Background: Rice is an important human staple food vulnerable to heavy metal contamination due to its unique physiology and growth environment. High yield with low heavy metal contamination is a common but highly challenging goal for rice breeders worldwide due to lack of genetic knowledge. To identify candidate QTLs for rice yield and heavy metal content, sixteen ionomic traits and thirteen agronomic traits of the USDA Rice mini-core collection were analyzed using both univariate and multivariate GWAS methods in this study. The USDA Rice Mini-Core Collection contains about 1% of the whole Rice Collection of the National Small Grains Collection (NSGC), USA.Results: Using the p-value <1.53×10-8, this criterium p-value was determined by the Bonferroni correction for p-value of 0.05, 106, 47, and 97 QTLs were identified for ionomics in flooded environment, unflooded environment, and agronomic traits, respectively. A large number of QTLs coincide well with previous report results while many of the QTLs are new QTLs, suggesting the efficiency of GWAS methods and the reliability of this study. Our results further showed that each of the four GWAS methods can identify unique as well as common QTLs. When univariate methods failed to identify QTLs for a trait, the multivariate methods frequently detected QTLs. However, when many QTLs were detected by univariate methods, the number of QTLs detected by multivariate methods were reduced in many cases. These analyses suggest that using multiple GWAS methods can complement each other in QTL identification. In addition, several candidate genes involved in ionomic and agronomic traits control were identified by analyzing the sequences of the candidate QTL regions.Conclusions: Significant QTLs for heavy metal, mineral, and agronomic traits are presented in the rice genome and some of them have been fine mapped in the rice genome in this study. This research provides novel insights into the genetic basis of both ionomic and agronomic variations in rice, establishing an important foundation for further studies on reducing heavy-metal contamination and improving crop yields. In addition, the comparison analysis of the GAWS methods showed that each method has unique feature and different method can complement each other.

Rice is an important human staple food vulnerable to heavy metal contamination due to its unique physiology and growth environment. High yield with low heavy metal contamination is a common but highly challenging goal for rice breeders worldwide due to lack of genetic knowledge. In this report, a comprehensive GWAS analyses for ionomic and agronomic traits based on 3,259,478 SNPs were performed using two univariate methods and two multivariate methods. Under the criterium p-value <1.53×10-8, 106, 47, and 97 QTLs were identified for ionomics in flooded environment, unflooded environment, and agronomic traits, respectively. Detailed analysis of the QTLs revealed that many of the identified QTLs are co-localized with the QTLs reported in prior ionomic and agronomic studies or posited near the genes with known functions in the related traits, suggesting that our GWAS analyses are reliable. Our results further showed that each of the four GWAS methods can identify unique as well as common QTLs. When univariate methods failed to identify QTLs for a trait, the multivariate methods frequently detected QTLs. However, when many QTLs were detected by univariate methods, the number of QTLs detected by multivariate methods were reduced in many cases. These analyses suggest that using multiple GWAS methods can complement each other in QTL identification and some methods may be more powerful with less false discovery rate. In addition, several candidate genes involved in ionomic and agronomic traits control were identified by sequence analysis of the QTL regions. This research provides novel insights into the genetic basis of both ionomic and agronomic variations in rice, establishing an important foundation for further studies on reducing heavy-metal contamination and improving crop yields.

BackgroundRice is an important human staple food vulnerable to heavy metal contamination leading to serious concerns. High yield with low heavy metal contamination is a common but highly challenging goal for rice breeders worldwide due to lack of genetic knowledge and markers. ResultsTo identify candidate QTLs and develop molecular markers for rice yield and heavy metal content, a total of 191 accessions from the USDA Rice mini-core collection with over 3.2 million SNPs were employed to investigate the QTLs. Sixteen ionomic and thirteen agronomic traits were analyzed utilizing two univariate (GLM and MLM) and two multivariate (MLMM and FarmCPU) GWAS methods. 106, 47, and 97 QTLs were identified for ionomics flooded, ionomics unflooded, and agronomic traits, respectively, with the criterium of p-value <1.53×10-8, which was determined by the Bonferroni correction for p-value of 0.05. While 49 (~20%) of the 250 QTLs were coinciding with previous reported QTLs/genes, about 201 (~80%) were new. In addition, several new candidate genes involved in ionomic and agronomic traits control were identified by analyzing the DNA sequence, gene expression, and the homologs of the QTL regions. Our results further showed that each of the four GWAS methods can identify unique as well as common QTLs, suggesting that using multiple GWAS methods can complement each other in QTL identification, especially by combining univariate and multivariate methods. ConclusionsWhile 49 previously reported QTLs/genes were rediscovered, over 200 new QTLs for ionomic and agronomic traits were found in the rice genome. Moreover, multiple new candidate genes for agronomic and ionomic traits were identified. This research provides novel insights into the genetic basis of both ionomic and agronomic variations in rice, establishing the foundation for marker development in breeding and further investigation on reducing heavy-metal contamination and improving crop yields. Finally, the comparative analysis of the GWAS methods showed that each method has unique features and different methods can complement each other.