Studies On QTL Detection Methods Based On Experimental Design For A Large Number Of Plant Varieties

Studies on the methods of quantitative trait locus (QTL) analysis are of great significance due to the important economic value of quantitative traits. Present work aimed at developmenet of new methods with high QTL detection power in experiment design. To control experimental error, two experimental designs for a large number of plant lines are adopted. By means of theoretical deduction and simulation, the proposed methods of QTL detection with high precision have been validated. Main results are as follows:1) A QTL detection method based on Lattice design:Lattice design can control experimental error for trials with a huge amount of genotypes. By taking the advantage of error reduction of Lattice design, we proposed a composite interval mapping model for QTL detection. We compared the precision of QTL detection under Lattice design with that under randomized block design by simulation experiments in which genomic loci, QTL effect, block error and other conditions were simulated. The results indicated that the QTL detection power of Lattice design is much larger in comparison with the QTL detection model of randomized block design, and the false discovery rate (FDR) of QTL model with Lattice design is much smaller. As an example, the RIL population experimental data for soybean biomass trait was analyzed with the Lattice design model; the resusts showed that the new method can detect more QTLs than methods with average phenotypic value as respond variable for QTL detection. This may probably be caused by efficiently controlling of experimental error.2) A QTL detection method based on block in replication design:Block in replication design can also control experimental error for trials with a huge amount of genotypes. By taking advantage of error reduction of Block in replication design, we suggest a composite interval mapping model for QTL detection. We compared the precision of QTL detection under Block in replication design with that under randomized block design by simulation experiments in which genomic loci, QTL effect, block error and other conditions were simulated. Simulation results indicated that the QTL detection power of Block in replication design is much larger than that with the QTL detection model of randomized block design, and the FDR of QTL model under Block in replication design is much smaller.We proposed the two composite interval mapping models under experimental design that can efficiently control experimental errors. By using the two proposed models, the precision of QTL detection can be improved. Meanwhile, the simulation results indicated that controlling experimental error is very important for QTL mapping.