View source for IC4R007-GWAS-2015-25785447

== Project Title ==
''' Genome-Wide Association Mapping for Yield and Other Agronomic Traits in an Elite Breeding Population of Tropical Rice (Oryza sativa) '''

==The Background of This Project==
[[File:IC4R007-GWAS-2015-25785447-1.PNG|500px|thumb|right|'''Figure 1''' ''Selected Manhattan plots for flowering time (FLW, top), length-breadth ratio (LBR, top middle), plant height (PH, bottom middle), and grain yield (YLD, bottom). Dashed line shows the 0.1 FDR '']]
* Genome-wide association mapping studies (GWAS) are frequently used to detect QTL in diverse collections of crop germplasm, based on historic recombination events and linkage disequilibrium across the genome. Generally, diversity panels genotyped with high density SNP panels are utilized in order to assay a wide range of alleles and haplotypes and to monitor recombination breakpoints across the genome. By contrast, GWAS have not generally been performed in breeding populations.
* Developing new rice varieties that yield well with fewer inputs and under more stressful and unpredictable climatic conditions is essential for the future of food security, and is the major challenge for today's rice breeders. Fortunately, the rapid development of new sequencing technologies has created the opportunity to enhance our understanding of the genetic basis of crop productivity. The utilization of this genetic information offers the plant breeding community a range of modern tools and methods for addressing these challenges.
* Genome wide association studies (GWAS) have been widely used to identify QTL underlying quantitative traits in humans and animals, and has recently also become a popular method of mapping QTL in plants. Association mapping identifies QTL based on the historic recombination in a panel of diverse germplasm via the presence of linkage disequilibrium (LD) between SNPs and QTL, i.e., the non-random association of alleles. A high density marker panel that covers the genome is required in order to monitor the density of recombination breakpoints in the population.
* Genome-wide prediction, or genomic selection (GS), refers to the process of using genome-wide DNA markers to predict which individuals in a breeding population are most valuable as parents of the next generation of offspring. GS takes the same inputs as GWAS, a phenotype dataset and genotype dataset on a population of lines of interest to plant breeders. As such, it is possible to perform GWAS and GS on the same population, where all that is needed is additional computation analysis. Such an undertaking has clear advantages. The genetic architecture revealed by association mapping can be used to inform the GS models—for example, if highly significant SNPs are revealed by a GWAS, these SNPs could be fit as fixed effects in a GS model, and experimenting with different types of genomic selection statistical methods (i.e., linear versus non-linear, additive versus non-additive) can corroborate inferences about the genetic architecture of a trait.

==Plant Culture & Treatment==
* 363 elite breeding lines were selected for genotyping from the International Rice Research Institute (IRRI) irrigated rice breeding program based on the planned inclusion of the lines in the 2011 Multi-Environment Testing Program and in the 2011 and 2012 Replicated Yield Trials (RYT) at IRRI (Los Baños). Approximately half of the lines were also included in the 2009–2010 RYTs at IRRI (S1 Table). The other lines were promoted from the observational yield trial (OYT) to the RYT in 2011. The lines were all derived from the pedigree breeding method.
* A total of 19 agronomic, morphological, grain and yield-related traits were evaluated in the panel in 2012 (dry and wet seasons). A detailed analysis of these traits will be reported elsewhere. Briefly, all measurements were performed using two replications with 5 plants per plot, except plot yield, which was measured using a 6 m 2 harvest area. All measurements were made on a quantitative scale, with the exception of lodging score and pani- cle exertion rate, which were nominal. Methods used followed IRRI’s standard evaluation system (SES) or other routinely used protocols. Predicted means and variance components were calculated using linear mixed models in Genstat v. 16. Entries were considered as fixed effects and replication and entry by replication interactions were considered as random effects. A detailed trait analysis will be reported elsewhere.

==Research Findings==
* A total of 52 QTL were identified for 11 of the 19 agronomic traits evaluated in this study. Peaks for the other eight traits did not pass the significance threshold as determined by a false discovery rate (FDR) of 0.1; all traits were evaluated during both the 2012 dry and wet seasons in Los Baños, Philippines ('''Figures. 1–2''') (materials and methods). The Manhattan plots for flowering time (FLW), plant height (PH), and length-breadth ratio (LBR) in the dry season and yield (YLD) in the wet season are shown in '''Figure. 1'''.
[[File:IC4R007-GWAS-2015-25785447-2.PNG|1000px|thumb|center|'''Figure 2''' '' Physical map of significant GWAS QTL. '']]
* Performing association mapping on a panel of adapted breeding lines rather than on a diversity panel provides the opportunity to apply the results directly to breeding programs. Unlike the results from studies using diversity panels, our association mapping results can be readily used to identify favorable or unfavorable haplotypes that are currently segregating in our elite breeding material. These haplotypes could be used to determine the most suitable parents for crossing in order to exploit transgressive segregation and/or to increase the frequency with which favorable haplotypes appear in the progeny. MAS for favorable haplotypes among the progeny would allow us to increase breeding efficiency and decrease cost by reducing the number of plants advanced to the next generation of breeding or that need to be phenotyped. In this way, we aim to increase the rate of genetic improvement by increasing gain from selection.
* The lack of perfect association between the haplotypes described above and their respective phenotypes, as well as the limited number of QTL identified for many of the phenotypes in this study, highlight the complexity of the genetic architecture underlying many agronomic traits in rice. Even the haplotypes for flowering time, for which we identified a large effect QTL, did not explain 100% of the phenotypic variance. This poses a problem for the implementation of MAS in rice breeding programs. While genetic gain could potentially be increased over phenotypic selection alone, the probability of eliminating favorable individuals or selecting unfavorable individuals will limit overall breeding progress.

== Labs working on this Project ==
* International Rice Research Institute, Los Baños, Philippines
* Department of Plant Breeding and Genetics, Cornell University, Ithaca, NY, United States of America
* Crop Science Cluster, University of the Philippines Los Baños, Los Baños, Philippines
* International Center for Tropical Agriculture, Cali, Colombia

==Corresponding Author==
* ''' Susan R. McCouch'''(srm4@cornell.edu)