Gains in QTL detection using an ultra-high density SNP map based on population sequencing relative to traditional RFLP/SSR markers

Abstract

Huge efforts have been invested in the last two decades to dissect the genetic bases of complex traits including yields of many crop plants, through quantitative trait locus (QTL) analyses. However, almost all the studies were based on linkage maps constructed using low-throughput molecular markers, e.g. restriction fragment length polymorphisms (RFLPs) and simple sequence repeats (SSRs), thus are mostly of low density and not able to provide precise and complete information about the numbers and locations of the genes or QTLs controlling the traits. In this study, we constructed an ultra-high density genetic map based on high quality single nucleotide polymorphisms (SNPs) from low-coverage sequences of a recombinant inbred line (RIL) population of rice, generated using new sequencing technology. The quality of the map was assessed by validating the positions of several cloned genes including GS3 and GW5/qSW5, two major QTLs for grain length and grain width respectively, and OsC1, a qualitative trait locus for pigmentation. In all the cases the loci could be precisely resolved to the bins where the genes are located, indicating high quality and accuracy of the map. The SNP map was used to perform QTL analysis for yield and three yield-component traits, number of tillers per plant, number of grains per panicle and grain weight, using data from field trials conducted over years, in comparison to QTL mapping based on RFLPs/SSRs. The SNP map detected more QTLs especially for grain weight, with precise map locations, demonstrating advantages in detecting power and resolution relative to the RFLP/SSR map. Thus this study provided an example for ultra-high density map construction using sequencing technology. Moreover, the results obtained are helpful for understanding the genetic bases of the yield traits and for fine mapping and cloning of QTLs.

Figure 1. Recombination bin map constructed using high quality SNPs from sequencing genotyping of the RIL population.

(A) Whole map of 1,619 recombination bins for the 210 RILs. Chromosomes are separated by vertical gray lines. (B) The map of the first 50 bins on chromosome 1 for the first 20 RILs. The vertical gray lines indicate the recombination breakpoints. A region between two vertical lines across all RILs is recognized as a recombination bin. Physical positions are based on rice TIGR6.1. Red, Zhenshan 97 genotype; Blue, Minghui 63 genotype.

Figure 2. Comparative genotyping of R001 on chromosome 1 with different markers.

(A) RFLPs and SSRs. (B) Microarray-based SFPs. (C) Bin map based on SNPs constructed in this study. All positions are transformed to physical positions according to rice TIGR6.1.

Figure 3. Cosegregation analysis of the trait values of apicule color and genotypes of the recombination bins.

The x-axis shows the positions of the bins distributed on rice 12 chromosomes. Chromosomes are separated by the vertical gray lines. The y-axis indicates the number of RILs that the apicule color cosegregating with the genotypes at each bin. The peak is at Bin868 on chromosome 6, showing complete cosegregation for all the 210 RILs.

Figure 4. Precise locations of GS3 and GW5/qSW5 in the SNP bin map.

(A) LOD curves of QTL mapping for grain length on chromosome 3. Short lines on x-axis indicate the genetic positions of the bins. (B) Physical mapping of GS3. Short lines on x-axis indicate the boundaries of the bins. The exact position of GS3 is indicated by the black dash line. (C) LOD curves of QTL mapping for grain width on chromosome 5. Short lines on x-axis indicate the genetic positions of the bins. (D) Physical mapping of GW5/qSW5. Short lines on x-axis indicate the boundaries of the bins. The exact position of GW5/qSW5 is indicated by the black dash line. Red curves indicate the data from 1998 and blue curves indicate the data from 2008.

Figure 5. QTL mapping for yield and yield-component traits using the SNP bin map.

The phenotype data are from Xing et al collected in 1997 (Xing1997) and 1998 (Xing1998), and Hua et al – collected in 1998 (Hua1998) and 1999 (Hua1999). Four traits, grain yield/plant, tillers/plant, grains/panicle and grain weight, are shown from top to bottom. A triangle indicates a QTL detected above the LOD threshold by the permutation test (1000 permutations, P = 0.05) in only one trail. An arrow indicates a QTL identified in at least two trials.

Figure 6. Comparison of QTL mapping for gn3 using different maps.

LOD curves for number of grains per panicle in gn3 region on chromosome 3 are shown. The phenotype data are from Xing et al collected in 1997 (black lines) and 1998 (red lines), and Hua et al – collected in 1998 (green lines) and 1999 (blue lines). Physical positions are indicated in x-axis. (A) Using the SNP bin map. The short lines on x-axis indicate the positions of the recombination breakpoints. (B) Using the RFLP/SSR map. The short lines on x-axis indicate the positions of the markers.

Figure 7. QTL mapping for grain length and grain width using the SNP bin map.

Red lines show the LOD curves for the phenotypic data from Tan et al collected in 1998 and blue lines show the LOD curves for the phenotypic data collected in 2008. A triangle indicates a QTL detected above the LOD threshold by the permutation test (1000 permutations, P = 0.05) in only one year. An arrow indicates a QTL identified in two years.