through a single‐locus model and then used empirical Bayes and the likelihood ratio test in a multi‐locus model to capture true QTNs. Multiple studies have confirmed that 3VmrMLM has a better detection ability for dissecting complex traits compared to other models (He et al., 2022; Yu et al., 2022; Zhang et al., 2022). In this study, 3VmrMLM was used to identify QTNs for two drought‐related traits at the germination stage in 165 rice accessions.…”

“…This framework can be used to detect and estimate the genetic effects for each genotype when controlling for all the possible polygenic backgrounds (Li, Zhang, Xiang et al., 2022). The 3VmrMLM method has been applied to identify potential QTNs and unbiasedly estimate their effects on complex quantitative traits in plants, such as coleoptile length (Wei et al., 2022), epicotyl length (Hong et al., 2022), flowering time (Han et al., 2022), yield (Rabieyan et al., 2022; Wang et al., 2022; Zhang et al., 2022), seed oil content (Zuo et al., 2022), tocopherol content (Yu et al., 2022), and amino acid level (He et al., 2022). Therefore, 3VmrMLM could be considered as an alternative method for identifying key genes involved in seed germination.…”