Genome-wide response to selection and genetic basis of cold tolerance in rice (Oryza sativaL.)

Abstract

BackgroundCold stress is an important factor limiting rice yield in many areas of high latitude and altitude. Considerable efforts have been taken to genetically dissect cold tolerance (CT) in rice using DNA markers. Because of possible epistasis and gene × environment interactions associated with identified quantitative trait loci, the results of these genetic studies have unfortunately not been directly applicable to marker-assisted selection for improved rice CT. In this study, we demonstrated the utility of a selective introgression strategy for simultaneous improvement and genetic dissection of rice seedling CT.ResultsA set of japonica introgression lines (ILs) with significantly improved seedling CT were developed from four backcross populations based on two rounds of selection. Genetic characterization of these cold-tolerant ILs revealed two important aspects of genome-wide responses to strong phenotypic selection for rice CT: (1) significant over-introgression of donor alleles at 57 loci in 29 functional genetic units (FGUs) across the rice genome and (2) pronounced non-random associations between or among alleles at many unlinked CT loci. Linkage disequilibrium analyses of the detected CT loci allowed us to construct putative genetic networks (multi-locus structures) underlying the seedling CT of rice. Each network consisted of a single FGU, with high introgression as the putative regulator plus two to three groups of highly associated downstream FGUs. A bioinformatics search of rice genomic regions harboring these putative regulators identified a small set of candidate regulatory genes that are known to be involved in plant stress response.ConclusionsOur results suggest that CT in rice is controlled by multiple pathways. Genetic complementarity between parental-derived functional alleles at many loci within a given pathway provides an appropriate explanation for the commonly observed hidden diversity and transgressive segregation of CT and other complex traits in rice.

Genome‐Wide Responses to Selection and Genetic Networks Underlying Submergence Tolerance in Rice

“…The third feature of the genetic networks was that ST FGUs in the downstream of the genetic networks accounted for the majority of detected ST FGUs or loci (Supplemental Table S6), meeting the theoretical expectation and principle of hierarchy (Zhang et al, 2011). However, while most of the downstream loci were detected in multiple populations, their associations with those putative regulators were less consistent across different genetic backgrounds.…”

“…Thus, those loci in the same AGs or highly associated FGUs in the same branches of the detected genetic networks were most likely loci involved in the same positively regulated pathway for enhanced ST, which were co‐responding to selection. However, the level of NRAs and dramatically increased homozygosity or loss of heterozygosity (LOH) at the detected ST loci in either BC 2 F 2 or BC 3 F 2 ILs were much stronger than the theoretical expectations (Zhang et al, 2011) and could not be easily explained by phenotypic selection acting on ST alleles of incomplete dominance assuming the Mendelian segregation. In particular, AGs involving many unlinked loci with high IF and complete homozygosity, such as agST J3 , agST K1 , and agST J5 , were extremely unlikely in the two BC 3 F 2 populations under the Mendelian segregation.…”

“…Based on the LD results (Supplemental Table S2) and the principle of hierarchy (Zhang et al, 2011), we were able to construct putative genetic networks, each containing all ST FGUs identified in ILs from each BC population and their corresponding graphical genotypes at the detected ST FGUs (Fig. 2A–K).…”

“…Loci with thicker lines were those of high introgression and putative activators in the upstream of the genetic networks (Fig. 2a‐l) and those with an asterisk (*) are putative repressors according to the molecular‐quantitative genetics theory (Zhang et al, 2011). The large rectangular boxes cover regions where important putative regulatory loci were detected in multiple populations and searched for positional regulatory candidate genes (Supplemental Table S4) from the NCBI database of the annotated rice genome.…”

“…Two concepts, FGUs and the principle of hierarchy, are defined here based on the two most common types of functional relationships between genes acting in the same signaling pathways affecting complex traits (Zhang et al, 2011). Hierarchy reflects the one‐way functional dependency (FD) of genes in downstream pathways on their upstream regulators, while an FGU represents the mutual FD among a group of genes functioning at each level of signaling pathways, which affect phenotypes in a manner of “house of cards” or complete complementarity.…”