An improved 7K SNP array, the C7AIR, provides a wealth of validated SNP markers for rice breeding and genetics studies

Abstract

Single nucleotide polymorphisms (SNPs) are highly abundant, amendable to high-throughput genotyping, and useful for a number of breeding and genetics applications in crops. SNP frequencies vary depending on the species and populations under study, and therefore target SNPs need to be carefully selected to be informative for each application. While multiple SNP genotyping systems are available for rice (Oryza sativa L. and its relatives), they vary in their informativeness, cost, marker density, speed, flexibility, and data quality. In this study, we report the development and performance of the Cornell-IR LD Rice Array (C7AIR), a second-generation SNP array containing 7,098 markers that improves upon the previously released C6AIR. The C7AIR is designed to detect genome-wide polymorphisms within and between subpopulations of O. sativa, as well as O. glaberrima, O. rufipogon and O. nivara. The C7AIR combines top-performing SNPs from several previous rice arrays, including 4,007 SNPs from the C6AIR, 2,056 SNPs from the High Density Rice Array (HDRA), 910 SNPs from the 384-SNP GoldenGate sets, 189 SNPs from the 44K array selected to add information content for elite U.S. tropical japonica rice varieties, and 8 trait-specific SNPs. To demonstrate its utility, we carried out a genome-wide association analysis for plant height, employing the C7AIR across a diversity panel of 189 rice accessions and identified 20 QTLs contributing to plant height. The C7AIR SNP chip has so far been used for genotyping >10,000 rice samples. It successfully differentiates the five subpopulations of Oryza sativa, identifies introgressions from wild and exotic relatives, and is useful for quantitative trait loci (QTL) and association mapping in diverse materials. Moreover, data from the C7AIR provides valuable information that can be used to select informative and reliable SNP markers for conversion to lower-cost genotyping platforms for genomic selection and other downstream applications in breeding.

Fig 1. Distribution of gap size between a SNP and its neighbors.

The majority of SNPs are within 100 kb of their neighbor, as seen by the histogram of SNP markers on the y-axis and density in basepairs on the x-axis.

Fig 2. p10 GC versus call rate for all samples.

The Illumina p10GC quality metric is shown on the y-axis and the SNP call rate on the x-axis; most samples had p10 GC values >0.45 and call rates > 0.94.

Fig 3. Phylogenetic tree from 6,514 SNPs across 551 Oryza sativa accessions.

The five subgroups of O. sativa are shown: indica: dark purple, aus: light purple, aromatic: green, temperate japonica: turquoise, tropical japonica: blue, admixed: black.

Fig 4. VanRaeden kinship heat map.

The relatedness of individuals compared to each other is shown, with the legend identifying subgroups of O. sativa.

Fig 5. GWAS results for plant height using SNPs from the C7AIR.

The manhattan plot shows the level of significance for SNPs correlated with plant height from the GWAS analysis (numbers across the x-axis indicate the chromosome). The Q-Q plot is shown at right. The blue line indicates the p <0.001 threshold and the green line indicates the Bonferroni correction threshold.