Identification and characterization of quantitative trait loci for grain yield and its components under different nitrogen fertilization levels in rice (Oryza sativa L.)

A set of rice (Oryza sativa L.) recombinant inbred lines from a cross between Zhenshan 97 (indica) and HR5 (indica) was planted for four different growing seasons in two locations at three nitrogen (N) fertilization levels (N300, 300 kg urea/ha; N150, 150 kg urea/ha; and N0, 0 kg urea/ha). Grain yield and its components were evaluated, including grain yield per plant (GYPP), panicle number per plant (PNPP), grain number per panicle (GNPP), filled grains per panicle (FGPP), spikelet fertility percentage (SFP) and 100-grain weight (HGW). Correlation and path analysis indicated that SFP had the greatest contribution to GYPP at the N300 and N150 levels, but FGPP contributed the most to GYPP at the N0 level. Quantitative trait loci (QTL) were mapped based on a mixed linear model; genetic components (main effects, epistatic effects and QTL-by-environment interactions) were estimated separately. Six to 15 QTL with main effects were detected for each trait except SFP. Clusters of main-effect QTL associated with PNPP, GNPP, SFP and HGW were observed in regions on chromosomes 1, 2, 3, 5, 7 and 10. The main-effect QTL (qGYPP-4b and qGNPP-12) were only detected at the N0 level and explained 10.9 and 10.2% of the total phenotypic variation, respectively. A total of 33 digenic interactions among grain yield and its components were also identified. The identification of genomic regions associated with yield and its components at different nitrogen levels will be useful in marker-assisted selection for improving the nitrogen use efficiency of rice.